X-Git-Url: http://git.droids-corp.org/?a=blobdiff_plain;f=drivers%2Fnet%2Fmlx5%2Fmlx5_flow.c;h=c914a7120cca61e092ee7c5b9b7525bd579fc9fd;hb=250e2ed8d85d038ce864052ebd6f9af51db40df2;hp=0ee34001a0f6a2a5c2f4114f5ef8501cf2473aff;hpb=1d1f909cf52d160227a0c87f9e95d2734cc4c58e;p=dpdk.git diff --git a/drivers/net/mlx5/mlx5_flow.c b/drivers/net/mlx5/mlx5_flow.c index 0ee34001a0..c914a7120c 100644 --- a/drivers/net/mlx5/mlx5_flow.c +++ b/drivers/net/mlx5/mlx5_flow.c @@ -3,16 +3,15 @@ * Copyright 2016 Mellanox Technologies, Ltd */ -#include -#include #include #include #include #include +#include #include #include -#include +#include #include #include #include @@ -29,20 +28,43 @@ #include "mlx5.h" #include "mlx5_flow.h" #include "mlx5_flow_os.h" -#include "mlx5_rxtx.h" +#include "mlx5_rx.h" +#include "mlx5_tx.h" #include "mlx5_common_os.h" #include "rte_pmd_mlx5.h" +struct tunnel_default_miss_ctx { + uint16_t *queue; + __extension__ + union { + struct rte_flow_action_rss action_rss; + struct rte_flow_action_queue miss_queue; + struct rte_flow_action_jump miss_jump; + uint8_t raw[0]; + }; +}; + +static int +flow_tunnel_add_default_miss(struct rte_eth_dev *dev, + struct rte_flow *flow, + const struct rte_flow_attr *attr, + const struct rte_flow_action *app_actions, + uint32_t flow_idx, + const struct mlx5_flow_tunnel *tunnel, + struct tunnel_default_miss_ctx *ctx, + struct rte_flow_error *error); static struct mlx5_flow_tunnel * mlx5_find_tunnel_id(struct rte_eth_dev *dev, uint32_t id); static void mlx5_flow_tunnel_free(struct rte_eth_dev *dev, struct mlx5_flow_tunnel *tunnel); -static const struct mlx5_flow_tbl_data_entry * -tunnel_mark_decode(struct rte_eth_dev *dev, uint32_t mark); -static int -mlx5_get_flow_tunnel(struct rte_eth_dev *dev, - const struct rte_flow_tunnel *app_tunnel, - struct mlx5_flow_tunnel **tunnel); +static uint32_t +tunnel_flow_group_to_flow_table(struct rte_eth_dev *dev, + const struct mlx5_flow_tunnel *tunnel, + uint32_t group, uint32_t *table, + struct rte_flow_error *error); + +static struct mlx5_flow_workspace *mlx5_flow_push_thread_workspace(void); +static void mlx5_flow_pop_thread_workspace(void); /** Device flow drivers. */ @@ -52,7 +74,7 @@ const struct mlx5_flow_driver_ops mlx5_flow_null_drv_ops; const struct mlx5_flow_driver_ops *flow_drv_ops[] = { [MLX5_FLOW_TYPE_MIN] = &mlx5_flow_null_drv_ops, -#ifdef HAVE_IBV_FLOW_DV_SUPPORT +#if defined(HAVE_IBV_FLOW_DV_SUPPORT) || !defined(HAVE_INFINIBAND_VERBS_H) [MLX5_FLOW_TYPE_DV] = &mlx5_flow_dv_drv_ops, #endif [MLX5_FLOW_TYPE_VERBS] = &mlx5_flow_verbs_drv_ops, @@ -78,8 +100,25 @@ struct mlx5_flow_expand_node { * RSS types bit-field associated with this node * (see ETH_RSS_* definitions). */ + uint64_t node_flags; + /**< + * Bit-fields that define how the node is used in the expansion. + * (see MLX5_EXPANSION_NODE_* definitions). + */ }; +/* Optional expand field. The expansion alg will not go deeper. */ +#define MLX5_EXPANSION_NODE_OPTIONAL (UINT64_C(1) << 0) + +/* The node is not added implicitly as expansion to the flow pattern. + * If the node type does not match the flow pattern item type, the + * expansion alg will go deeper to its next items. + * In the current implementation, the list of next nodes indexes can + * have up to one node with this flag set and it has to be the last + * node index (before the list terminator). + */ +#define MLX5_EXPANSION_NODE_EXPLICIT (UINT64_C(1) << 1) + /** Object returned by mlx5_flow_expand_rss(). */ struct mlx5_flow_expand_rss { uint32_t entries; @@ -90,6 +129,35 @@ struct mlx5_flow_expand_rss { } entry[]; }; +static void +mlx5_dbg__print_pattern(const struct rte_flow_item *item); + +static bool +mlx5_flow_is_rss_expandable_item(const struct rte_flow_item *item) +{ + switch (item->type) { + case RTE_FLOW_ITEM_TYPE_ETH: + case RTE_FLOW_ITEM_TYPE_VLAN: + case RTE_FLOW_ITEM_TYPE_IPV4: + case RTE_FLOW_ITEM_TYPE_IPV6: + case RTE_FLOW_ITEM_TYPE_UDP: + case RTE_FLOW_ITEM_TYPE_TCP: + case RTE_FLOW_ITEM_TYPE_VXLAN: + case RTE_FLOW_ITEM_TYPE_NVGRE: + case RTE_FLOW_ITEM_TYPE_GRE: + case RTE_FLOW_ITEM_TYPE_GENEVE: + case RTE_FLOW_ITEM_TYPE_MPLS: + case RTE_FLOW_ITEM_TYPE_VXLAN_GPE: + case RTE_FLOW_ITEM_TYPE_GRE_KEY: + case RTE_FLOW_ITEM_TYPE_IPV6_FRAG_EXT: + case RTE_FLOW_ITEM_TYPE_GTP: + return true; + default: + break; + } + return false; +} + static enum rte_flow_item_type mlx5_flow_expand_rss_item_complete(const struct rte_flow_item *item) { @@ -191,6 +259,28 @@ mlx5_flow_expand_rss_item_complete(const struct rte_flow_item *item) return ret; } +static const int * +mlx5_flow_expand_rss_skip_explicit(const struct mlx5_flow_expand_node graph[], + const int *next_node) +{ + const struct mlx5_flow_expand_node *node = NULL; + const int *next = next_node; + + while (next && *next) { + /* + * Skip the nodes with the MLX5_EXPANSION_NODE_EXPLICIT + * flag set, because they were not found in the flow pattern. + */ + node = &graph[*next]; + if (!(node->node_flags & MLX5_EXPANSION_NODE_EXPLICIT)) + break; + next = node->next; + } + return next; +} + +#define MLX5_RSS_EXP_ELT_N 16 + /** * Expand RSS flows into several possible flows according to the RSS hash * fields requested and the driver capabilities. @@ -214,6 +304,7 @@ mlx5_flow_expand_rss_item_complete(const struct rte_flow_item *item) * set, the following errors are defined: * * -E2BIG: graph-depth @p graph is too deep. + * -EINVAL: @p size has not enough space for expanded pattern. */ static int mlx5_flow_expand_rss(struct mlx5_flow_expand_rss *buf, size_t size, @@ -221,13 +312,12 @@ mlx5_flow_expand_rss(struct mlx5_flow_expand_rss *buf, size_t size, const struct mlx5_flow_expand_node graph[], int graph_root_index) { - const int elt_n = 8; const struct rte_flow_item *item; const struct mlx5_flow_expand_node *node = &graph[graph_root_index]; const int *next_node; - const int *stack[elt_n]; + const int *stack[MLX5_RSS_EXP_ELT_N]; int stack_pos = 0; - struct rte_flow_item flow_items[elt_n]; + struct rte_flow_item flow_items[MLX5_RSS_EXP_ELT_N]; unsigned int i; size_t lsize; size_t user_pattern_size = 0; @@ -240,20 +330,30 @@ mlx5_flow_expand_rss(struct mlx5_flow_expand_rss *buf, size_t size, memset(&missed_item, 0, sizeof(missed_item)); lsize = offsetof(struct mlx5_flow_expand_rss, entry) + - elt_n * sizeof(buf->entry[0]); - if (lsize <= size) { - buf->entry[0].priority = 0; - buf->entry[0].pattern = (void *)&buf->entry[elt_n]; - buf->entries = 0; - addr = buf->entry[0].pattern; - } + MLX5_RSS_EXP_ELT_N * sizeof(buf->entry[0]); + if (lsize > size) + return -EINVAL; + buf->entry[0].priority = 0; + buf->entry[0].pattern = (void *)&buf->entry[MLX5_RSS_EXP_ELT_N]; + buf->entries = 0; + addr = buf->entry[0].pattern; for (item = pattern; item->type != RTE_FLOW_ITEM_TYPE_END; item++) { - if (item->type != RTE_FLOW_ITEM_TYPE_VOID) - last_item = item; - for (i = 0; node->next && node->next[i]; ++i) { + if (!mlx5_flow_is_rss_expandable_item(item)) { + user_pattern_size += sizeof(*item); + continue; + } + last_item = item; + i = 0; + while (node->next && node->next[i]) { next = &graph[node->next[i]]; if (next->type == item->type) break; + if (next->node_flags & MLX5_EXPANSION_NODE_EXPLICIT) { + node = next; + i = 0; + } else { + ++i; + } } if (next) node = next; @@ -261,12 +361,12 @@ mlx5_flow_expand_rss(struct mlx5_flow_expand_rss *buf, size_t size, } user_pattern_size += sizeof(*item); /* Handle END item. */ lsize += user_pattern_size; + if (lsize > size) + return -EINVAL; /* Copy the user pattern in the first entry of the buffer. */ - if (lsize <= size) { - rte_memcpy(addr, pattern, user_pattern_size); - addr = (void *)(((uintptr_t)addr) + user_pattern_size); - buf->entries = 1; - } + rte_memcpy(addr, pattern, user_pattern_size); + addr = (void *)(((uintptr_t)addr) + user_pattern_size); + buf->entries = 1; /* Start expanding. */ memset(flow_items, 0, sizeof(flow_items)); user_pattern_size -= sizeof(*item); @@ -296,7 +396,9 @@ mlx5_flow_expand_rss(struct mlx5_flow_expand_rss *buf, size_t size, elt = 2; /* missed item + item end. */ node = next; lsize += elt * sizeof(*item) + user_pattern_size; - if ((node->rss_types & types) && lsize <= size) { + if (lsize > size) + return -EINVAL; + if (node->rss_types & types) { buf->entry[buf->entries].priority = 1; buf->entry[buf->entries].pattern = addr; buf->entries++; @@ -309,12 +411,14 @@ mlx5_flow_expand_rss(struct mlx5_flow_expand_rss *buf, size_t size, } } memset(flow_items, 0, sizeof(flow_items)); - next_node = node->next; + next_node = mlx5_flow_expand_rss_skip_explicit(graph, + node->next); stack[stack_pos] = next_node; node = next_node ? &graph[*next_node] : NULL; while (node) { flow_items[stack_pos].type = node->type; if (node->rss_types & types) { + size_t n; /* * compute the number of items to copy from the * expansion and copy it. @@ -324,73 +428,65 @@ mlx5_flow_expand_rss(struct mlx5_flow_expand_rss *buf, size_t size, elt = stack_pos + 2; flow_items[stack_pos + 1].type = RTE_FLOW_ITEM_TYPE_END; lsize += elt * sizeof(*item) + user_pattern_size; - if (lsize <= size) { - size_t n = elt * sizeof(*item); - - buf->entry[buf->entries].priority = - stack_pos + 1 + missed; - buf->entry[buf->entries].pattern = addr; - buf->entries++; - rte_memcpy(addr, buf->entry[0].pattern, - user_pattern_size); - addr = (void *)(((uintptr_t)addr) + - user_pattern_size); - rte_memcpy(addr, &missed_item, - missed * sizeof(*item)); - addr = (void *)(((uintptr_t)addr) + - missed * sizeof(*item)); - rte_memcpy(addr, flow_items, n); - addr = (void *)(((uintptr_t)addr) + n); - } + if (lsize > size) + return -EINVAL; + n = elt * sizeof(*item); + buf->entry[buf->entries].priority = + stack_pos + 1 + missed; + buf->entry[buf->entries].pattern = addr; + buf->entries++; + rte_memcpy(addr, buf->entry[0].pattern, + user_pattern_size); + addr = (void *)(((uintptr_t)addr) + + user_pattern_size); + rte_memcpy(addr, &missed_item, + missed * sizeof(*item)); + addr = (void *)(((uintptr_t)addr) + + missed * sizeof(*item)); + rte_memcpy(addr, flow_items, n); + addr = (void *)(((uintptr_t)addr) + n); } /* Go deeper. */ - if (node->next) { - next_node = node->next; - if (stack_pos++ == elt_n) { + if (!(node->node_flags & MLX5_EXPANSION_NODE_OPTIONAL) && + node->next) { + next_node = mlx5_flow_expand_rss_skip_explicit(graph, + node->next); + if (stack_pos++ == MLX5_RSS_EXP_ELT_N) { rte_errno = E2BIG; return -rte_errno; } stack[stack_pos] = next_node; } else if (*(next_node + 1)) { /* Follow up with the next possibility. */ + next_node = mlx5_flow_expand_rss_skip_explicit(graph, + ++next_node); + } else if (!stack_pos) { + /* + * Completing the traverse over the different paths. + * The next_node is advanced to the terminator. + */ ++next_node; } else { /* Move to the next path. */ - if (stack_pos) + while (stack_pos) { next_node = stack[--stack_pos]; - next_node++; + next_node++; + if (*next_node) + break; + } + next_node = mlx5_flow_expand_rss_skip_explicit(graph, + next_node); stack[stack_pos] = next_node; } - node = *next_node ? &graph[*next_node] : NULL; + node = next_node && *next_node ? &graph[*next_node] : NULL; }; - /* no expanded flows but we have missed item, create one rule for it */ - if (buf->entries == 1 && missed != 0) { - elt = 2; - lsize += elt * sizeof(*item) + user_pattern_size; - if (lsize <= size) { - buf->entry[buf->entries].priority = 1; - buf->entry[buf->entries].pattern = addr; - buf->entries++; - flow_items[0].type = missed_item.type; - flow_items[1].type = RTE_FLOW_ITEM_TYPE_END; - rte_memcpy(addr, buf->entry[0].pattern, - user_pattern_size); - addr = (void *)(((uintptr_t)addr) + user_pattern_size); - rte_memcpy(addr, flow_items, elt * sizeof(*item)); - addr = (void *)(((uintptr_t)addr) + - elt * sizeof(*item)); - } - } return lsize; } enum mlx5_expansion { MLX5_EXPANSION_ROOT, MLX5_EXPANSION_ROOT_OUTER, - MLX5_EXPANSION_ROOT_ETH_VLAN, - MLX5_EXPANSION_ROOT_OUTER_ETH_VLAN, MLX5_EXPANSION_OUTER_ETH, - MLX5_EXPANSION_OUTER_ETH_VLAN, MLX5_EXPANSION_OUTER_VLAN, MLX5_EXPANSION_OUTER_IPV4, MLX5_EXPANSION_OUTER_IPV4_UDP, @@ -401,9 +497,10 @@ enum mlx5_expansion { MLX5_EXPANSION_VXLAN, MLX5_EXPANSION_VXLAN_GPE, MLX5_EXPANSION_GRE, + MLX5_EXPANSION_NVGRE, + MLX5_EXPANSION_GRE_KEY, MLX5_EXPANSION_MPLS, MLX5_EXPANSION_ETH, - MLX5_EXPANSION_ETH_VLAN, MLX5_EXPANSION_VLAN, MLX5_EXPANSION_IPV4, MLX5_EXPANSION_IPV4_UDP, @@ -411,6 +508,8 @@ enum mlx5_expansion { MLX5_EXPANSION_IPV6, MLX5_EXPANSION_IPV6_UDP, MLX5_EXPANSION_IPV6_TCP, + MLX5_EXPANSION_IPV6_FRAG_EXT, + MLX5_EXPANSION_GTP }; /** Supported expansion of items. */ @@ -427,23 +526,7 @@ static const struct mlx5_flow_expand_node mlx5_support_expansion[] = { MLX5_EXPANSION_OUTER_IPV6), .type = RTE_FLOW_ITEM_TYPE_END, }, - [MLX5_EXPANSION_ROOT_ETH_VLAN] = { - .next = MLX5_FLOW_EXPAND_RSS_NEXT(MLX5_EXPANSION_ETH_VLAN), - .type = RTE_FLOW_ITEM_TYPE_END, - }, - [MLX5_EXPANSION_ROOT_OUTER_ETH_VLAN] = { - .next = MLX5_FLOW_EXPAND_RSS_NEXT - (MLX5_EXPANSION_OUTER_ETH_VLAN), - .type = RTE_FLOW_ITEM_TYPE_END, - }, [MLX5_EXPANSION_OUTER_ETH] = { - .next = MLX5_FLOW_EXPAND_RSS_NEXT(MLX5_EXPANSION_OUTER_IPV4, - MLX5_EXPANSION_OUTER_IPV6, - MLX5_EXPANSION_MPLS), - .type = RTE_FLOW_ITEM_TYPE_ETH, - .rss_types = 0, - }, - [MLX5_EXPANSION_OUTER_ETH_VLAN] = { .next = MLX5_FLOW_EXPAND_RSS_NEXT(MLX5_EXPANSION_OUTER_VLAN), .type = RTE_FLOW_ITEM_TYPE_ETH, .rss_types = 0, @@ -452,12 +535,14 @@ static const struct mlx5_flow_expand_node mlx5_support_expansion[] = { .next = MLX5_FLOW_EXPAND_RSS_NEXT(MLX5_EXPANSION_OUTER_IPV4, MLX5_EXPANSION_OUTER_IPV6), .type = RTE_FLOW_ITEM_TYPE_VLAN, + .node_flags = MLX5_EXPANSION_NODE_EXPLICIT, }, [MLX5_EXPANSION_OUTER_IPV4] = { .next = MLX5_FLOW_EXPAND_RSS_NEXT (MLX5_EXPANSION_OUTER_IPV4_UDP, MLX5_EXPANSION_OUTER_IPV4_TCP, MLX5_EXPANSION_GRE, + MLX5_EXPANSION_NVGRE, MLX5_EXPANSION_IPV4, MLX5_EXPANSION_IPV6), .type = RTE_FLOW_ITEM_TYPE_IPV4, @@ -466,7 +551,9 @@ static const struct mlx5_flow_expand_node mlx5_support_expansion[] = { }, [MLX5_EXPANSION_OUTER_IPV4_UDP] = { .next = MLX5_FLOW_EXPAND_RSS_NEXT(MLX5_EXPANSION_VXLAN, - MLX5_EXPANSION_VXLAN_GPE), + MLX5_EXPANSION_VXLAN_GPE, + MLX5_EXPANSION_MPLS, + MLX5_EXPANSION_GTP), .type = RTE_FLOW_ITEM_TYPE_UDP, .rss_types = ETH_RSS_NONFRAG_IPV4_UDP, }, @@ -479,14 +566,18 @@ static const struct mlx5_flow_expand_node mlx5_support_expansion[] = { (MLX5_EXPANSION_OUTER_IPV6_UDP, MLX5_EXPANSION_OUTER_IPV6_TCP, MLX5_EXPANSION_IPV4, - MLX5_EXPANSION_IPV6), + MLX5_EXPANSION_IPV6, + MLX5_EXPANSION_GRE, + MLX5_EXPANSION_NVGRE), .type = RTE_FLOW_ITEM_TYPE_IPV6, .rss_types = ETH_RSS_IPV6 | ETH_RSS_FRAG_IPV6 | ETH_RSS_NONFRAG_IPV6_OTHER, }, [MLX5_EXPANSION_OUTER_IPV6_UDP] = { .next = MLX5_FLOW_EXPAND_RSS_NEXT(MLX5_EXPANSION_VXLAN, - MLX5_EXPANSION_VXLAN_GPE), + MLX5_EXPANSION_VXLAN_GPE, + MLX5_EXPANSION_MPLS, + MLX5_EXPANSION_GTP), .type = RTE_FLOW_ITEM_TYPE_UDP, .rss_types = ETH_RSS_NONFRAG_IPV6_UDP, }, @@ -507,20 +598,31 @@ static const struct mlx5_flow_expand_node mlx5_support_expansion[] = { .type = RTE_FLOW_ITEM_TYPE_VXLAN_GPE, }, [MLX5_EXPANSION_GRE] = { - .next = MLX5_FLOW_EXPAND_RSS_NEXT(MLX5_EXPANSION_IPV4), + .next = MLX5_FLOW_EXPAND_RSS_NEXT(MLX5_EXPANSION_IPV4, + MLX5_EXPANSION_IPV6, + MLX5_EXPANSION_GRE_KEY, + MLX5_EXPANSION_MPLS), .type = RTE_FLOW_ITEM_TYPE_GRE, }, + [MLX5_EXPANSION_GRE_KEY] = { + .next = MLX5_FLOW_EXPAND_RSS_NEXT(MLX5_EXPANSION_IPV4, + MLX5_EXPANSION_IPV6, + MLX5_EXPANSION_MPLS), + .type = RTE_FLOW_ITEM_TYPE_GRE_KEY, + .node_flags = MLX5_EXPANSION_NODE_OPTIONAL, + }, + [MLX5_EXPANSION_NVGRE] = { + .next = MLX5_FLOW_EXPAND_RSS_NEXT(MLX5_EXPANSION_ETH), + .type = RTE_FLOW_ITEM_TYPE_NVGRE, + }, [MLX5_EXPANSION_MPLS] = { .next = MLX5_FLOW_EXPAND_RSS_NEXT(MLX5_EXPANSION_IPV4, - MLX5_EXPANSION_IPV6), + MLX5_EXPANSION_IPV6, + MLX5_EXPANSION_ETH), .type = RTE_FLOW_ITEM_TYPE_MPLS, + .node_flags = MLX5_EXPANSION_NODE_OPTIONAL, }, [MLX5_EXPANSION_ETH] = { - .next = MLX5_FLOW_EXPAND_RSS_NEXT(MLX5_EXPANSION_IPV4, - MLX5_EXPANSION_IPV6), - .type = RTE_FLOW_ITEM_TYPE_ETH, - }, - [MLX5_EXPANSION_ETH_VLAN] = { .next = MLX5_FLOW_EXPAND_RSS_NEXT(MLX5_EXPANSION_VLAN), .type = RTE_FLOW_ITEM_TYPE_ETH, }, @@ -528,6 +630,7 @@ static const struct mlx5_flow_expand_node mlx5_support_expansion[] = { .next = MLX5_FLOW_EXPAND_RSS_NEXT(MLX5_EXPANSION_IPV4, MLX5_EXPANSION_IPV6), .type = RTE_FLOW_ITEM_TYPE_VLAN, + .node_flags = MLX5_EXPANSION_NODE_EXPLICIT, }, [MLX5_EXPANSION_IPV4] = { .next = MLX5_FLOW_EXPAND_RSS_NEXT(MLX5_EXPANSION_IPV4_UDP, @@ -546,7 +649,8 @@ static const struct mlx5_flow_expand_node mlx5_support_expansion[] = { }, [MLX5_EXPANSION_IPV6] = { .next = MLX5_FLOW_EXPAND_RSS_NEXT(MLX5_EXPANSION_IPV6_UDP, - MLX5_EXPANSION_IPV6_TCP), + MLX5_EXPANSION_IPV6_TCP, + MLX5_EXPANSION_IPV6_FRAG_EXT), .type = RTE_FLOW_ITEM_TYPE_IPV6, .rss_types = ETH_RSS_IPV6 | ETH_RSS_FRAG_IPV6 | ETH_RSS_NONFRAG_IPV6_OTHER, @@ -559,191 +663,61 @@ static const struct mlx5_flow_expand_node mlx5_support_expansion[] = { .type = RTE_FLOW_ITEM_TYPE_TCP, .rss_types = ETH_RSS_NONFRAG_IPV6_TCP, }, + [MLX5_EXPANSION_IPV6_FRAG_EXT] = { + .type = RTE_FLOW_ITEM_TYPE_IPV6_FRAG_EXT, + }, + [MLX5_EXPANSION_GTP] = { + .next = MLX5_FLOW_EXPAND_RSS_NEXT(MLX5_EXPANSION_IPV4, + MLX5_EXPANSION_IPV6), + .type = RTE_FLOW_ITEM_TYPE_GTP, + }, }; -static struct rte_flow_shared_action * -mlx5_shared_action_create(struct rte_eth_dev *dev, - const struct rte_flow_shared_action_conf *conf, +static struct rte_flow_action_handle * +mlx5_action_handle_create(struct rte_eth_dev *dev, + const struct rte_flow_indir_action_conf *conf, const struct rte_flow_action *action, struct rte_flow_error *error); -static int mlx5_shared_action_destroy +static int mlx5_action_handle_destroy (struct rte_eth_dev *dev, - struct rte_flow_shared_action *shared_action, + struct rte_flow_action_handle *handle, struct rte_flow_error *error); -static int mlx5_shared_action_update +static int mlx5_action_handle_update (struct rte_eth_dev *dev, - struct rte_flow_shared_action *shared_action, - const struct rte_flow_action *action, + struct rte_flow_action_handle *handle, + const void *update, struct rte_flow_error *error); -static int mlx5_shared_action_query +static int mlx5_action_handle_query (struct rte_eth_dev *dev, - const struct rte_flow_shared_action *action, + const struct rte_flow_action_handle *handle, void *data, struct rte_flow_error *error); -static inline bool -mlx5_flow_tunnel_validate(struct rte_eth_dev *dev, - struct rte_flow_tunnel *tunnel, - const char *err_msg) -{ - err_msg = NULL; - if (!is_tunnel_offload_active(dev)) { - err_msg = "tunnel offload was not activated"; - goto out; - } else if (!tunnel) { - err_msg = "no application tunnel"; - goto out; - } - - switch (tunnel->type) { - default: - err_msg = "unsupported tunnel type"; - goto out; - case RTE_FLOW_ITEM_TYPE_VXLAN: - break; - } - -out: - return !err_msg; -} - - static int mlx5_flow_tunnel_decap_set(struct rte_eth_dev *dev, struct rte_flow_tunnel *app_tunnel, struct rte_flow_action **actions, uint32_t *num_of_actions, - struct rte_flow_error *error) -{ - int ret; - struct mlx5_flow_tunnel *tunnel; - const char *err_msg = NULL; - bool verdict = mlx5_flow_tunnel_validate(dev, app_tunnel, err_msg); - - if (!verdict) - return rte_flow_error_set(error, EINVAL, - RTE_FLOW_ERROR_TYPE_ACTION_CONF, NULL, - err_msg); - ret = mlx5_get_flow_tunnel(dev, app_tunnel, &tunnel); - if (ret < 0) { - return rte_flow_error_set(error, ret, - RTE_FLOW_ERROR_TYPE_ACTION_CONF, NULL, - "failed to initialize pmd tunnel"); - } - *actions = &tunnel->action; - *num_of_actions = 1; - return 0; -} - + struct rte_flow_error *error); static int mlx5_flow_tunnel_match(struct rte_eth_dev *dev, struct rte_flow_tunnel *app_tunnel, struct rte_flow_item **items, uint32_t *num_of_items, - struct rte_flow_error *error) -{ - int ret; - struct mlx5_flow_tunnel *tunnel; - const char *err_msg = NULL; - bool verdict = mlx5_flow_tunnel_validate(dev, app_tunnel, err_msg); - - if (!verdict) - return rte_flow_error_set(error, EINVAL, - RTE_FLOW_ERROR_TYPE_HANDLE, NULL, - err_msg); - ret = mlx5_get_flow_tunnel(dev, app_tunnel, &tunnel); - if (ret < 0) { - return rte_flow_error_set(error, ret, - RTE_FLOW_ERROR_TYPE_HANDLE, NULL, - "failed to initialize pmd tunnel"); - } - *items = &tunnel->item; - *num_of_items = 1; - return 0; -} - + struct rte_flow_error *error); static int -mlx5_flow_item_release(struct rte_eth_dev *dev, - struct rte_flow_item *pmd_items, - uint32_t num_items, struct rte_flow_error *err) -{ - struct mlx5_flow_tunnel_hub *thub = mlx5_tunnel_hub(dev); - struct mlx5_flow_tunnel *tun; - - rte_spinlock_lock(&thub->sl); - LIST_FOREACH(tun, &thub->tunnels, chain) { - if (&tun->item == pmd_items) { - LIST_REMOVE(tun, chain); - break; - } - } - rte_spinlock_unlock(&thub->sl); - if (!tun || num_items != 1) - return rte_flow_error_set(err, EINVAL, - RTE_FLOW_ERROR_TYPE_HANDLE, NULL, - "invalid argument"); - if (!__atomic_sub_fetch(&tun->refctn, 1, __ATOMIC_RELAXED)) - mlx5_flow_tunnel_free(dev, tun); - return 0; -} - +mlx5_flow_tunnel_item_release(struct rte_eth_dev *dev, + struct rte_flow_item *pmd_items, + uint32_t num_items, struct rte_flow_error *err); static int -mlx5_flow_action_release(struct rte_eth_dev *dev, - struct rte_flow_action *pmd_actions, - uint32_t num_actions, struct rte_flow_error *err) -{ - struct mlx5_flow_tunnel_hub *thub = mlx5_tunnel_hub(dev); - struct mlx5_flow_tunnel *tun; - - rte_spinlock_lock(&thub->sl); - LIST_FOREACH(tun, &thub->tunnels, chain) { - if (&tun->action == pmd_actions) { - LIST_REMOVE(tun, chain); - break; - } - } - rte_spinlock_unlock(&thub->sl); - if (!tun || num_actions != 1) - return rte_flow_error_set(err, EINVAL, - RTE_FLOW_ERROR_TYPE_HANDLE, NULL, - "invalid argument"); - if (!__atomic_sub_fetch(&tun->refctn, 1, __ATOMIC_RELAXED)) - mlx5_flow_tunnel_free(dev, tun); - - return 0; -} - +mlx5_flow_tunnel_action_release(struct rte_eth_dev *dev, + struct rte_flow_action *pmd_actions, + uint32_t num_actions, + struct rte_flow_error *err); static int mlx5_flow_tunnel_get_restore_info(struct rte_eth_dev *dev, struct rte_mbuf *m, struct rte_flow_restore_info *info, - struct rte_flow_error *err) -{ - uint64_t ol_flags = m->ol_flags; - const struct mlx5_flow_tbl_data_entry *tble; - const uint64_t mask = PKT_RX_FDIR | PKT_RX_FDIR_ID; - - if ((ol_flags & mask) != mask) - goto err; - tble = tunnel_mark_decode(dev, m->hash.fdir.hi); - if (!tble) { - DRV_LOG(DEBUG, "port %u invalid miss tunnel mark %#x", - dev->data->port_id, m->hash.fdir.hi); - goto err; - } - MLX5_ASSERT(tble->tunnel); - memcpy(&info->tunnel, &tble->tunnel->app_tunnel, sizeof(info->tunnel)); - info->group_id = tble->group_id; - info->flags = RTE_FLOW_RESTORE_INFO_TUNNEL | - RTE_FLOW_RESTORE_INFO_GROUP_ID | - RTE_FLOW_RESTORE_INFO_ENCAPSULATED; - - return 0; - -err: - return rte_flow_error_set(err, EINVAL, - RTE_FLOW_ERROR_TYPE_UNSPECIFIED, NULL, - "failed to get restore info"); -} + struct rte_flow_error *err); static const struct rte_flow_ops mlx5_flow_ops = { .validate = mlx5_flow_validate, @@ -754,14 +728,14 @@ static const struct rte_flow_ops mlx5_flow_ops = { .query = mlx5_flow_query, .dev_dump = mlx5_flow_dev_dump, .get_aged_flows = mlx5_flow_get_aged_flows, - .shared_action_create = mlx5_shared_action_create, - .shared_action_destroy = mlx5_shared_action_destroy, - .shared_action_update = mlx5_shared_action_update, - .shared_action_query = mlx5_shared_action_query, + .action_handle_create = mlx5_action_handle_create, + .action_handle_destroy = mlx5_action_handle_destroy, + .action_handle_update = mlx5_action_handle_update, + .action_handle_query = mlx5_action_handle_query, .tunnel_decap_set = mlx5_flow_tunnel_decap_set, .tunnel_match = mlx5_flow_tunnel_match, - .tunnel_action_decap_release = mlx5_flow_action_release, - .tunnel_item_release = mlx5_flow_item_release, + .tunnel_action_decap_release = mlx5_flow_tunnel_action_release, + .tunnel_item_release = mlx5_flow_tunnel_item_release, .get_restore_info = mlx5_flow_tunnel_get_restore_info, }; @@ -814,11 +788,6 @@ static struct mlx5_flow_tunnel_info tunnels_info[] = { }, }; -/* Key of thread specific flow workspace data. */ -static pthread_key_t key_workspace; - -/* Thread specific flow workspace data once initialization data. */ -static pthread_once_t key_workspace_init; /** @@ -885,9 +854,9 @@ mlx5_flow_get_reg_id(struct rte_eth_dev *dev, return REG_C_0; } break; - case MLX5_MTR_SFX: + case MLX5_MTR_ID: /* - * If meter color and flow match share one register, flow match + * If meter color and meter id share one register, flow match * should use the meter color register for match. */ if (priv->mtr_reg_share) @@ -896,6 +865,9 @@ mlx5_flow_get_reg_id(struct rte_eth_dev *dev, return priv->mtr_color_reg != REG_C_2 ? REG_C_2 : REG_C_3; case MLX5_MTR_COLOR: + case MLX5_ASO_FLOW_HIT: + case MLX5_ASO_CONNTRACK: + /* All features use the same REG_C. */ MLX5_ASSERT(priv->mtr_color_reg != REG_NON); return priv->mtr_color_reg; case MLX5_COPY_MARK: @@ -915,7 +887,7 @@ mlx5_flow_get_reg_id(struct rte_eth_dev *dev, start_reg = priv->mtr_color_reg != REG_C_2 ? REG_C_2 : (priv->mtr_reg_share ? REG_C_3 : REG_C_4); skip_mtr_reg = !!(priv->mtr_en && start_reg == REG_C_2); - if (id > (REG_C_7 - start_reg)) + if (id > (uint32_t)(REG_C_7 - start_reg)) return rte_flow_error_set(error, EINVAL, RTE_FLOW_ERROR_TYPE_ITEM, NULL, "invalid tag id"); @@ -931,7 +903,7 @@ mlx5_flow_get_reg_id(struct rte_eth_dev *dev, */ if (skip_mtr_reg && config->flow_mreg_c [id + start_reg - REG_C_0] >= priv->mtr_color_reg) { - if (id >= (REG_C_7 - start_reg)) + if (id >= (uint32_t)(REG_C_7 - start_reg)) return rte_flow_error_set(error, EINVAL, RTE_FLOW_ERROR_TYPE_ITEM, NULL, "invalid tag id"); @@ -978,17 +950,69 @@ mlx5_flow_ext_mreg_supported(struct rte_eth_dev *dev) } /** - * Verify the @p item specifications (spec, last, mask) are compatible with the - * NIC capabilities. + * Get the lowest priority. * - * @param[in] item - * Item specification. - * @param[in] mask - * @p item->mask or flow default bit-masks. - * @param[in] nic_mask - * Bit-masks covering supported fields by the NIC to compare with user mask. - * @param[in] size - * Bit-masks size in bytes. + * @param[in] dev + * Pointer to the Ethernet device structure. + * @param[in] attributes + * Pointer to device flow rule attributes. + * + * @return + * The value of lowest priority of flow. + */ +uint32_t +mlx5_get_lowest_priority(struct rte_eth_dev *dev, + const struct rte_flow_attr *attr) +{ + struct mlx5_priv *priv = dev->data->dev_private; + + if (!attr->group && !attr->transfer) + return priv->config.flow_prio - 2; + return MLX5_NON_ROOT_FLOW_MAX_PRIO - 1; +} + +/** + * Calculate matcher priority of the flow. + * + * @param[in] dev + * Pointer to the Ethernet device structure. + * @param[in] attr + * Pointer to device flow rule attributes. + * @param[in] subpriority + * The priority based on the items. + * @return + * The matcher priority of the flow. + */ +uint16_t +mlx5_get_matcher_priority(struct rte_eth_dev *dev, + const struct rte_flow_attr *attr, + uint32_t subpriority) +{ + uint16_t priority = (uint16_t)attr->priority; + struct mlx5_priv *priv = dev->data->dev_private; + + if (!attr->group && !attr->transfer) { + if (attr->priority == MLX5_FLOW_LOWEST_PRIO_INDICATOR) + priority = priv->config.flow_prio - 1; + return mlx5_os_flow_adjust_priority(dev, priority, subpriority); + } + if (attr->priority == MLX5_FLOW_LOWEST_PRIO_INDICATOR) + priority = MLX5_NON_ROOT_FLOW_MAX_PRIO; + return priority * 3 + subpriority; +} + +/** + * Verify the @p item specifications (spec, last, mask) are compatible with the + * NIC capabilities. + * + * @param[in] item + * Item specification. + * @param[in] mask + * @p item->mask or flow default bit-masks. + * @param[in] nic_mask + * Bit-masks covering supported fields by the NIC to compare with user mask. + * @param[in] size + * Bit-masks size in bytes. * @param[in] range_accepted * True if range of values is accepted for specific fields, false otherwise. * @param[out] error @@ -1112,24 +1136,36 @@ flow_rxq_tunnel_ptype_update(struct mlx5_rxq_ctrl *rxq_ctrl) * @param[in] dev_handle * Pointer to device flow handle structure. */ -static void +void flow_drv_rxq_flags_set(struct rte_eth_dev *dev, struct mlx5_flow_handle *dev_handle) { struct mlx5_priv *priv = dev->data->dev_private; const int mark = dev_handle->mark; const int tunnel = !!(dev_handle->layers & MLX5_FLOW_LAYER_TUNNEL); - struct mlx5_hrxq *hrxq; + struct mlx5_ind_table_obj *ind_tbl = NULL; unsigned int i; - if (dev_handle->fate_action != MLX5_FLOW_FATE_QUEUE) - return; - hrxq = mlx5_ipool_get(priv->sh->ipool[MLX5_IPOOL_HRXQ], + if (dev_handle->fate_action == MLX5_FLOW_FATE_QUEUE) { + struct mlx5_hrxq *hrxq; + + hrxq = mlx5_ipool_get(priv->sh->ipool[MLX5_IPOOL_HRXQ], dev_handle->rix_hrxq); - if (!hrxq) + if (hrxq) + ind_tbl = hrxq->ind_table; + } else if (dev_handle->fate_action == MLX5_FLOW_FATE_SHARED_RSS) { + struct mlx5_shared_action_rss *shared_rss; + + shared_rss = mlx5_ipool_get + (priv->sh->ipool[MLX5_IPOOL_RSS_SHARED_ACTIONS], + dev_handle->rix_srss); + if (shared_rss) + ind_tbl = shared_rss->ind_tbl; + } + if (!ind_tbl) return; - for (i = 0; i != hrxq->ind_table->queues_n; ++i) { - int idx = hrxq->ind_table->queues[i]; + for (i = 0; i != ind_tbl->queues_n; ++i) { + int idx = ind_tbl->queues[i]; struct mlx5_rxq_ctrl *rxq_ctrl = container_of((*priv->rxqs)[idx], struct mlx5_rxq_ctrl, rxq); @@ -1201,18 +1237,30 @@ flow_drv_rxq_flags_trim(struct rte_eth_dev *dev, struct mlx5_priv *priv = dev->data->dev_private; const int mark = dev_handle->mark; const int tunnel = !!(dev_handle->layers & MLX5_FLOW_LAYER_TUNNEL); - struct mlx5_hrxq *hrxq; + struct mlx5_ind_table_obj *ind_tbl = NULL; unsigned int i; - if (dev_handle->fate_action != MLX5_FLOW_FATE_QUEUE) - return; - hrxq = mlx5_ipool_get(priv->sh->ipool[MLX5_IPOOL_HRXQ], + if (dev_handle->fate_action == MLX5_FLOW_FATE_QUEUE) { + struct mlx5_hrxq *hrxq; + + hrxq = mlx5_ipool_get(priv->sh->ipool[MLX5_IPOOL_HRXQ], dev_handle->rix_hrxq); - if (!hrxq) + if (hrxq) + ind_tbl = hrxq->ind_table; + } else if (dev_handle->fate_action == MLX5_FLOW_FATE_SHARED_RSS) { + struct mlx5_shared_action_rss *shared_rss; + + shared_rss = mlx5_ipool_get + (priv->sh->ipool[MLX5_IPOOL_RSS_SHARED_ACTIONS], + dev_handle->rix_srss); + if (shared_rss) + ind_tbl = shared_rss->ind_tbl; + } + if (!ind_tbl) return; MLX5_ASSERT(dev->data->dev_started); - for (i = 0; i != hrxq->ind_table->queues_n; ++i) { - int idx = hrxq->ind_table->queues[i]; + for (i = 0; i != ind_tbl->queues_n; ++i) { + int idx = ind_tbl->queues[i]; struct mlx5_rxq_ctrl *rxq_ctrl = container_of((*priv->rxqs)[idx], struct mlx5_rxq_ctrl, rxq); @@ -1313,10 +1361,12 @@ mlx5_flow_rxq_dynf_metadata_set(struct rte_eth_dev *dev) data->dynf_meta = 0; data->flow_meta_mask = 0; data->flow_meta_offset = -1; + data->flow_meta_port_mask = 0; } else { data->dynf_meta = 1; data->flow_meta_mask = rte_flow_dynf_metadata_mask; data->flow_meta_offset = rte_flow_dynf_metadata_offs; + data->flow_meta_port_mask = priv->sh->dv_meta_mask; } } } @@ -1530,6 +1580,7 @@ mlx5_validate_action_rss(struct rte_eth_dev *dev, { struct mlx5_priv *priv = dev->data->dev_private; const struct rte_flow_action_rss *rss = action->conf; + enum mlx5_rxq_type rxq_type = MLX5_RXQ_TYPE_UNDEFINED; unsigned int i; if (rss->func != RTE_ETH_HASH_FUNCTION_DEFAULT && @@ -1595,6 +1646,8 @@ mlx5_validate_action_rss(struct rte_eth_dev *dev, RTE_FLOW_ERROR_TYPE_ACTION_CONF, NULL, "No queues configured"); for (i = 0; i != rss->queue_num; ++i) { + struct mlx5_rxq_ctrl *rxq_ctrl; + if (rss->queue[i] >= priv->rxqs_n) return rte_flow_error_set (error, EINVAL, @@ -1604,6 +1657,15 @@ mlx5_validate_action_rss(struct rte_eth_dev *dev, return rte_flow_error_set (error, EINVAL, RTE_FLOW_ERROR_TYPE_ACTION_CONF, &rss->queue[i], "queue is not configured"); + rxq_ctrl = container_of((*priv->rxqs)[rss->queue[i]], + struct mlx5_rxq_ctrl, rxq); + if (i == 0) + rxq_type = rxq_ctrl->type; + if (rxq_type != rxq_ctrl->type) + return rte_flow_error_set + (error, ENOTSUP, RTE_FLOW_ERROR_TYPE_ACTION_CONF, + &rss->queue[i], + "combining hairpin and regular RSS queues is not supported"); } return 0; } @@ -1663,6 +1725,13 @@ mlx5_flow_validate_action_rss(const struct rte_flow_action *action, RTE_FLOW_ERROR_TYPE_ACTION_CONF, NULL, "RSS on eCPRI is not supported now"); } + if ((item_flags & MLX5_FLOW_LAYER_MPLS) && + !(item_flags & + (MLX5_FLOW_LAYER_INNER_L2 | MLX5_FLOW_LAYER_INNER_L3)) && + rss->level > 1) + return rte_flow_error_set(error, EINVAL, + RTE_FLOW_ERROR_TYPE_ITEM, NULL, + "MPLS inner RSS needs to specify inner L2/L3 items after MPLS in pattern"); return 0; } @@ -1729,6 +1798,37 @@ mlx5_flow_validate_action_count(struct rte_eth_dev *dev __rte_unused, return 0; } +/* + * Validate the ASO CT action. + * + * @param[in] dev + * Pointer to the Ethernet device structure. + * @param[in] conntrack + * Pointer to the CT action profile. + * @param[out] error + * Pointer to error structure. + * + * @return + * 0 on success, a negative errno value otherwise and rte_errno is set. + */ +int +mlx5_validate_action_ct(struct rte_eth_dev *dev, + const struct rte_flow_action_conntrack *conntrack, + struct rte_flow_error *error) +{ + RTE_SET_USED(dev); + + if (conntrack->state > RTE_FLOW_CONNTRACK_STATE_TIME_WAIT) + return rte_flow_error_set(error, EINVAL, + RTE_FLOW_ERROR_TYPE_ACTION, NULL, + "Invalid CT state"); + if (conntrack->last_index > RTE_FLOW_CONNTRACK_FLAG_RST) + return rte_flow_error_set(error, EINVAL, + RTE_FLOW_ERROR_TYPE_ACTION, NULL, + "Invalid last TCP packet flag"); + return 0; +} + /** * Verify the @p attributes will be correctly understood by the NIC and store * them in the @p flow if everything is correct. @@ -1755,7 +1855,7 @@ mlx5_flow_validate_attributes(struct rte_eth_dev *dev, return rte_flow_error_set(error, ENOTSUP, RTE_FLOW_ERROR_TYPE_ATTR_GROUP, NULL, "groups is not supported"); - if (attributes->priority != MLX5_FLOW_PRIO_RSVD && + if (attributes->priority != MLX5_FLOW_LOWEST_PRIO_INDICATOR && attributes->priority >= priority_max) return rte_flow_error_set(error, ENOTSUP, RTE_FLOW_ERROR_TYPE_ATTR_PRIORITY, @@ -1935,6 +2035,10 @@ mlx5_flow_validate_item_eth(const struct rte_flow_item *item, return rte_flow_error_set(error, EINVAL, RTE_FLOW_ERROR_TYPE_ITEM, item, "L2 layer should not follow VLAN"); + if (item_flags & MLX5_FLOW_LAYER_GTP) + return rte_flow_error_set(error, EINVAL, + RTE_FLOW_ERROR_TYPE_ITEM, item, + "L2 layer should not follow GTP"); if (!mask) mask = &rte_flow_item_eth_mask; ret = mlx5_flow_item_acceptable(item, (const uint8_t *)mask, @@ -2090,7 +2194,7 @@ mlx5_flow_validate_item_ipv4(const struct rte_flow_item *item, RTE_FLOW_ERROR_TYPE_ITEM, item, "IPv4 cannot follow L2/VLAN layer " "which ether type is not IPv4"); - if (item_flags & MLX5_FLOW_LAYER_IPIP) { + if (item_flags & MLX5_FLOW_LAYER_TUNNEL) { if (mask && spec) next_proto = mask->hdr.next_proto_id & spec->hdr.next_proto_id; @@ -2198,7 +2302,7 @@ mlx5_flow_validate_item_ipv6(const struct rte_flow_item *item, "which ether type is not IPv6"); if (mask && mask->hdr.proto == UINT8_MAX && spec) next_proto = spec->hdr.proto; - if (item_flags & MLX5_FLOW_LAYER_IPV6_ENCAP) { + if (item_flags & MLX5_FLOW_LAYER_TUNNEL) { if (next_proto == IPPROTO_IPIP || next_proto == IPPROTO_IPV6) return rte_flow_error_set(error, EINVAL, RTE_FLOW_ERROR_TYPE_ITEM, @@ -2361,12 +2465,16 @@ mlx5_flow_validate_item_tcp(const struct rte_flow_item *item, /** * Validate VXLAN item. * + * @param[in] dev + * Pointer to the Ethernet device structure. + * @param[in] udp_dport + * UDP destination port * @param[in] item * Item specification. * @param[in] item_flags * Bit-fields that holds the items detected until now. - * @param[in] target_protocol - * The next protocol in the previous item. + * @param[in] attr + * Flow rule attributes. * @param[out] error * Pointer to error structure. * @@ -2374,24 +2482,33 @@ mlx5_flow_validate_item_tcp(const struct rte_flow_item *item, * 0 on success, a negative errno value otherwise and rte_errno is set. */ int -mlx5_flow_validate_item_vxlan(const struct rte_flow_item *item, +mlx5_flow_validate_item_vxlan(struct rte_eth_dev *dev, + uint16_t udp_dport, + const struct rte_flow_item *item, uint64_t item_flags, + const struct rte_flow_attr *attr, struct rte_flow_error *error) { const struct rte_flow_item_vxlan *spec = item->spec; const struct rte_flow_item_vxlan *mask = item->mask; int ret; + struct mlx5_priv *priv = dev->data->dev_private; union vni { uint32_t vlan_id; uint8_t vni[4]; } id = { .vlan_id = 0, }; - + const struct rte_flow_item_vxlan nic_mask = { + .vni = "\xff\xff\xff", + .rsvd1 = 0xff, + }; + const struct rte_flow_item_vxlan *valid_mask; if (item_flags & MLX5_FLOW_LAYER_TUNNEL) return rte_flow_error_set(error, ENOTSUP, RTE_FLOW_ERROR_TYPE_ITEM, item, "multiple tunnel layers not" " supported"); + valid_mask = &rte_flow_item_vxlan_mask; /* * Verify only UDPv4 is present as defined in * https://tools.ietf.org/html/rfc7348 @@ -2402,9 +2519,21 @@ mlx5_flow_validate_item_vxlan(const struct rte_flow_item *item, "no outer UDP layer found"); if (!mask) mask = &rte_flow_item_vxlan_mask; + + if (priv->sh->steering_format_version != + MLX5_STEERING_LOGIC_FORMAT_CONNECTX_5 || + !udp_dport || udp_dport == MLX5_UDP_PORT_VXLAN) { + /* FDB domain & NIC domain non-zero group */ + if ((attr->transfer || attr->group) && priv->sh->misc5_cap) + valid_mask = &nic_mask; + /* Group zero in NIC domain */ + if (!attr->group && !attr->transfer && + priv->sh->tunnel_header_0_1) + valid_mask = &nic_mask; + } ret = mlx5_flow_item_acceptable (item, (const uint8_t *)mask, - (const uint8_t *)&rte_flow_item_vxlan_mask, + (const uint8_t *)valid_mask, sizeof(struct rte_flow_item_vxlan), MLX5_ITEM_RANGE_NOT_ACCEPTED, error); if (ret < 0) @@ -2708,6 +2837,149 @@ mlx5_flow_validate_item_geneve(const struct rte_flow_item *item, return 0; } +/** + * Validate Geneve TLV option item. + * + * @param[in] item + * Item specification. + * @param[in] last_item + * Previous validated item in the pattern items. + * @param[in] geneve_item + * Previous GENEVE item specification. + * @param[in] dev + * Pointer to the rte_eth_dev structure. + * @param[out] error + * Pointer to error structure. + * + * @return + * 0 on success, a negative errno value otherwise and rte_errno is set. + */ +int +mlx5_flow_validate_item_geneve_opt(const struct rte_flow_item *item, + uint64_t last_item, + const struct rte_flow_item *geneve_item, + struct rte_eth_dev *dev, + struct rte_flow_error *error) +{ + struct mlx5_priv *priv = dev->data->dev_private; + struct mlx5_dev_ctx_shared *sh = priv->sh; + struct mlx5_geneve_tlv_option_resource *geneve_opt_resource; + struct mlx5_hca_attr *hca_attr = &priv->config.hca_attr; + uint8_t data_max_supported = + hca_attr->max_geneve_tlv_option_data_len * 4; + struct mlx5_dev_config *config = &priv->config; + const struct rte_flow_item_geneve *geneve_spec; + const struct rte_flow_item_geneve *geneve_mask; + const struct rte_flow_item_geneve_opt *spec = item->spec; + const struct rte_flow_item_geneve_opt *mask = item->mask; + unsigned int i; + unsigned int data_len; + uint8_t tlv_option_len; + uint16_t optlen_m, optlen_v; + const struct rte_flow_item_geneve_opt full_mask = { + .option_class = RTE_BE16(0xffff), + .option_type = 0xff, + .option_len = 0x1f, + }; + + if (!mask) + mask = &rte_flow_item_geneve_opt_mask; + if (!spec) + return rte_flow_error_set + (error, ENOTSUP, RTE_FLOW_ERROR_TYPE_ITEM, item, + "Geneve TLV opt class/type/length must be specified"); + if ((uint32_t)spec->option_len > MLX5_GENEVE_OPTLEN_MASK) + return rte_flow_error_set + (error, ENOTSUP, RTE_FLOW_ERROR_TYPE_ITEM, item, + "Geneve TLV opt length exceeeds the limit (31)"); + /* Check if class type and length masks are full. */ + if (full_mask.option_class != mask->option_class || + full_mask.option_type != mask->option_type || + full_mask.option_len != (mask->option_len & full_mask.option_len)) + return rte_flow_error_set + (error, ENOTSUP, RTE_FLOW_ERROR_TYPE_ITEM, item, + "Geneve TLV opt class/type/length masks must be full"); + /* Check if length is supported */ + if ((uint32_t)spec->option_len > + config->hca_attr.max_geneve_tlv_option_data_len) + return rte_flow_error_set + (error, ENOTSUP, RTE_FLOW_ERROR_TYPE_ITEM, item, + "Geneve TLV opt length not supported"); + if (config->hca_attr.max_geneve_tlv_options > 1) + DRV_LOG(DEBUG, + "max_geneve_tlv_options supports more than 1 option"); + /* Check GENEVE item preceding. */ + if (!geneve_item || !(last_item & MLX5_FLOW_LAYER_GENEVE)) + return rte_flow_error_set + (error, ENOTSUP, RTE_FLOW_ERROR_TYPE_ITEM, item, + "Geneve opt item must be preceded with Geneve item"); + geneve_spec = geneve_item->spec; + geneve_mask = geneve_item->mask ? geneve_item->mask : + &rte_flow_item_geneve_mask; + /* Check if GENEVE TLV option size doesn't exceed option length */ + if (geneve_spec && (geneve_mask->ver_opt_len_o_c_rsvd0 || + geneve_spec->ver_opt_len_o_c_rsvd0)) { + tlv_option_len = spec->option_len & mask->option_len; + optlen_v = rte_be_to_cpu_16(geneve_spec->ver_opt_len_o_c_rsvd0); + optlen_v = MLX5_GENEVE_OPTLEN_VAL(optlen_v); + optlen_m = rte_be_to_cpu_16(geneve_mask->ver_opt_len_o_c_rsvd0); + optlen_m = MLX5_GENEVE_OPTLEN_VAL(optlen_m); + if ((optlen_v & optlen_m) <= tlv_option_len) + return rte_flow_error_set + (error, ENOTSUP, RTE_FLOW_ERROR_TYPE_ITEM, item, + "GENEVE TLV option length exceeds optlen"); + } + /* Check if length is 0 or data is 0. */ + if (spec->data == NULL || spec->option_len == 0) + return rte_flow_error_set + (error, ENOTSUP, RTE_FLOW_ERROR_TYPE_ITEM, item, + "Geneve TLV opt with zero data/length not supported"); + /* Check not all data & mask are 0. */ + data_len = spec->option_len * 4; + if (mask->data == NULL) { + for (i = 0; i < data_len; i++) + if (spec->data[i]) + break; + if (i == data_len) + return rte_flow_error_set(error, ENOTSUP, + RTE_FLOW_ERROR_TYPE_ITEM, item, + "Can't match on Geneve option data 0"); + } else { + for (i = 0; i < data_len; i++) + if (spec->data[i] & mask->data[i]) + break; + if (i == data_len) + return rte_flow_error_set(error, ENOTSUP, + RTE_FLOW_ERROR_TYPE_ITEM, item, + "Can't match on Geneve option data and mask 0"); + /* Check data mask supported. */ + for (i = data_max_supported; i < data_len ; i++) + if (mask->data[i]) + return rte_flow_error_set(error, ENOTSUP, + RTE_FLOW_ERROR_TYPE_ITEM, item, + "Data mask is of unsupported size"); + } + /* Check GENEVE option is supported in NIC. */ + if (!config->hca_attr.geneve_tlv_opt) + return rte_flow_error_set + (error, ENOTSUP, RTE_FLOW_ERROR_TYPE_ITEM, item, + "Geneve TLV opt not supported"); + /* Check if we already have geneve option with different type/class. */ + rte_spinlock_lock(&sh->geneve_tlv_opt_sl); + geneve_opt_resource = sh->geneve_tlv_option_resource; + if (geneve_opt_resource != NULL) + if (geneve_opt_resource->option_class != spec->option_class || + geneve_opt_resource->option_type != spec->option_type || + geneve_opt_resource->length != spec->option_len) { + rte_spinlock_unlock(&sh->geneve_tlv_opt_sl); + return rte_flow_error_set(error, ENOTSUP, + RTE_FLOW_ERROR_TYPE_ITEM, item, + "Only one Geneve TLV option supported"); + } + rte_spinlock_unlock(&sh->geneve_tlv_opt_sl); + return 0; +} + /** * Validate MPLS item. * @@ -2743,10 +3015,10 @@ mlx5_flow_validate_item_mpls(struct rte_eth_dev *dev __rte_unused, "MPLS not supported or" " disabled in firmware" " configuration."); - /* MPLS over IP, UDP, GRE is allowed */ - if (!(prev_layer & (MLX5_FLOW_LAYER_OUTER_L3 | - MLX5_FLOW_LAYER_OUTER_L4_UDP | - MLX5_FLOW_LAYER_GRE))) + /* MPLS over UDP, GRE is allowed */ + if (!(prev_layer & (MLX5_FLOW_LAYER_OUTER_L4_UDP | + MLX5_FLOW_LAYER_GRE | + MLX5_FLOW_LAYER_GRE_KEY))) return rte_flow_error_set(error, EINVAL, RTE_FLOW_ERROR_TYPE_ITEM, item, "protocol filtering not compatible" @@ -2917,30 +3189,6 @@ mlx5_flow_validate_item_ecpri(const struct rte_flow_item *item, MLX5_ITEM_RANGE_NOT_ACCEPTED, error); } -/** - * Release resource related QUEUE/RSS action split. - * - * @param dev - * Pointer to Ethernet device. - * @param flow - * Flow to release id's from. - */ -static void -flow_mreg_split_qrss_release(struct rte_eth_dev *dev, - struct rte_flow *flow) -{ - struct mlx5_priv *priv = dev->data->dev_private; - uint32_t handle_idx; - struct mlx5_flow_handle *dev_handle; - - SILIST_FOREACH(priv->sh->ipool[MLX5_IPOOL_MLX5_FLOW], flow->dev_handles, - handle_idx, dev_handle, next) - if (dev_handle->split_flow_id) - mlx5_ipool_free(priv->sh->ipool - [MLX5_IPOOL_RSS_EXPANTION_FLOW_ID], - dev_handle->split_flow_id); -} - static int flow_null_validate(struct rte_eth_dev *dev __rte_unused, const struct rte_flow_attr *attr __rte_unused, @@ -3236,34 +3484,139 @@ flow_drv_destroy(struct rte_eth_dev *dev, struct rte_flow *flow) const struct mlx5_flow_driver_ops *fops; enum mlx5_flow_drv_type type = flow->drv_type; - flow_mreg_split_qrss_release(dev, flow); MLX5_ASSERT(type > MLX5_FLOW_TYPE_MIN && type < MLX5_FLOW_TYPE_MAX); fops = flow_get_drv_ops(type); fops->destroy(dev, flow); } +/** + * Flow driver find RSS policy tbl API. This abstracts calling driver + * specific functions. Parent flow (rte_flow) should have driver + * type (drv_type). It will find the RSS policy table that has the rss_desc. + * + * @param[in] dev + * Pointer to Ethernet device. + * @param[in, out] flow + * Pointer to flow structure. + * @param[in] policy + * Pointer to meter policy table. + * @param[in] rss_desc + * Pointer to rss_desc + */ +static struct mlx5_flow_meter_sub_policy * +flow_drv_meter_sub_policy_rss_prepare(struct rte_eth_dev *dev, + struct rte_flow *flow, + struct mlx5_flow_meter_policy *policy, + struct mlx5_flow_rss_desc *rss_desc[MLX5_MTR_RTE_COLORS]) +{ + const struct mlx5_flow_driver_ops *fops; + enum mlx5_flow_drv_type type = flow->drv_type; + + MLX5_ASSERT(type > MLX5_FLOW_TYPE_MIN && type < MLX5_FLOW_TYPE_MAX); + fops = flow_get_drv_ops(type); + return fops->meter_sub_policy_rss_prepare(dev, policy, rss_desc); +} + +/** + * Flow driver color tag rule API. This abstracts calling driver + * specific functions. Parent flow (rte_flow) should have driver + * type (drv_type). It will create the color tag rules in hierarchy meter. + * + * @param[in] dev + * Pointer to Ethernet device. + * @param[in, out] flow + * Pointer to flow structure. + * @param[in] fm + * Pointer to flow meter structure. + * @param[in] src_port + * The src port this extra rule should use. + * @param[in] item + * The src port id match item. + * @param[out] error + * Pointer to error structure. + */ +static int +flow_drv_mtr_hierarchy_rule_create(struct rte_eth_dev *dev, + struct rte_flow *flow, + struct mlx5_flow_meter_info *fm, + int32_t src_port, + const struct rte_flow_item *item, + struct rte_flow_error *error) +{ + const struct mlx5_flow_driver_ops *fops; + enum mlx5_flow_drv_type type = flow->drv_type; + + MLX5_ASSERT(type > MLX5_FLOW_TYPE_MIN && type < MLX5_FLOW_TYPE_MAX); + fops = flow_get_drv_ops(type); + return fops->meter_hierarchy_rule_create(dev, fm, + src_port, item, error); +} + /** * Get RSS action from the action list. * + * @param[in] dev + * Pointer to Ethernet device. * @param[in] actions * Pointer to the list of actions. + * @param[in] flow + * Parent flow structure pointer. * * @return * Pointer to the RSS action if exist, else return NULL. */ static const struct rte_flow_action_rss* -flow_get_rss_action(const struct rte_flow_action actions[]) +flow_get_rss_action(struct rte_eth_dev *dev, + const struct rte_flow_action actions[]) { + struct mlx5_priv *priv = dev->data->dev_private; + const struct rte_flow_action_rss *rss = NULL; + for (; actions->type != RTE_FLOW_ACTION_TYPE_END; actions++) { switch (actions->type) { case RTE_FLOW_ACTION_TYPE_RSS: - return (const struct rte_flow_action_rss *) - actions->conf; + rss = actions->conf; + break; + case RTE_FLOW_ACTION_TYPE_SAMPLE: + { + const struct rte_flow_action_sample *sample = + actions->conf; + const struct rte_flow_action *act = sample->actions; + for (; act->type != RTE_FLOW_ACTION_TYPE_END; act++) + if (act->type == RTE_FLOW_ACTION_TYPE_RSS) + rss = act->conf; + break; + } + case RTE_FLOW_ACTION_TYPE_METER: + { + uint32_t mtr_idx; + struct mlx5_flow_meter_info *fm; + struct mlx5_flow_meter_policy *policy; + const struct rte_flow_action_meter *mtr = actions->conf; + + fm = mlx5_flow_meter_find(priv, mtr->mtr_id, &mtr_idx); + if (fm && !fm->def_policy) { + policy = mlx5_flow_meter_policy_find(dev, + fm->policy_id, NULL); + MLX5_ASSERT(policy); + if (policy->is_hierarchy) { + policy = + mlx5_flow_meter_hierarchy_get_final_policy(dev, + policy); + if (!policy) + return NULL; + } + if (policy->is_rss) + rss = + policy->act_cnt[RTE_COLOR_GREEN].rss->conf; + } + break; + } default: break; } } - return NULL; + return rss; } /** @@ -3289,31 +3642,31 @@ flow_aso_age_get_by_idx(struct rte_eth_dev *dev, uint32_t age_idx) return &pool->actions[offset - 1]; } -/* maps shared action to translated non shared in some actions array */ -struct mlx5_translated_shared_action { - struct rte_flow_shared_action *action; /**< Shared action */ - int index; /**< Index in related array of rte_flow_action */ +/* maps indirect action to translated direct in some actions array */ +struct mlx5_translated_action_handle { + struct rte_flow_action_handle *action; /**< Indirect action handle. */ + int index; /**< Index in related array of rte_flow_action. */ }; /** - * Translates actions of type RTE_FLOW_ACTION_TYPE_SHARED to related - * non shared action if translation possible. - * This functionality used to run same execution path for both shared & non - * shared actions on flow create. All necessary preparations for shared - * action handling should be preformed on *shared* actions list returned + * Translates actions of type RTE_FLOW_ACTION_TYPE_INDIRECT to related + * direct action if translation possible. + * This functionality used to run same execution path for both direct and + * indirect actions on flow create. All necessary preparations for indirect + * action handling should be performed on *handle* actions list returned * from this call. * * @param[in] dev * Pointer to Ethernet device. * @param[in] actions * List of actions to translate. - * @param[out] shared - * List to store translated shared actions. - * @param[in, out] shared_n - * Size of *shared* array. On return should be updated with number of shared - * actions retrieved from the *actions* list. + * @param[out] handle + * List to store translated indirect action object handles. + * @param[in, out] indir_n + * Size of *handle* array. On return should be updated with number of + * indirect actions retrieved from the *actions* list. * @param[out] translated_actions - * List of actions where all shared actions were translated to non shared + * List of actions where all indirect actions were translated to direct * if possible. NULL if no translation took place. * @param[out] error * Pointer to the error structure. @@ -3322,10 +3675,10 @@ struct mlx5_translated_shared_action { * 0 on success, a negative errno value otherwise and rte_errno is set. */ static int -flow_shared_actions_translate(struct rte_eth_dev *dev, +flow_action_handles_translate(struct rte_eth_dev *dev, const struct rte_flow_action actions[], - struct mlx5_translated_shared_action *shared, - int *shared_n, + struct mlx5_translated_action_handle *handle, + int *indir_n, struct rte_flow_action **translated_actions, struct rte_flow_error *error) { @@ -3334,23 +3687,23 @@ flow_shared_actions_translate(struct rte_eth_dev *dev, size_t actions_size; int n; int copied_n = 0; - struct mlx5_translated_shared_action *shared_end = NULL; + struct mlx5_translated_action_handle *handle_end = NULL; for (n = 0; actions[n].type != RTE_FLOW_ACTION_TYPE_END; n++) { - if (actions[n].type != RTE_FLOW_ACTION_TYPE_SHARED) + if (actions[n].type != RTE_FLOW_ACTION_TYPE_INDIRECT) continue; - if (copied_n == *shared_n) { + if (copied_n == *indir_n) { return rte_flow_error_set (error, EINVAL, RTE_FLOW_ERROR_TYPE_ACTION_NUM, NULL, "too many shared actions"); } - rte_memcpy(&shared[copied_n].action, &actions[n].conf, + rte_memcpy(&handle[copied_n].action, &actions[n].conf, sizeof(actions[n].conf)); - shared[copied_n].index = n; + handle[copied_n].index = n; copied_n++; } n++; - *shared_n = copied_n; + *indir_n = copied_n; if (!copied_n) return 0; actions_size = sizeof(struct rte_flow_action) * n; @@ -3360,28 +3713,43 @@ flow_shared_actions_translate(struct rte_eth_dev *dev, return -ENOMEM; } memcpy(translated, actions, actions_size); - for (shared_end = shared + copied_n; shared < shared_end; shared++) { + for (handle_end = handle + copied_n; handle < handle_end; handle++) { struct mlx5_shared_action_rss *shared_rss; - uint32_t act_idx = (uint32_t)(uintptr_t)shared->action; - uint32_t type = act_idx >> MLX5_SHARED_ACTION_TYPE_OFFSET; - uint32_t idx = act_idx & ((1u << MLX5_SHARED_ACTION_TYPE_OFFSET) - - 1); + uint32_t act_idx = (uint32_t)(uintptr_t)handle->action; + uint32_t type = act_idx >> MLX5_INDIRECT_ACTION_TYPE_OFFSET; + uint32_t idx = act_idx & + ((1u << MLX5_INDIRECT_ACTION_TYPE_OFFSET) - 1); switch (type) { - case MLX5_SHARED_ACTION_TYPE_RSS: + case MLX5_INDIRECT_ACTION_TYPE_RSS: shared_rss = mlx5_ipool_get (priv->sh->ipool[MLX5_IPOOL_RSS_SHARED_ACTIONS], idx); - translated[shared->index].type = + translated[handle->index].type = RTE_FLOW_ACTION_TYPE_RSS; - translated[shared->index].conf = + translated[handle->index].conf = &shared_rss->origin; break; - case MLX5_SHARED_ACTION_TYPE_AGE: + case MLX5_INDIRECT_ACTION_TYPE_COUNT: + translated[handle->index].type = + (enum rte_flow_action_type) + MLX5_RTE_FLOW_ACTION_TYPE_COUNT; + translated[handle->index].conf = (void *)(uintptr_t)idx; + break; + case MLX5_INDIRECT_ACTION_TYPE_AGE: if (priv->sh->flow_hit_aso_en) { - translated[shared->index].type = + translated[handle->index].type = (enum rte_flow_action_type) MLX5_RTE_FLOW_ACTION_TYPE_AGE; - translated[shared->index].conf = + translated[handle->index].conf = + (void *)(uintptr_t)idx; + break; + } + /* Fall-through */ + case MLX5_INDIRECT_ACTION_TYPE_CT: + if (priv->sh->ct_aso_en) { + translated[handle->index].type = + RTE_FLOW_ACTION_TYPE_CONNTRACK; + translated[handle->index].conf = (void *)(uintptr_t)idx; break; } @@ -3390,7 +3758,7 @@ flow_shared_actions_translate(struct rte_eth_dev *dev, mlx5_free(translated); return rte_flow_error_set (error, EINVAL, RTE_FLOW_ERROR_TYPE_ACTION, - NULL, "invalid shared action type"); + NULL, "invalid indirect action type"); } } *translated_actions = translated; @@ -3412,21 +3780,21 @@ flow_shared_actions_translate(struct rte_eth_dev *dev, */ static uint32_t flow_get_shared_rss_action(struct rte_eth_dev *dev, - struct mlx5_translated_shared_action *shared, + struct mlx5_translated_action_handle *handle, int shared_n) { - struct mlx5_translated_shared_action *shared_end; + struct mlx5_translated_action_handle *handle_end; struct mlx5_priv *priv = dev->data->dev_private; struct mlx5_shared_action_rss *shared_rss; - for (shared_end = shared + shared_n; shared < shared_end; shared++) { - uint32_t act_idx = (uint32_t)(uintptr_t)shared->action; - uint32_t type = act_idx >> MLX5_SHARED_ACTION_TYPE_OFFSET; + for (handle_end = handle + shared_n; handle < handle_end; handle++) { + uint32_t act_idx = (uint32_t)(uintptr_t)handle->action; + uint32_t type = act_idx >> MLX5_INDIRECT_ACTION_TYPE_OFFSET; uint32_t idx = act_idx & - ((1u << MLX5_SHARED_ACTION_TYPE_OFFSET) - 1); + ((1u << MLX5_INDIRECT_ACTION_TYPE_OFFSET) - 1); switch (type) { - case MLX5_SHARED_ACTION_TYPE_RSS: + case MLX5_INDIRECT_ACTION_TYPE_RSS: shared_rss = mlx5_ipool_get (priv->sh->ipool[MLX5_IPOOL_RSS_SHARED_ACTIONS], idx); @@ -3441,20 +3809,8 @@ flow_get_shared_rss_action(struct rte_eth_dev *dev, } static unsigned int -find_graph_root(const struct rte_flow_item pattern[], uint32_t rss_level) +find_graph_root(uint32_t rss_level) { - const struct rte_flow_item *item; - unsigned int has_vlan = 0; - - for (item = pattern; item->type != RTE_FLOW_ITEM_TYPE_END; item++) { - if (item->type == RTE_FLOW_ITEM_TYPE_VLAN) { - has_vlan = 1; - break; - } - } - if (has_vlan) - return rss_level < 2 ? MLX5_EXPANSION_ROOT_ETH_VLAN : - MLX5_EXPANSION_ROOT_OUTER_ETH_VLAN; return rss_level < 2 ? MLX5_EXPANSION_ROOT : MLX5_EXPANSION_ROOT_OUTER; } @@ -3557,32 +3913,104 @@ flow_parse_metadata_split_actions_info(const struct rte_flow_action actions[], return actions_n + 1; } +/** + * Check if the action will change packet. + * + * @param dev + * Pointer to Ethernet device. + * @param[in] type + * action type. + * + * @return + * true if action will change packet, false otherwise. + */ +static bool flow_check_modify_action_type(struct rte_eth_dev *dev, + enum rte_flow_action_type type) +{ + struct mlx5_priv *priv = dev->data->dev_private; + + switch (type) { + case RTE_FLOW_ACTION_TYPE_SET_MAC_SRC: + case RTE_FLOW_ACTION_TYPE_SET_MAC_DST: + case RTE_FLOW_ACTION_TYPE_SET_IPV4_SRC: + case RTE_FLOW_ACTION_TYPE_SET_IPV4_DST: + case RTE_FLOW_ACTION_TYPE_SET_IPV6_SRC: + case RTE_FLOW_ACTION_TYPE_SET_IPV6_DST: + case RTE_FLOW_ACTION_TYPE_SET_TP_SRC: + case RTE_FLOW_ACTION_TYPE_SET_TP_DST: + case RTE_FLOW_ACTION_TYPE_DEC_TTL: + case RTE_FLOW_ACTION_TYPE_SET_TTL: + case RTE_FLOW_ACTION_TYPE_INC_TCP_SEQ: + case RTE_FLOW_ACTION_TYPE_DEC_TCP_SEQ: + case RTE_FLOW_ACTION_TYPE_INC_TCP_ACK: + case RTE_FLOW_ACTION_TYPE_DEC_TCP_ACK: + case RTE_FLOW_ACTION_TYPE_SET_IPV4_DSCP: + case RTE_FLOW_ACTION_TYPE_SET_IPV6_DSCP: + case RTE_FLOW_ACTION_TYPE_SET_META: + case RTE_FLOW_ACTION_TYPE_SET_TAG: + case RTE_FLOW_ACTION_TYPE_OF_POP_VLAN: + case RTE_FLOW_ACTION_TYPE_OF_PUSH_VLAN: + case RTE_FLOW_ACTION_TYPE_OF_SET_VLAN_VID: + case RTE_FLOW_ACTION_TYPE_OF_SET_VLAN_PCP: + case RTE_FLOW_ACTION_TYPE_VXLAN_ENCAP: + case RTE_FLOW_ACTION_TYPE_VXLAN_DECAP: + case RTE_FLOW_ACTION_TYPE_NVGRE_ENCAP: + case RTE_FLOW_ACTION_TYPE_NVGRE_DECAP: + case RTE_FLOW_ACTION_TYPE_RAW_ENCAP: + case RTE_FLOW_ACTION_TYPE_RAW_DECAP: + case RTE_FLOW_ACTION_TYPE_MODIFY_FIELD: + return true; + case RTE_FLOW_ACTION_TYPE_FLAG: + case RTE_FLOW_ACTION_TYPE_MARK: + if (priv->config.dv_xmeta_en != MLX5_XMETA_MODE_LEGACY) + return true; + else + return false; + default: + return false; + } +} + /** * Check meter action from the action list. * + * @param dev + * Pointer to Ethernet device. * @param[in] actions * Pointer to the list of actions. - * @param[out] mtr + * @param[out] has_mtr * Pointer to the meter exist flag. + * @param[out] has_modify + * Pointer to the flag showing there's packet change action. + * @param[out] meter_id + * Pointer to the meter id. * * @return * Total number of actions. */ static int -flow_check_meter_action(const struct rte_flow_action actions[], uint32_t *mtr) +flow_check_meter_action(struct rte_eth_dev *dev, + const struct rte_flow_action actions[], + bool *has_mtr, bool *has_modify, uint32_t *meter_id) { + const struct rte_flow_action_meter *mtr = NULL; int actions_n = 0; - MLX5_ASSERT(mtr); - *mtr = 0; + MLX5_ASSERT(has_mtr); + *has_mtr = false; for (; actions->type != RTE_FLOW_ACTION_TYPE_END; actions++) { switch (actions->type) { case RTE_FLOW_ACTION_TYPE_METER: - *mtr = 1; + mtr = actions->conf; + *meter_id = mtr->mtr_id; + *has_mtr = true; break; default: break; } + if (!*has_mtr) + *has_modify |= flow_check_modify_action_type(dev, + actions->type); actions_n++; } /* Count RTE_FLOW_ACTION_TYPE_END. */ @@ -3628,7 +4056,7 @@ flow_check_hairpin_split(struct rte_eth_dev *dev, if (queue == NULL) return 0; conf = mlx5_rxq_get_hairpin_conf(dev, queue->index); - if (conf != NULL && !!conf->tx_explicit) + if (conf == NULL || conf->tx_explicit != 0) return 0; queue_action = 1; action_n++; @@ -3638,7 +4066,7 @@ flow_check_hairpin_split(struct rte_eth_dev *dev, if (rss == NULL || rss->queue_num == 0) return 0; conf = mlx5_rxq_get_hairpin_conf(dev, rss->queue[0]); - if (conf != NULL && !!conf->tx_explicit) + if (conf == NULL || conf->tx_explicit != 0) return 0; queue_action = 1; action_n++; @@ -3653,9 +4081,7 @@ flow_check_hairpin_split(struct rte_eth_dev *dev, break; case RTE_FLOW_ACTION_TYPE_RAW_ENCAP: raw_encap = actions->conf; - if (raw_encap->size > - (sizeof(struct rte_flow_item_eth) + - sizeof(struct rte_flow_item_ipv4))) + if (raw_encap->size > MLX5_ENCAPSULATION_DECISION_SIZE) split++; action_n++; break; @@ -3671,28 +4097,38 @@ flow_check_hairpin_split(struct rte_eth_dev *dev, /* Declare flow create/destroy prototype in advance. */ static uint32_t -flow_list_create(struct rte_eth_dev *dev, uint32_t *list, +flow_list_create(struct rte_eth_dev *dev, enum mlx5_flow_type type, const struct rte_flow_attr *attr, const struct rte_flow_item items[], const struct rte_flow_action actions[], bool external, struct rte_flow_error *error); static void -flow_list_destroy(struct rte_eth_dev *dev, uint32_t *list, +flow_list_destroy(struct rte_eth_dev *dev, enum mlx5_flow_type type, uint32_t flow_idx); -struct mlx5_hlist_entry * -flow_dv_mreg_create_cb(struct mlx5_hlist *list, uint64_t key, - void *cb_ctx) +int +flow_dv_mreg_match_cb(void *tool_ctx __rte_unused, + struct mlx5_list_entry *entry, void *cb_ctx) +{ + struct mlx5_flow_cb_ctx *ctx = cb_ctx; + struct mlx5_flow_mreg_copy_resource *mcp_res = + container_of(entry, typeof(*mcp_res), hlist_ent); + + return mcp_res->mark_id != *(uint32_t *)(ctx->data); +} + +struct mlx5_list_entry * +flow_dv_mreg_create_cb(void *tool_ctx, void *cb_ctx) { - struct rte_eth_dev *dev = list->ctx; + struct rte_eth_dev *dev = tool_ctx; struct mlx5_priv *priv = dev->data->dev_private; struct mlx5_flow_cb_ctx *ctx = cb_ctx; struct mlx5_flow_mreg_copy_resource *mcp_res; struct rte_flow_error *error = ctx->error; uint32_t idx = 0; int ret; - uint32_t mark_id = key; + uint32_t mark_id = *(uint32_t *)(ctx->data); struct rte_flow_attr attr = { .group = MLX5_FLOW_MREG_CP_TABLE_GROUP, .ingress = 1, @@ -3758,7 +4194,7 @@ flow_dv_mreg_create_cb(struct mlx5_hlist *list, uint64_t key, }; } else { /* Default rule, wildcard match. */ - attr.priority = MLX5_FLOW_PRIO_RSVD; + attr.priority = MLX5_FLOW_LOWEST_PRIO_INDICATOR; items[0] = (struct rte_flow_item){ .type = RTE_FLOW_ITEM_TYPE_END, }; @@ -3782,14 +4218,15 @@ flow_dv_mreg_create_cb(struct mlx5_hlist *list, uint64_t key, return NULL; } mcp_res->idx = idx; + mcp_res->mark_id = mark_id; /* * The copy Flows are not included in any list. There * ones are referenced from other Flows and can not * be applied, removed, deleted in ardbitrary order * by list traversing. */ - mcp_res->rix_flow = flow_list_create(dev, NULL, &attr, items, - actions, false, error); + mcp_res->rix_flow = flow_list_create(dev, MLX5_FLOW_TYPE_MCP, + &attr, items, actions, false, error); if (!mcp_res->rix_flow) { mlx5_ipool_free(priv->sh->ipool[MLX5_IPOOL_MCP], idx); return NULL; @@ -3797,14 +4234,44 @@ flow_dv_mreg_create_cb(struct mlx5_hlist *list, uint64_t key, return &mcp_res->hlist_ent; } -/** - * Add a flow of copying flow metadata registers in RX_CP_TBL. - * - * As mark_id is unique, if there's already a registered flow for the mark_id, - * return by increasing the reference counter of the resource. Otherwise, create - * the resource (mcp_res) and flow. - * - * Flow looks like, +struct mlx5_list_entry * +flow_dv_mreg_clone_cb(void *tool_ctx, struct mlx5_list_entry *oentry, + void *cb_ctx __rte_unused) +{ + struct rte_eth_dev *dev = tool_ctx; + struct mlx5_priv *priv = dev->data->dev_private; + struct mlx5_flow_mreg_copy_resource *mcp_res; + uint32_t idx = 0; + + mcp_res = mlx5_ipool_malloc(priv->sh->ipool[MLX5_IPOOL_MCP], &idx); + if (!mcp_res) { + rte_errno = ENOMEM; + return NULL; + } + memcpy(mcp_res, oentry, sizeof(*mcp_res)); + mcp_res->idx = idx; + return &mcp_res->hlist_ent; +} + +void +flow_dv_mreg_clone_free_cb(void *tool_ctx, struct mlx5_list_entry *entry) +{ + struct mlx5_flow_mreg_copy_resource *mcp_res = + container_of(entry, typeof(*mcp_res), hlist_ent); + struct rte_eth_dev *dev = tool_ctx; + struct mlx5_priv *priv = dev->data->dev_private; + + mlx5_ipool_free(priv->sh->ipool[MLX5_IPOOL_MCP], mcp_res->idx); +} + +/** + * Add a flow of copying flow metadata registers in RX_CP_TBL. + * + * As mark_id is unique, if there's already a registered flow for the mark_id, + * return by increasing the reference counter of the resource. Otherwise, create + * the resource (mcp_res) and flow. + * + * Flow looks like, * - If ingress port is ANY and reg_c[1] is mark_id, * flow_tag := mark_id, reg_b := reg_c[0] and jump to RX_ACT_TBL. * @@ -3827,10 +4294,11 @@ flow_mreg_add_copy_action(struct rte_eth_dev *dev, uint32_t mark_id, struct rte_flow_error *error) { struct mlx5_priv *priv = dev->data->dev_private; - struct mlx5_hlist_entry *entry; + struct mlx5_list_entry *entry; struct mlx5_flow_cb_ctx ctx = { .dev = dev, .error = error, + .data = &mark_id, }; /* Check if already registered. */ @@ -3843,15 +4311,15 @@ flow_mreg_add_copy_action(struct rte_eth_dev *dev, uint32_t mark_id, } void -flow_dv_mreg_remove_cb(struct mlx5_hlist *list, struct mlx5_hlist_entry *entry) +flow_dv_mreg_remove_cb(void *tool_ctx, struct mlx5_list_entry *entry) { struct mlx5_flow_mreg_copy_resource *mcp_res = - container_of(entry, typeof(*mcp_res), hlist_ent); - struct rte_eth_dev *dev = list->ctx; + container_of(entry, typeof(*mcp_res), hlist_ent); + struct rte_eth_dev *dev = tool_ctx; struct mlx5_priv *priv = dev->data->dev_private; MLX5_ASSERT(mcp_res->rix_flow); - flow_list_destroy(dev, NULL, mcp_res->rix_flow); + flow_list_destroy(dev, MLX5_FLOW_TYPE_MCP, mcp_res->rix_flow); mlx5_ipool_free(priv->sh->ipool[MLX5_IPOOL_MCP], mcp_res->idx); } @@ -3893,14 +4361,17 @@ flow_mreg_del_copy_action(struct rte_eth_dev *dev, static void flow_mreg_del_default_copy_action(struct rte_eth_dev *dev) { - struct mlx5_hlist_entry *entry; + struct mlx5_list_entry *entry; struct mlx5_priv *priv = dev->data->dev_private; + struct mlx5_flow_cb_ctx ctx; + uint32_t mark_id; /* Check if default flow is registered. */ if (!priv->mreg_cp_tbl) return; - entry = mlx5_hlist_lookup(priv->mreg_cp_tbl, - MLX5_DEFAULT_COPY_ID, NULL); + mark_id = MLX5_DEFAULT_COPY_ID; + ctx.data = &mark_id; + entry = mlx5_hlist_lookup(priv->mreg_cp_tbl, mark_id, &ctx); if (!entry) return; mlx5_hlist_unregister(priv->mreg_cp_tbl, entry); @@ -3926,6 +4397,8 @@ flow_mreg_add_default_copy_action(struct rte_eth_dev *dev, { struct mlx5_priv *priv = dev->data->dev_private; struct mlx5_flow_mreg_copy_resource *mcp_res; + struct mlx5_flow_cb_ctx ctx; + uint32_t mark_id; /* Check whether extensive metadata feature is engaged. */ if (!priv->config.dv_flow_en || @@ -3937,9 +4410,11 @@ flow_mreg_add_default_copy_action(struct rte_eth_dev *dev, * Add default mreg copy flow may be called multiple time, but * only be called once in stop. Avoid register it twice. */ - if (mlx5_hlist_lookup(priv->mreg_cp_tbl, MLX5_DEFAULT_COPY_ID, NULL)) + mark_id = MLX5_DEFAULT_COPY_ID; + ctx.data = &mark_id; + if (mlx5_hlist_lookup(priv->mreg_cp_tbl, mark_id, &ctx)) return 0; - mcp_res = flow_mreg_add_copy_action(dev, MLX5_DEFAULT_COPY_ID, error); + mcp_res = flow_mreg_add_copy_action(dev, mark_id, error); if (!mcp_res) return -rte_errno; return 0; @@ -4090,9 +4565,7 @@ flow_hairpin_split(struct rte_eth_dev *dev, break; case RTE_FLOW_ACTION_TYPE_RAW_ENCAP: raw_encap = actions->conf; - if (raw_encap->size > - (sizeof(struct rte_flow_item_eth) + - sizeof(struct rte_flow_item_ipv4))) { + if (raw_encap->size > MLX5_ENCAPSULATION_DECISION_SIZE) { memcpy(actions_tx, actions, sizeof(struct rte_flow_action)); actions_tx++; @@ -4105,9 +4578,7 @@ flow_hairpin_split(struct rte_eth_dev *dev, break; case RTE_FLOW_ACTION_TYPE_RAW_DECAP: raw_decap = actions->conf; - if (raw_decap->size < - (sizeof(struct rte_flow_item_eth) + - sizeof(struct rte_flow_item_ipv4))) { + if (raw_decap->size < MLX5_ENCAPSULATION_DECISION_SIZE) { memcpy(actions_tx, actions, sizeof(struct rte_flow_action)); actions_tx++; @@ -4132,9 +4603,11 @@ flow_hairpin_split(struct rte_eth_dev *dev, rte_memcpy(actions_rx, actions, sizeof(struct rte_flow_action)); actions_rx++; set_tag = (void *)actions_rx; - set_tag->id = mlx5_flow_get_reg_id(dev, MLX5_HAIRPIN_RX, 0, NULL); + *set_tag = (struct mlx5_rte_flow_action_set_tag) { + .id = mlx5_flow_get_reg_id(dev, MLX5_HAIRPIN_RX, 0, NULL), + .data = flow_id, + }; MLX5_ASSERT(set_tag->id > REG_NON); - set_tag->data = flow_id; tag_action->conf = set_tag; /* Create Tx item list. */ rte_memcpy(actions_tx, actions, sizeof(struct rte_flow_action)); @@ -4158,142 +4631,6 @@ flow_hairpin_split(struct rte_eth_dev *dev, return 0; } -__extension__ -union tunnel_offload_mark { - uint32_t val; - struct { - uint32_t app_reserve:8; - uint32_t table_id:15; - uint32_t transfer:1; - uint32_t _unused_:8; - }; -}; - -struct tunnel_default_miss_ctx { - uint16_t *queue; - __extension__ - union { - struct rte_flow_action_rss action_rss; - struct rte_flow_action_queue miss_queue; - struct rte_flow_action_jump miss_jump; - uint8_t raw[0]; - }; -}; - -static int -flow_tunnel_add_default_miss(struct rte_eth_dev *dev, - struct rte_flow *flow, - const struct rte_flow_attr *attr, - const struct rte_flow_action *app_actions, - uint32_t flow_idx, - struct tunnel_default_miss_ctx *ctx, - struct rte_flow_error *error) -{ - struct mlx5_priv *priv = dev->data->dev_private; - struct mlx5_flow *dev_flow; - struct rte_flow_attr miss_attr = *attr; - const struct mlx5_flow_tunnel *tunnel = app_actions[0].conf; - const struct rte_flow_item miss_items[2] = { - { - .type = RTE_FLOW_ITEM_TYPE_ETH, - .spec = NULL, - .last = NULL, - .mask = NULL - }, - { - .type = RTE_FLOW_ITEM_TYPE_END, - .spec = NULL, - .last = NULL, - .mask = NULL - } - }; - union tunnel_offload_mark mark_id; - struct rte_flow_action_mark miss_mark; - struct rte_flow_action miss_actions[3] = { - [0] = { .type = RTE_FLOW_ACTION_TYPE_MARK, .conf = &miss_mark }, - [2] = { .type = RTE_FLOW_ACTION_TYPE_END, .conf = NULL } - }; - const struct rte_flow_action_jump *jump_data; - uint32_t i, flow_table = 0; /* prevent compilation warning */ - struct flow_grp_info grp_info = { - .external = 1, - .transfer = attr->transfer, - .fdb_def_rule = !!priv->fdb_def_rule, - .std_tbl_fix = 0, - }; - int ret; - - if (!attr->transfer) { - uint32_t q_size; - - miss_actions[1].type = RTE_FLOW_ACTION_TYPE_RSS; - q_size = priv->reta_idx_n * sizeof(ctx->queue[0]); - ctx->queue = mlx5_malloc(MLX5_MEM_SYS | MLX5_MEM_ZERO, q_size, - 0, SOCKET_ID_ANY); - if (!ctx->queue) - return rte_flow_error_set - (error, ENOMEM, - RTE_FLOW_ERROR_TYPE_ACTION_CONF, - NULL, "invalid default miss RSS"); - ctx->action_rss.func = RTE_ETH_HASH_FUNCTION_DEFAULT, - ctx->action_rss.level = 0, - ctx->action_rss.types = priv->rss_conf.rss_hf, - ctx->action_rss.key_len = priv->rss_conf.rss_key_len, - ctx->action_rss.queue_num = priv->reta_idx_n, - ctx->action_rss.key = priv->rss_conf.rss_key, - ctx->action_rss.queue = ctx->queue; - if (!priv->reta_idx_n || !priv->rxqs_n) - return rte_flow_error_set - (error, EINVAL, - RTE_FLOW_ERROR_TYPE_ACTION_CONF, - NULL, "invalid port configuration"); - if (!(dev->data->dev_conf.rxmode.mq_mode & ETH_MQ_RX_RSS_FLAG)) - ctx->action_rss.types = 0; - for (i = 0; i != priv->reta_idx_n; ++i) - ctx->queue[i] = (*priv->reta_idx)[i]; - } else { - miss_actions[1].type = RTE_FLOW_ACTION_TYPE_JUMP; - ctx->miss_jump.group = MLX5_TNL_MISS_FDB_JUMP_GRP; - } - miss_actions[1].conf = (typeof(miss_actions[1].conf))ctx->raw; - for (; app_actions->type != RTE_FLOW_ACTION_TYPE_JUMP; app_actions++); - jump_data = app_actions->conf; - miss_attr.priority = MLX5_TNL_MISS_RULE_PRIORITY; - miss_attr.group = jump_data->group; - ret = mlx5_flow_group_to_table(dev, tunnel, jump_data->group, - &flow_table, grp_info, error); - if (ret) - return rte_flow_error_set(error, EINVAL, - RTE_FLOW_ERROR_TYPE_ACTION_CONF, - NULL, "invalid tunnel id"); - mark_id.app_reserve = 0; - mark_id.table_id = tunnel_flow_tbl_to_id(flow_table); - mark_id.transfer = !!attr->transfer; - mark_id._unused_ = 0; - miss_mark.id = mark_id.val; - dev_flow = flow_drv_prepare(dev, flow, &miss_attr, - miss_items, miss_actions, flow_idx, error); - if (!dev_flow) - return -rte_errno; - dev_flow->flow = flow; - dev_flow->external = true; - dev_flow->tunnel = tunnel; - /* Subflow object was created, we must include one in the list. */ - SILIST_INSERT(&flow->dev_handles, dev_flow->handle_idx, - dev_flow->handle, next); - DRV_LOG(DEBUG, - "port %u tunnel type=%d id=%u miss rule priority=%u group=%u", - dev->data->port_id, tunnel->app_tunnel.type, - tunnel->tunnel_id, miss_attr.priority, miss_attr.group); - ret = flow_drv_translate(dev, dev_flow, &miss_attr, miss_items, - miss_actions, error); - if (!ret) - ret = flow_mreg_update_copy_table(dev, flow, miss_actions, - error); - - return ret; -} - /** * The last stage of splitting chain, just creates the subflow * without any modification. @@ -4350,9 +4687,131 @@ flow_create_split_inner(struct rte_eth_dev *dev, dev_flow->handle->mark = 1; if (sub_flow) *sub_flow = dev_flow; +#ifdef HAVE_IBV_FLOW_DV_SUPPORT + dev_flow->dv.table_id = flow_split_info->table_id; +#endif return flow_drv_translate(dev, dev_flow, attr, items, actions, error); } +/** + * Get the sub policy of a meter. + * + * @param[in] dev + * Pointer to Ethernet device. + * @param[in] flow + * Parent flow structure pointer. + * @param wks + * Pointer to thread flow work space. + * @param[in] attr + * Flow rule attributes. + * @param[in] items + * Pattern specification (list terminated by the END pattern item). + * @param[out] error + * Perform verbose error reporting if not NULL. + * + * @return + * Pointer to the meter sub policy, NULL otherwise and rte_errno is set. + */ +static struct mlx5_flow_meter_sub_policy * +get_meter_sub_policy(struct rte_eth_dev *dev, + struct rte_flow *flow, + struct mlx5_flow_workspace *wks, + const struct rte_flow_attr *attr, + const struct rte_flow_item items[], + struct rte_flow_error *error) +{ + struct mlx5_flow_meter_policy *policy; + struct mlx5_flow_meter_policy *final_policy; + struct mlx5_flow_meter_sub_policy *sub_policy = NULL; + + policy = wks->policy; + final_policy = policy->is_hierarchy ? wks->final_policy : policy; + if (final_policy->is_rss || final_policy->is_queue) { + struct mlx5_flow_rss_desc rss_desc_v[MLX5_MTR_RTE_COLORS]; + struct mlx5_flow_rss_desc *rss_desc[MLX5_MTR_RTE_COLORS] = {0}; + uint32_t i; + + /* + * This is a tmp dev_flow, + * no need to register any matcher for it in translate. + */ + wks->skip_matcher_reg = 1; + for (i = 0; i < MLX5_MTR_RTE_COLORS; i++) { + struct mlx5_flow dev_flow = {0}; + struct mlx5_flow_handle dev_handle = { {0} }; + uint8_t fate = final_policy->act_cnt[i].fate_action; + + if (fate == MLX5_FLOW_FATE_SHARED_RSS) { + const struct rte_flow_action_rss *rss_act = + final_policy->act_cnt[i].rss->conf; + struct rte_flow_action rss_actions[2] = { + [0] = { + .type = RTE_FLOW_ACTION_TYPE_RSS, + .conf = rss_act, + }, + [1] = { + .type = RTE_FLOW_ACTION_TYPE_END, + .conf = NULL, + } + }; + + dev_flow.handle = &dev_handle; + dev_flow.ingress = attr->ingress; + dev_flow.flow = flow; + dev_flow.external = 0; +#ifdef HAVE_IBV_FLOW_DV_SUPPORT + dev_flow.dv.transfer = attr->transfer; +#endif + /** + * Translate RSS action to get rss hash fields. + */ + if (flow_drv_translate(dev, &dev_flow, attr, + items, rss_actions, error)) + goto exit; + rss_desc_v[i] = wks->rss_desc; + rss_desc_v[i].key_len = MLX5_RSS_HASH_KEY_LEN; + rss_desc_v[i].hash_fields = + dev_flow.hash_fields; + rss_desc_v[i].queue_num = + rss_desc_v[i].hash_fields ? + rss_desc_v[i].queue_num : 1; + rss_desc_v[i].tunnel = + !!(dev_flow.handle->layers & + MLX5_FLOW_LAYER_TUNNEL); + /* Use the RSS queues in the containers. */ + rss_desc_v[i].queue = + (uint16_t *)(uintptr_t)rss_act->queue; + rss_desc[i] = &rss_desc_v[i]; + } else if (fate == MLX5_FLOW_FATE_QUEUE) { + /* This is queue action. */ + rss_desc_v[i] = wks->rss_desc; + rss_desc_v[i].key_len = 0; + rss_desc_v[i].hash_fields = 0; + rss_desc_v[i].queue = + &final_policy->act_cnt[i].queue; + rss_desc_v[i].queue_num = 1; + rss_desc[i] = &rss_desc_v[i]; + } else { + rss_desc[i] = NULL; + } + } + sub_policy = flow_drv_meter_sub_policy_rss_prepare(dev, + flow, policy, rss_desc); + } else { + enum mlx5_meter_domain mtr_domain = + attr->transfer ? MLX5_MTR_DOMAIN_TRANSFER : + (attr->egress ? MLX5_MTR_DOMAIN_EGRESS : + MLX5_MTR_DOMAIN_INGRESS); + sub_policy = policy->sub_policys[mtr_domain][0]; + } + if (!sub_policy) + rte_flow_error_set(error, EINVAL, + RTE_FLOW_ERROR_TYPE_UNSPECIFIED, NULL, + "Failed to get meter sub-policy."); +exit: + return sub_policy; +} + /** * Split the meter flow. * @@ -4365,8 +4824,14 @@ flow_create_split_inner(struct rte_eth_dev *dev, * header will be in the prefix sub flow, as not to take the * L3 tunnel header into account. * - * @param dev + * @param[in] dev * Pointer to Ethernet device. + * @param[in] flow + * Parent flow structure pointer. + * @param wks + * Pointer to thread flow work space. + * @param[in] attr + * Flow rule attributes. * @param[in] items * Pattern specification (list terminated by the END pattern item). * @param[out] sfx_items @@ -4377,133 +4842,248 @@ flow_create_split_inner(struct rte_eth_dev *dev, * Suffix flow actions. * @param[out] actions_pre * Prefix flow actions. - * @param[out] pattern_sfx - * The pattern items for the suffix flow. - * @param[out] tag_sfx - * Pointer to suffix flow tag. + * @param[out] mtr_flow_id + * Pointer to meter flow id. + * @param[out] error + * Perform verbose error reporting if not NULL. * * @return - * 0 on success. + * 0 on success, a negative errno value otherwise and rte_errno is set. */ static int flow_meter_split_prep(struct rte_eth_dev *dev, - const struct rte_flow_item items[], - struct rte_flow_item sfx_items[], - const struct rte_flow_action actions[], - struct rte_flow_action actions_sfx[], - struct rte_flow_action actions_pre[]) + struct rte_flow *flow, + struct mlx5_flow_workspace *wks, + const struct rte_flow_attr *attr, + const struct rte_flow_item items[], + struct rte_flow_item sfx_items[], + const struct rte_flow_action actions[], + struct rte_flow_action actions_sfx[], + struct rte_flow_action actions_pre[], + uint32_t *mtr_flow_id, + struct rte_flow_error *error) { struct mlx5_priv *priv = dev->data->dev_private; + struct mlx5_flow_meter_info *fm = wks->fm; struct rte_flow_action *tag_action = NULL; struct rte_flow_item *tag_item; struct mlx5_rte_flow_action_set_tag *set_tag; - struct rte_flow_error error; const struct rte_flow_action_raw_encap *raw_encap; const struct rte_flow_action_raw_decap *raw_decap; - struct mlx5_rte_flow_item_tag *tag_spec; - struct mlx5_rte_flow_item_tag *tag_mask; + struct mlx5_rte_flow_item_tag *tag_item_spec; + struct mlx5_rte_flow_item_tag *tag_item_mask; uint32_t tag_id = 0; - bool copy_vlan = false; + struct rte_flow_item *vlan_item_dst = NULL; + const struct rte_flow_item *vlan_item_src = NULL; + struct rte_flow_action *hw_mtr_action; + struct rte_flow_action *action_pre_head = NULL; + int32_t flow_src_port = priv->representor_id; + bool mtr_first; + uint8_t mtr_id_offset = priv->mtr_reg_share ? MLX5_MTR_COLOR_BITS : 0; + uint8_t mtr_reg_bits = priv->mtr_reg_share ? + MLX5_MTR_IDLE_BITS_IN_COLOR_REG : MLX5_REG_BITS; + uint32_t flow_id = 0; + uint32_t flow_id_reversed = 0; + uint8_t flow_id_bits = 0; + int shift; + + /* Prepare the suffix subflow items. */ + tag_item = sfx_items++; + for (; items->type != RTE_FLOW_ITEM_TYPE_END; items++) { + struct mlx5_priv *port_priv; + const struct rte_flow_item_port_id *pid_v; + int item_type = items->type; + switch (item_type) { + case RTE_FLOW_ITEM_TYPE_PORT_ID: + pid_v = items->spec; + MLX5_ASSERT(pid_v); + port_priv = mlx5_port_to_eswitch_info(pid_v->id, false); + if (!port_priv) + return rte_flow_error_set(error, + rte_errno, + RTE_FLOW_ERROR_TYPE_ITEM_SPEC, + pid_v, + "Failed to get port info."); + flow_src_port = port_priv->representor_id; + if (!fm->def_policy && wks->policy->is_hierarchy && + flow_src_port != priv->representor_id) { + if (flow_drv_mtr_hierarchy_rule_create(dev, + flow, fm, + flow_src_port, + items, + error)) + return -rte_errno; + } + memcpy(sfx_items, items, sizeof(*sfx_items)); + sfx_items++; + break; + case RTE_FLOW_ITEM_TYPE_VLAN: + /* Determine if copy vlan item below. */ + vlan_item_src = items; + vlan_item_dst = sfx_items++; + vlan_item_dst->type = RTE_FLOW_ITEM_TYPE_VOID; + break; + default: + break; + } + } + sfx_items->type = RTE_FLOW_ITEM_TYPE_END; + sfx_items++; + mtr_first = priv->sh->meter_aso_en && + (attr->egress || (attr->transfer && flow_src_port != UINT16_MAX)); + /* For ASO meter, meter must be before tag in TX direction. */ + if (mtr_first) { + action_pre_head = actions_pre++; + /* Leave space for tag action. */ + tag_action = actions_pre++; + } /* Prepare the actions for prefix and suffix flow. */ for (; actions->type != RTE_FLOW_ACTION_TYPE_END; actions++) { - struct rte_flow_action **action_cur = NULL; + struct rte_flow_action *action_cur = NULL; switch (actions->type) { case RTE_FLOW_ACTION_TYPE_METER: - /* Add the extra tag action first. */ - tag_action = actions_pre; - tag_action->type = (enum rte_flow_action_type) - MLX5_RTE_FLOW_ACTION_TYPE_TAG; - actions_pre++; - action_cur = &actions_pre; + if (mtr_first) { + action_cur = action_pre_head; + } else { + /* Leave space for tag action. */ + tag_action = actions_pre++; + action_cur = actions_pre++; + } break; case RTE_FLOW_ACTION_TYPE_VXLAN_DECAP: case RTE_FLOW_ACTION_TYPE_NVGRE_DECAP: - action_cur = &actions_pre; + action_cur = actions_pre++; break; case RTE_FLOW_ACTION_TYPE_RAW_ENCAP: raw_encap = actions->conf; if (raw_encap->size < MLX5_ENCAPSULATION_DECISION_SIZE) - action_cur = &actions_pre; + action_cur = actions_pre++; break; case RTE_FLOW_ACTION_TYPE_RAW_DECAP: raw_decap = actions->conf; if (raw_decap->size > MLX5_ENCAPSULATION_DECISION_SIZE) - action_cur = &actions_pre; + action_cur = actions_pre++; break; case RTE_FLOW_ACTION_TYPE_OF_PUSH_VLAN: case RTE_FLOW_ACTION_TYPE_OF_SET_VLAN_VID: - copy_vlan = true; + if (vlan_item_dst && vlan_item_src) { + memcpy(vlan_item_dst, vlan_item_src, + sizeof(*vlan_item_dst)); + /* + * Convert to internal match item, it is used + * for vlan push and set vid. + */ + vlan_item_dst->type = (enum rte_flow_item_type) + MLX5_RTE_FLOW_ITEM_TYPE_VLAN; + } break; default: break; } if (!action_cur) - action_cur = &actions_sfx; - memcpy(*action_cur, actions, sizeof(struct rte_flow_action)); - (*action_cur)++; + action_cur = (fm->def_policy) ? + actions_sfx++ : actions_pre++; + memcpy(action_cur, actions, sizeof(struct rte_flow_action)); } /* Add end action to the actions. */ actions_sfx->type = RTE_FLOW_ACTION_TYPE_END; - actions_pre->type = RTE_FLOW_ACTION_TYPE_END; - actions_pre++; - /* Set the tag. */ - set_tag = (void *)actions_pre; - set_tag->id = mlx5_flow_get_reg_id(dev, MLX5_MTR_SFX, 0, &error); - mlx5_ipool_malloc(priv->sh->ipool[MLX5_IPOOL_RSS_EXPANTION_FLOW_ID], - &tag_id); - if (tag_id >= (1 << (sizeof(tag_id) * 8 - MLX5_MTR_COLOR_BITS))) { - DRV_LOG(ERR, "Port %u meter flow id exceed max limit.", - dev->data->port_id); - mlx5_ipool_free(priv->sh->ipool - [MLX5_IPOOL_RSS_EXPANTION_FLOW_ID], tag_id); - return 0; - } else if (!tag_id) { - return 0; - } - set_tag->data = tag_id << MLX5_MTR_COLOR_BITS; - assert(tag_action); - tag_action->conf = set_tag; - /* Prepare the suffix subflow items. */ - tag_item = sfx_items++; - for (; items->type != RTE_FLOW_ITEM_TYPE_END; items++) { - int item_type = items->type; + if (priv->sh->meter_aso_en) { + /** + * For ASO meter, need to add an extra jump action explicitly, + * to jump from meter to policer table. + */ + struct mlx5_flow_meter_sub_policy *sub_policy; + struct mlx5_flow_tbl_data_entry *tbl_data; - switch (item_type) { - case RTE_FLOW_ITEM_TYPE_PORT_ID: - memcpy(sfx_items, items, sizeof(*sfx_items)); - sfx_items++; - break; - case RTE_FLOW_ITEM_TYPE_VLAN: - if (copy_vlan) { - memcpy(sfx_items, items, sizeof(*sfx_items)); - /* - * Convert to internal match item, it is used - * for vlan push and set vid. - */ - sfx_items->type = (enum rte_flow_item_type) - MLX5_RTE_FLOW_ITEM_TYPE_VLAN; - sfx_items++; - } - break; - default: - break; + if (!fm->def_policy) { + sub_policy = get_meter_sub_policy(dev, flow, wks, + attr, items, error); + if (!sub_policy) + return -rte_errno; + } else { + enum mlx5_meter_domain mtr_domain = + attr->transfer ? MLX5_MTR_DOMAIN_TRANSFER : + (attr->egress ? MLX5_MTR_DOMAIN_EGRESS : + MLX5_MTR_DOMAIN_INGRESS); + + sub_policy = + &priv->sh->mtrmng->def_policy[mtr_domain]->sub_policy; } + tbl_data = container_of(sub_policy->tbl_rsc, + struct mlx5_flow_tbl_data_entry, tbl); + hw_mtr_action = actions_pre++; + hw_mtr_action->type = (enum rte_flow_action_type) + MLX5_RTE_FLOW_ACTION_TYPE_JUMP; + hw_mtr_action->conf = tbl_data->jump.action; } - sfx_items->type = RTE_FLOW_ITEM_TYPE_END; - sfx_items++; - tag_spec = (struct mlx5_rte_flow_item_tag *)sfx_items; - tag_spec->data = tag_id << MLX5_MTR_COLOR_BITS; - tag_spec->id = mlx5_flow_get_reg_id(dev, MLX5_MTR_SFX, 0, &error); - tag_mask = tag_spec + 1; - tag_mask->data = 0xffffff00; + actions_pre->type = RTE_FLOW_ACTION_TYPE_END; + actions_pre++; + if (!tag_action) + return rte_flow_error_set(error, ENOMEM, + RTE_FLOW_ERROR_TYPE_UNSPECIFIED, + NULL, "No tag action space."); + if (!mtr_flow_id) { + tag_action->type = RTE_FLOW_ACTION_TYPE_VOID; + goto exit; + } + /* Only default-policy Meter creates mtr flow id. */ + if (fm->def_policy) { + mlx5_ipool_malloc(fm->flow_ipool, &tag_id); + if (!tag_id) + return rte_flow_error_set(error, ENOMEM, + RTE_FLOW_ERROR_TYPE_UNSPECIFIED, NULL, + "Failed to allocate meter flow id."); + flow_id = tag_id - 1; + flow_id_bits = (!flow_id) ? 1 : + (MLX5_REG_BITS - __builtin_clz(flow_id)); + if ((flow_id_bits + priv->sh->mtrmng->max_mtr_bits) > + mtr_reg_bits) { + mlx5_ipool_free(fm->flow_ipool, tag_id); + return rte_flow_error_set(error, EINVAL, + RTE_FLOW_ERROR_TYPE_UNSPECIFIED, NULL, + "Meter flow id exceeds max limit."); + } + if (flow_id_bits > priv->sh->mtrmng->max_mtr_flow_bits) + priv->sh->mtrmng->max_mtr_flow_bits = flow_id_bits; + } + /* Build tag actions and items for meter_id/meter flow_id. */ + set_tag = (struct mlx5_rte_flow_action_set_tag *)actions_pre; + tag_item_spec = (struct mlx5_rte_flow_item_tag *)sfx_items; + tag_item_mask = tag_item_spec + 1; + /* Both flow_id and meter_id share the same register. */ + *set_tag = (struct mlx5_rte_flow_action_set_tag) { + .id = (enum modify_reg)mlx5_flow_get_reg_id(dev, MLX5_MTR_ID, + 0, error), + .offset = mtr_id_offset, + .length = mtr_reg_bits, + .data = flow->meter, + }; + /* + * The color Reg bits used by flow_id are growing from + * msb to lsb, so must do bit reverse for flow_id val in RegC. + */ + for (shift = 0; shift < flow_id_bits; shift++) + flow_id_reversed = (flow_id_reversed << 1) | + ((flow_id >> shift) & 0x1); + set_tag->data |= + flow_id_reversed << (mtr_reg_bits - flow_id_bits); + tag_item_spec->id = set_tag->id; + tag_item_spec->data = set_tag->data << mtr_id_offset; + tag_item_mask->data = UINT32_MAX << mtr_id_offset; + tag_action->type = (enum rte_flow_action_type) + MLX5_RTE_FLOW_ACTION_TYPE_TAG; + tag_action->conf = set_tag; tag_item->type = (enum rte_flow_item_type) - MLX5_RTE_FLOW_ITEM_TYPE_TAG; - tag_item->spec = tag_spec; + MLX5_RTE_FLOW_ITEM_TYPE_TAG; + tag_item->spec = tag_item_spec; tag_item->last = NULL; - tag_item->mask = tag_mask; - return tag_id; + tag_item->mask = tag_item_mask; +exit: + if (mtr_flow_id) + *mtr_flow_id = tag_id; + return 0; } /** @@ -4700,6 +5280,8 @@ flow_mreg_tx_copy_prep(struct rte_eth_dev *dev, * Pointer to the position of the matched action if exists, otherwise is -1. * @param[out] qrss_action_pos * Pointer to the position of the Queue/RSS action if exists, otherwise is -1. + * @param[out] modify_after_mirror + * Pointer to the flag of modify action after FDB mirroring. * * @return * > 0 the total number of actions. @@ -4709,14 +5291,16 @@ static int flow_check_match_action(const struct rte_flow_action actions[], const struct rte_flow_attr *attr, enum rte_flow_action_type action, - int *match_action_pos, int *qrss_action_pos) + int *match_action_pos, int *qrss_action_pos, + int *modify_after_mirror) { const struct rte_flow_action_sample *sample; + const struct rte_flow_action_raw_decap *decap; int actions_n = 0; - int jump_flag = 0; uint32_t ratio = 0; int sub_type = 0; int flag = 0; + int fdb_mirror = 0; *match_action_pos = -1; *qrss_action_pos = -1; @@ -4725,27 +5309,79 @@ flow_check_match_action(const struct rte_flow_action actions[], flag = 1; *match_action_pos = actions_n; } - if (actions->type == RTE_FLOW_ACTION_TYPE_QUEUE || - actions->type == RTE_FLOW_ACTION_TYPE_RSS) + switch (actions->type) { + case RTE_FLOW_ACTION_TYPE_QUEUE: + case RTE_FLOW_ACTION_TYPE_RSS: *qrss_action_pos = actions_n; - if (actions->type == RTE_FLOW_ACTION_TYPE_JUMP) - jump_flag = 1; - if (actions->type == RTE_FLOW_ACTION_TYPE_SAMPLE) { + break; + case RTE_FLOW_ACTION_TYPE_SAMPLE: sample = actions->conf; ratio = sample->ratio; sub_type = ((const struct rte_flow_action *) (sample->actions))->type; + if (ratio == 1 && attr->transfer) + fdb_mirror = 1; + break; + case RTE_FLOW_ACTION_TYPE_SET_MAC_SRC: + case RTE_FLOW_ACTION_TYPE_SET_MAC_DST: + case RTE_FLOW_ACTION_TYPE_SET_IPV4_SRC: + case RTE_FLOW_ACTION_TYPE_SET_IPV4_DST: + case RTE_FLOW_ACTION_TYPE_SET_IPV6_SRC: + case RTE_FLOW_ACTION_TYPE_SET_IPV6_DST: + case RTE_FLOW_ACTION_TYPE_SET_TP_SRC: + case RTE_FLOW_ACTION_TYPE_SET_TP_DST: + case RTE_FLOW_ACTION_TYPE_DEC_TTL: + case RTE_FLOW_ACTION_TYPE_SET_TTL: + case RTE_FLOW_ACTION_TYPE_INC_TCP_SEQ: + case RTE_FLOW_ACTION_TYPE_DEC_TCP_SEQ: + case RTE_FLOW_ACTION_TYPE_INC_TCP_ACK: + case RTE_FLOW_ACTION_TYPE_DEC_TCP_ACK: + case RTE_FLOW_ACTION_TYPE_SET_IPV4_DSCP: + case RTE_FLOW_ACTION_TYPE_SET_IPV6_DSCP: + case RTE_FLOW_ACTION_TYPE_FLAG: + case RTE_FLOW_ACTION_TYPE_MARK: + case RTE_FLOW_ACTION_TYPE_SET_META: + case RTE_FLOW_ACTION_TYPE_SET_TAG: + case RTE_FLOW_ACTION_TYPE_OF_POP_VLAN: + case RTE_FLOW_ACTION_TYPE_OF_PUSH_VLAN: + case RTE_FLOW_ACTION_TYPE_OF_SET_VLAN_VID: + case RTE_FLOW_ACTION_TYPE_OF_SET_VLAN_PCP: + case RTE_FLOW_ACTION_TYPE_VXLAN_DECAP: + case RTE_FLOW_ACTION_TYPE_NVGRE_DECAP: + case RTE_FLOW_ACTION_TYPE_MODIFY_FIELD: + case RTE_FLOW_ACTION_TYPE_METER: + if (fdb_mirror) + *modify_after_mirror = 1; + break; + case RTE_FLOW_ACTION_TYPE_RAW_DECAP: + decap = actions->conf; + while ((++actions)->type == RTE_FLOW_ACTION_TYPE_VOID) + ; + actions_n++; + if (actions->type == RTE_FLOW_ACTION_TYPE_RAW_ENCAP) { + const struct rte_flow_action_raw_encap *encap = + actions->conf; + if (decap->size <= + MLX5_ENCAPSULATION_DECISION_SIZE && + encap->size > + MLX5_ENCAPSULATION_DECISION_SIZE) + /* L3 encap. */ + break; + } + if (fdb_mirror) + *modify_after_mirror = 1; + break; + default: + break; } actions_n++; } - if (flag && action == RTE_FLOW_ACTION_TYPE_SAMPLE && attr->transfer) { - if (ratio == 1) { - /* JUMP Action not support for Mirroring; - * Mirroring support multi-destination; - */ - if (!jump_flag && sub_type != RTE_FLOW_ACTION_TYPE_END) - flag = 0; - } + if (flag && fdb_mirror && !*modify_after_mirror) { + /* FDB mirroring uses the destination array to implement + * instead of FLOW_SAMPLER object. + */ + if (sub_type != RTE_FLOW_ACTION_TYPE_END) + flag = 0; } /* Count RTE_FLOW_ACTION_TYPE_END. */ return flag ? actions_n + 1 : 0; @@ -4764,8 +5400,8 @@ flow_check_match_action(const struct rte_flow_action actions[], * * @param dev * Pointer to Ethernet device. - * @param[in] fdb_tx - * FDB egress flow flag. + * @param[in] add_tag + * Add extra tag action flag. * @param[out] sfx_items * Suffix flow match items (list terminated by the END pattern item). * @param[in] actions @@ -4780,6 +5416,8 @@ flow_check_match_action(const struct rte_flow_action actions[], * The sample action position. * @param[in] qrss_action_pos * The Queue/RSS action position. + * @param[in] jump_table + * Add extra jump action flag. * @param[out] error * Perform verbose error reporting if not NULL. * @@ -4789,7 +5427,7 @@ flow_check_match_action(const struct rte_flow_action actions[], */ static int flow_sample_split_prep(struct rte_eth_dev *dev, - uint32_t fdb_tx, + int add_tag, struct rte_flow_item sfx_items[], const struct rte_flow_action actions[], struct rte_flow_action actions_sfx[], @@ -4797,14 +5435,17 @@ flow_sample_split_prep(struct rte_eth_dev *dev, int actions_n, int sample_action_pos, int qrss_action_pos, + int jump_table, struct rte_flow_error *error) { struct mlx5_priv *priv = dev->data->dev_private; struct mlx5_rte_flow_action_set_tag *set_tag; struct mlx5_rte_flow_item_tag *tag_spec; struct mlx5_rte_flow_item_tag *tag_mask; + struct rte_flow_action_jump *jump_action; uint32_t tag_id = 0; int index; + int append_index = 0; int ret; if (sample_action_pos < 0) @@ -4812,34 +5453,6 @@ flow_sample_split_prep(struct rte_eth_dev *dev, RTE_FLOW_ERROR_TYPE_ACTION, NULL, "invalid position of sample " "action in list"); - if (!fdb_tx) { - /* Prepare the prefix tag action. */ - set_tag = (void *)(actions_pre + actions_n + 1); - ret = mlx5_flow_get_reg_id(dev, MLX5_APP_TAG, 0, error); - if (ret < 0) - return ret; - set_tag->id = ret; - mlx5_ipool_malloc(priv->sh->ipool - [MLX5_IPOOL_RSS_EXPANTION_FLOW_ID], &tag_id); - set_tag->data = tag_id; - /* Prepare the suffix subflow items. */ - tag_spec = (void *)(sfx_items + SAMPLE_SUFFIX_ITEM); - tag_spec->data = tag_id; - tag_spec->id = set_tag->id; - tag_mask = tag_spec + 1; - tag_mask->data = UINT32_MAX; - sfx_items[0] = (struct rte_flow_item){ - .type = (enum rte_flow_item_type) - MLX5_RTE_FLOW_ITEM_TYPE_TAG, - .spec = tag_spec, - .last = NULL, - .mask = tag_mask, - }; - sfx_items[1] = (struct rte_flow_item){ - .type = (enum rte_flow_item_type) - RTE_FLOW_ITEM_TYPE_END, - }; - } /* Prepare the actions for prefix and suffix flow. */ if (qrss_action_pos >= 0 && qrss_action_pos < sample_action_pos) { index = qrss_action_pos; @@ -4863,8 +5476,41 @@ flow_sample_split_prep(struct rte_eth_dev *dev, memcpy(actions_pre, actions, sizeof(struct rte_flow_action) * index); } - /* Add the extra tag action for NIC-RX and E-Switch ingress. */ - if (!fdb_tx) { + /* For CX5, add an extra tag action for NIC-RX and E-Switch ingress. + * For CX6DX and above, metadata registers Cx preserve their value, + * add an extra tag action for NIC-RX and E-Switch Domain. + */ + if (add_tag) { + /* Prepare the prefix tag action. */ + append_index++; + set_tag = (void *)(actions_pre + actions_n + append_index); + ret = mlx5_flow_get_reg_id(dev, MLX5_APP_TAG, 0, error); + if (ret < 0) + return ret; + mlx5_ipool_malloc(priv->sh->ipool + [MLX5_IPOOL_RSS_EXPANTION_FLOW_ID], &tag_id); + *set_tag = (struct mlx5_rte_flow_action_set_tag) { + .id = ret, + .data = tag_id, + }; + /* Prepare the suffix subflow items. */ + tag_spec = (void *)(sfx_items + SAMPLE_SUFFIX_ITEM); + tag_spec->data = tag_id; + tag_spec->id = set_tag->id; + tag_mask = tag_spec + 1; + tag_mask->data = UINT32_MAX; + sfx_items[0] = (struct rte_flow_item){ + .type = (enum rte_flow_item_type) + MLX5_RTE_FLOW_ITEM_TYPE_TAG, + .spec = tag_spec, + .last = NULL, + .mask = tag_mask, + }; + sfx_items[1] = (struct rte_flow_item){ + .type = (enum rte_flow_item_type) + RTE_FLOW_ITEM_TYPE_END, + }; + /* Prepare the tag action in prefix subflow. */ actions_pre[index++] = (struct rte_flow_action){ .type = (enum rte_flow_action_type) @@ -4875,6 +5521,22 @@ flow_sample_split_prep(struct rte_eth_dev *dev, memcpy(actions_pre + index, actions + sample_action_pos, sizeof(struct rte_flow_action)); index += 1; + /* For the modify action after the sample action in E-Switch mirroring, + * Add the extra jump action in prefix subflow and jump into the next + * table, then do the modify action in the new table. + */ + if (jump_table) { + /* Prepare the prefix jump action. */ + append_index++; + jump_action = (void *)(actions_pre + actions_n + append_index); + jump_action->group = jump_table; + actions_pre[index++] = + (struct rte_flow_action){ + .type = (enum rte_flow_action_type) + RTE_FLOW_ACTION_TYPE_JUMP, + .conf = jump_action, + }; + } actions_pre[index] = (struct rte_flow_action){ .type = (enum rte_flow_action_type) RTE_FLOW_ACTION_TYPE_END, @@ -4963,8 +5625,8 @@ flow_create_split_metadata(struct rte_eth_dev *dev, if (qrss) { /* Check if it is in meter suffix table. */ mtr_sfx = attr->group == (attr->transfer ? - (MLX5_FLOW_TABLE_LEVEL_SUFFIX - 1) : - MLX5_FLOW_TABLE_LEVEL_SUFFIX); + (MLX5_FLOW_TABLE_LEVEL_METER - 1) : + MLX5_FLOW_TABLE_LEVEL_METER); /* * Q/RSS action on NIC Rx should be split in order to pass by * the mreg copy table (RX_CP_TBL) and then it jumps to the @@ -5112,6 +5774,57 @@ exit: return ret; } +/** + * Create meter internal drop flow with the original pattern. + * + * @param dev + * Pointer to Ethernet device. + * @param[in] flow + * Parent flow structure pointer. + * @param[in] attr + * Flow rule attributes. + * @param[in] items + * Pattern specification (list terminated by the END pattern item). + * @param[in] flow_split_info + * Pointer to flow split info structure. + * @param[in] fm + * Pointer to flow meter structure. + * @param[out] error + * Perform verbose error reporting if not NULL. + * @return + * 0 on success, negative value otherwise + */ +static uint32_t +flow_meter_create_drop_flow_with_org_pattern(struct rte_eth_dev *dev, + struct rte_flow *flow, + const struct rte_flow_attr *attr, + const struct rte_flow_item items[], + struct mlx5_flow_split_info *flow_split_info, + struct mlx5_flow_meter_info *fm, + struct rte_flow_error *error) +{ + struct mlx5_flow *dev_flow = NULL; + struct rte_flow_attr drop_attr = *attr; + struct rte_flow_action drop_actions[3]; + struct mlx5_flow_split_info drop_split_info = *flow_split_info; + + MLX5_ASSERT(fm->drop_cnt); + drop_actions[0].type = + (enum rte_flow_action_type)MLX5_RTE_FLOW_ACTION_TYPE_COUNT; + drop_actions[0].conf = (void *)(uintptr_t)fm->drop_cnt; + drop_actions[1].type = RTE_FLOW_ACTION_TYPE_DROP; + drop_actions[1].conf = NULL; + drop_actions[2].type = RTE_FLOW_ACTION_TYPE_END; + drop_actions[2].conf = NULL; + drop_split_info.external = false; + drop_split_info.skip_scale |= 1 << MLX5_SCALE_FLOW_GROUP_BIT; + drop_split_info.table_id = MLX5_MTR_TABLE_ID_DROP; + drop_attr.group = MLX5_FLOW_TABLE_LEVEL_METER; + return flow_create_split_inner(dev, flow, &dev_flow, + &drop_attr, items, drop_actions, + &drop_split_info, error); +} + /** * The splitting for meter feature. * @@ -5149,25 +5862,82 @@ flow_create_split_meter(struct rte_eth_dev *dev, struct rte_flow_error *error) { struct mlx5_priv *priv = dev->data->dev_private; + struct mlx5_flow_workspace *wks = mlx5_flow_get_thread_workspace(); struct rte_flow_action *sfx_actions = NULL; struct rte_flow_action *pre_actions = NULL; struct rte_flow_item *sfx_items = NULL; struct mlx5_flow *dev_flow = NULL; struct rte_flow_attr sfx_attr = *attr; - uint32_t mtr = 0; - uint32_t mtr_tag_id = 0; + struct mlx5_flow_meter_info *fm = NULL; + uint8_t skip_scale_restore; + bool has_mtr = false; + bool has_modify = false; + bool set_mtr_reg = true; + bool is_mtr_hierarchy = false; + uint32_t meter_id = 0; + uint32_t mtr_idx = 0; + uint32_t mtr_flow_id = 0; size_t act_size; size_t item_size; int actions_n = 0; - int ret; + int ret = 0; if (priv->mtr_en) - actions_n = flow_check_meter_action(actions, &mtr); - if (mtr) { - /* The five prefix actions: meter, decap, encap, tag, end. */ - act_size = sizeof(struct rte_flow_action) * (actions_n + 5) + + actions_n = flow_check_meter_action(dev, actions, &has_mtr, + &has_modify, &meter_id); + if (has_mtr) { + if (flow->meter) { + fm = flow_dv_meter_find_by_idx(priv, flow->meter); + if (!fm) + return rte_flow_error_set(error, EINVAL, + RTE_FLOW_ERROR_TYPE_UNSPECIFIED, + NULL, "Meter not found."); + } else { + fm = mlx5_flow_meter_find(priv, meter_id, &mtr_idx); + if (!fm) + return rte_flow_error_set(error, EINVAL, + RTE_FLOW_ERROR_TYPE_UNSPECIFIED, + NULL, "Meter not found."); + ret = mlx5_flow_meter_attach(priv, fm, + &sfx_attr, error); + if (ret) + return -rte_errno; + flow->meter = mtr_idx; + } + MLX5_ASSERT(wks); + wks->fm = fm; + if (!fm->def_policy) { + wks->policy = mlx5_flow_meter_policy_find(dev, + fm->policy_id, + NULL); + MLX5_ASSERT(wks->policy); + if (wks->policy->is_hierarchy) { + wks->final_policy = + mlx5_flow_meter_hierarchy_get_final_policy(dev, + wks->policy); + if (!wks->final_policy) + return rte_flow_error_set(error, + EINVAL, + RTE_FLOW_ERROR_TYPE_ACTION, NULL, + "Failed to find terminal policy of hierarchy."); + is_mtr_hierarchy = true; + } + } + /* + * If it isn't default-policy Meter, and + * 1. There's no action in flow to change + * packet (modify/encap/decap etc.), OR + * 2. No drop count needed for this meter. + * 3. It's not meter hierarchy. + * Then no need to use regC to save meter id anymore. + */ + if (!fm->def_policy && !is_mtr_hierarchy && + (!has_modify || !fm->drop_cnt)) + set_mtr_reg = false; + /* Prefix actions: meter, decap, encap, tag, jump, end. */ + act_size = sizeof(struct rte_flow_action) * (actions_n + 6) + sizeof(struct mlx5_rte_flow_action_set_tag); - /* tag, vlan, port id, end. */ + /* Suffix items: tag, vlan, port id, end. */ #define METER_SUFFIX_ITEM 4 item_size = sizeof(struct rte_flow_item) * METER_SUFFIX_ITEM + sizeof(struct mlx5_rte_flow_item_tag) * 2; @@ -5180,31 +5950,56 @@ flow_create_split_meter(struct rte_eth_dev *dev, "meter flow"); sfx_items = (struct rte_flow_item *)((char *)sfx_actions + act_size); - pre_actions = sfx_actions + actions_n; - mtr_tag_id = flow_meter_split_prep(dev, items, sfx_items, - actions, sfx_actions, - pre_actions); - if (!mtr_tag_id) { + /* There's no suffix flow for meter of non-default policy. */ + if (!fm->def_policy) + pre_actions = sfx_actions + 1; + else + pre_actions = sfx_actions + actions_n; + ret = flow_meter_split_prep(dev, flow, wks, &sfx_attr, + items, sfx_items, actions, + sfx_actions, pre_actions, + (set_mtr_reg ? &mtr_flow_id : NULL), + error); + if (ret) { ret = -rte_errno; goto exit; } /* Add the prefix subflow. */ flow_split_info->prefix_mark = 0; + skip_scale_restore = flow_split_info->skip_scale; + flow_split_info->skip_scale |= + 1 << MLX5_SCALE_JUMP_FLOW_GROUP_BIT; ret = flow_create_split_inner(dev, flow, &dev_flow, attr, items, pre_actions, flow_split_info, error); + flow_split_info->skip_scale = skip_scale_restore; if (ret) { + if (mtr_flow_id) + mlx5_ipool_free(fm->flow_ipool, mtr_flow_id); ret = -rte_errno; goto exit; } - dev_flow->handle->split_flow_id = mtr_tag_id; + if (mtr_flow_id) { + dev_flow->handle->split_flow_id = mtr_flow_id; + dev_flow->handle->is_meter_flow_id = 1; + } + if (!fm->def_policy) { + if (!set_mtr_reg && fm->drop_cnt) + ret = + flow_meter_create_drop_flow_with_org_pattern(dev, flow, + &sfx_attr, items, + flow_split_info, + fm, error); + goto exit; + } /* Setting the sfx group atrr. */ sfx_attr.group = sfx_attr.transfer ? - (MLX5_FLOW_TABLE_LEVEL_SUFFIX - 1) : - MLX5_FLOW_TABLE_LEVEL_SUFFIX; + (MLX5_FLOW_TABLE_LEVEL_METER - 1) : + MLX5_FLOW_TABLE_LEVEL_METER; flow_split_info->prefix_layers = flow_get_prefix_layer_flags(dev_flow); flow_split_info->prefix_mark = dev_flow->handle->mark; + flow_split_info->table_id = MLX5_MTR_TABLE_ID_SUFFIX; } /* Add the prefix subflow. */ ret = flow_create_split_metadata(dev, flow, @@ -5269,7 +6064,6 @@ flow_create_split_sample(struct rte_eth_dev *dev, struct mlx5_flow_dv_sample_resource *sample_res; struct mlx5_flow_tbl_data_entry *sfx_tbl_data; struct mlx5_flow_tbl_resource *sfx_tbl; - union mlx5_flow_tbl_key sfx_table_key; #endif size_t act_size; size_t item_size; @@ -5278,12 +6072,17 @@ flow_create_split_sample(struct rte_eth_dev *dev, int actions_n = 0; int sample_action_pos; int qrss_action_pos; + int add_tag = 0; + int modify_after_mirror = 0; + uint16_t jump_table = 0; + const uint32_t next_ft_step = 1; int ret = 0; if (priv->sampler_en) actions_n = flow_check_match_action(actions, attr, RTE_FLOW_ACTION_TYPE_SAMPLE, - &sample_action_pos, &qrss_action_pos); + &sample_action_pos, &qrss_action_pos, + &modify_after_mirror); if (actions_n) { /* The prefix actions must includes sample, tag, end. */ act_size = sizeof(struct rte_flow_action) * (actions_n * 2 + 1) @@ -5297,21 +6096,33 @@ flow_create_split_sample(struct rte_eth_dev *dev, RTE_FLOW_ERROR_TYPE_ACTION, NULL, "no memory to split " "sample flow"); - /* The representor_id is -1 for uplink. */ - fdb_tx = (attr->transfer && priv->representor_id != -1); - if (!fdb_tx) + /* The representor_id is UINT16_MAX for uplink. */ + fdb_tx = (attr->transfer && priv->representor_id != UINT16_MAX); + /* + * When reg_c_preserve is set, metadata registers Cx preserve + * their value even through packet duplication. + */ + add_tag = (!fdb_tx || priv->config.hca_attr.reg_c_preserve); + if (add_tag) sfx_items = (struct rte_flow_item *)((char *)sfx_actions + act_size); + if (modify_after_mirror) + jump_table = attr->group * MLX5_FLOW_TABLE_FACTOR + + next_ft_step; pre_actions = sfx_actions + actions_n; - tag_id = flow_sample_split_prep(dev, fdb_tx, sfx_items, + tag_id = flow_sample_split_prep(dev, add_tag, sfx_items, actions, sfx_actions, pre_actions, actions_n, sample_action_pos, - qrss_action_pos, error); - if (tag_id < 0 || (!fdb_tx && !tag_id)) { + qrss_action_pos, jump_table, + error); + if (tag_id < 0 || (add_tag && !tag_id)) { ret = -rte_errno; goto exit; } + if (modify_after_mirror) + flow_split_info->skip_scale = + 1 << MLX5_SCALE_JUMP_FLOW_GROUP_BIT; /* Add the prefix subflow. */ ret = flow_create_split_inner(dev, flow, &dev_flow, attr, items, pre_actions, @@ -5322,24 +6133,29 @@ flow_create_split_sample(struct rte_eth_dev *dev, } dev_flow->handle->split_flow_id = tag_id; #ifdef HAVE_IBV_FLOW_DV_SUPPORT - /* Set the sfx group attr. */ - sample_res = (struct mlx5_flow_dv_sample_resource *) - dev_flow->dv.sample_res; - sfx_tbl = (struct mlx5_flow_tbl_resource *) - sample_res->normal_path_tbl; - sfx_tbl_data = container_of(sfx_tbl, - struct mlx5_flow_tbl_data_entry, tbl); - sfx_table_key.v64 = sfx_tbl_data->entry.key; - sfx_attr.group = sfx_attr.transfer ? - (sfx_table_key.table_id - 1) : - sfx_table_key.table_id; + if (!modify_after_mirror) { + /* Set the sfx group attr. */ + sample_res = (struct mlx5_flow_dv_sample_resource *) + dev_flow->dv.sample_res; + sfx_tbl = (struct mlx5_flow_tbl_resource *) + sample_res->normal_path_tbl; + sfx_tbl_data = container_of(sfx_tbl, + struct mlx5_flow_tbl_data_entry, + tbl); + sfx_attr.group = sfx_attr.transfer ? + (sfx_tbl_data->level - 1) : sfx_tbl_data->level; + } else { + MLX5_ASSERT(attr->transfer); + sfx_attr.group = jump_table; + } flow_split_info->prefix_layers = flow_get_prefix_layer_flags(dev_flow); flow_split_info->prefix_mark = dev_flow->handle->mark; /* Suffix group level already be scaled with factor, set - * skip_scale to 1 to avoid scale again in translation. + * MLX5_SCALE_FLOW_GROUP_BIT of skip_scale to 1 to avoid scale + * again in translation. */ - flow_split_info->skip_scale = 1; + flow_split_info->skip_scale = 1 << MLX5_SCALE_FLOW_GROUP_BIT; #endif } /* Add the suffix subflow. */ @@ -5405,22 +6221,14 @@ flow_create_split_outer(struct rte_eth_dev *dev, return ret; } -static struct mlx5_flow_tunnel * -flow_tunnel_from_rule(struct rte_eth_dev *dev, - const struct rte_flow_attr *attr, - const struct rte_flow_item items[], - const struct rte_flow_action actions[]) +static inline struct mlx5_flow_tunnel * +flow_tunnel_from_rule(const struct mlx5_flow *flow) { struct mlx5_flow_tunnel *tunnel; #pragma GCC diagnostic push #pragma GCC diagnostic ignored "-Wcast-qual" - if (is_flow_tunnel_match_rule(dev, attr, items, actions)) - tunnel = (struct mlx5_flow_tunnel *)items[0].spec; - else if (is_flow_tunnel_steer_rule(dev, attr, items, actions)) - tunnel = (struct mlx5_flow_tunnel *)actions[0].conf; - else - tunnel = NULL; + tunnel = (typeof(tunnel))flow->tunnel; #pragma GCC diagnostic pop return tunnel; @@ -5444,17 +6252,15 @@ flow_rss_workspace_adjust(struct mlx5_flow_workspace *wks, struct mlx5_flow_rss_desc *rss_desc, uint32_t nrssq_num) { - bool fidx = !!wks->flow_idx; - - if (likely(nrssq_num <= wks->rssq_num[fidx])) + if (likely(nrssq_num <= wks->rssq_num)) return 0; rss_desc->queue = realloc(rss_desc->queue, - sizeof(rss_desc->queue[0]) * RTE_ALIGN(nrssq_num, 2)); + sizeof(*rss_desc->queue) * RTE_ALIGN(nrssq_num, 2)); if (!rss_desc->queue) { rte_errno = ENOMEM; return -1; } - wks->rssq_num[fidx] = RTE_ALIGN(nrssq_num, 2); + wks->rssq_num = RTE_ALIGN(nrssq_num, 2); return 0; } @@ -5483,7 +6289,7 @@ flow_rss_workspace_adjust(struct mlx5_flow_workspace *wks, * A flow index on success, 0 otherwise and rte_errno is set. */ static uint32_t -flow_list_create(struct rte_eth_dev *dev, uint32_t *list, +flow_list_create(struct rte_eth_dev *dev, enum mlx5_flow_type type, const struct rte_flow_attr *attr, const struct rte_flow_item items[], const struct rte_flow_action original_actions[], @@ -5492,13 +6298,13 @@ flow_list_create(struct rte_eth_dev *dev, uint32_t *list, struct mlx5_priv *priv = dev->data->dev_private; struct rte_flow *flow = NULL; struct mlx5_flow *dev_flow; - const struct rte_flow_action_rss *rss; - struct mlx5_translated_shared_action - shared_actions[MLX5_MAX_SHARED_ACTIONS]; - int shared_actions_n = MLX5_MAX_SHARED_ACTIONS; + const struct rte_flow_action_rss *rss = NULL; + struct mlx5_translated_action_handle + indir_actions[MLX5_MAX_INDIRECT_ACTIONS]; + int indir_actions_n = MLX5_MAX_INDIRECT_ACTIONS; union { struct mlx5_flow_expand_rss buf; - uint8_t buffer[2048]; + uint8_t buffer[4096]; } expand_buffer; union { struct rte_flow_action actions[MLX5_MAX_SPLIT_ACTIONS]; @@ -5523,22 +6329,22 @@ flow_list_create(struct rte_eth_dev *dev, uint32_t *list, struct rte_flow_action *translated_actions = NULL; struct mlx5_flow_tunnel *tunnel; struct tunnel_default_miss_ctx default_miss_ctx = { 0, }; + struct mlx5_flow_workspace *wks = mlx5_flow_push_thread_workspace(); struct mlx5_flow_split_info flow_split_info = { .external = !!external, .skip_scale = 0, .flow_idx = 0, .prefix_mark = 0, - .prefix_layers = 0 + .prefix_layers = 0, + .table_id = 0 }; - struct mlx5_flow_workspace *wks = mlx5_flow_get_thread_workspace(); - bool fidx = !!wks->flow_idx; int ret; MLX5_ASSERT(wks); - rss_desc = &wks->rss_desc[fidx]; - ret = flow_shared_actions_translate(dev, original_actions, - shared_actions, - &shared_actions_n, + rss_desc = &wks->rss_desc; + ret = flow_action_handles_translate(dev, original_actions, + indir_actions, + &indir_actions_n, &translated_actions, error); if (ret < 0) { MLX5_ASSERT(translated_actions == NULL); @@ -5551,7 +6357,7 @@ flow_list_create(struct rte_eth_dev *dev, uint32_t *list, external, hairpin_flow, error); if (ret < 0) goto error_before_hairpin_split; - flow = mlx5_ipool_zmalloc(priv->sh->ipool[MLX5_IPOOL_RTE_FLOW], &idx); + flow = mlx5_ipool_zmalloc(priv->flows[type], &idx); if (!flow) { rte_errno = ENOMEM; goto error_before_hairpin_split; @@ -5571,7 +6377,9 @@ flow_list_create(struct rte_eth_dev *dev, uint32_t *list, MLX5_ASSERT(flow->drv_type > MLX5_FLOW_TYPE_MIN && flow->drv_type < MLX5_FLOW_TYPE_MAX); memset(rss_desc, 0, offsetof(struct mlx5_flow_rss_desc, queue)); - rss = flow_get_rss_action(p_actions_rx); + /* RSS Action only works on NIC RX domain */ + if (attr->ingress && !attr->transfer) + rss = flow_get_rss_action(dev, p_actions_rx); if (rss) { if (flow_rss_workspace_adjust(wks, rss_desc, rss->queue_num)) return 0; @@ -5587,27 +6395,22 @@ flow_list_create(struct rte_eth_dev *dev, uint32_t *list, if (rss && rss->types) { unsigned int graph_root; - graph_root = find_graph_root(items, rss->level); + graph_root = find_graph_root(rss->level); ret = mlx5_flow_expand_rss(buf, sizeof(expand_buffer.buffer), items, rss->types, mlx5_support_expansion, graph_root); MLX5_ASSERT(ret > 0 && (unsigned int)ret < sizeof(expand_buffer.buffer)); + if (rte_log_can_log(mlx5_logtype, RTE_LOG_DEBUG)) { + for (i = 0; i < buf->entries; ++i) + mlx5_dbg__print_pattern(buf->entry[i].pattern); + } } else { buf->entries = 1; buf->entry[0].pattern = (void *)(uintptr_t)items; } - rss_desc->shared_rss = flow_get_shared_rss_action(dev, shared_actions, - shared_actions_n); - /* - * Record the start index when there is a nested call. All sub-flows - * need to be translated before another calling. - * No need to use ping-pong buffer to save memory here. - */ - if (fidx) { - MLX5_ASSERT(!wks->flow_nested_idx); - wks->flow_nested_idx = fidx; - } + rss_desc->shared_rss = flow_get_shared_rss_action(dev, indir_actions, + indir_actions_n); for (i = 0; i < buf->entries; ++i) { /* Initialize flow split data. */ flow_split_info.prefix_layers = 0; @@ -5624,12 +6427,11 @@ flow_list_create(struct rte_eth_dev *dev, uint32_t *list, error); if (ret < 0) goto error; - if (is_flow_tunnel_steer_rule(dev, attr, - buf->entry[i].pattern, - p_actions_rx)) { + if (is_flow_tunnel_steer_rule(wks->flows[0].tof_type)) { ret = flow_tunnel_add_default_miss(dev, flow, attr, p_actions_rx, idx, + wks->flows[0].tunnel, &default_miss_ctx, error); if (ret < 0) { @@ -5675,33 +6477,27 @@ flow_list_create(struct rte_eth_dev *dev, uint32_t *list, goto error; } /* - * If the flow is external (from application) OR device is started, then - * the flow will be applied immediately. + * If the flow is external (from application) OR device is started, + * OR mreg discover, then apply immediately. */ - if (external || dev->data->dev_started) { + if (external || dev->data->dev_started || + (attr->group == MLX5_FLOW_MREG_CP_TABLE_GROUP && + attr->priority == MLX5_FLOW_LOWEST_PRIO_INDICATOR)) { ret = flow_drv_apply(dev, flow, error); if (ret < 0) goto error; } - if (list) { - rte_spinlock_lock(&priv->flow_list_lock); - ILIST_INSERT(priv->sh->ipool[MLX5_IPOOL_RTE_FLOW], list, idx, - flow, next); - rte_spinlock_unlock(&priv->flow_list_lock); - } + flow->type = type; flow_rxq_flags_set(dev, flow); rte_free(translated_actions); - /* Nested flow creation index recovery. */ - wks->flow_idx = wks->flow_nested_idx; - if (wks->flow_nested_idx) - wks->flow_nested_idx = 0; - tunnel = flow_tunnel_from_rule(dev, attr, items, actions); + tunnel = flow_tunnel_from_rule(wks->flows); if (tunnel) { flow->tunnel = 1; flow->tunnel_id = tunnel->tunnel_id; __atomic_add_fetch(&tunnel->refctn, 1, __ATOMIC_RELAXED); mlx5_free(default_miss_ctx.queue); } + mlx5_flow_pop_thread_workspace(); return idx; error: MLX5_ASSERT(flow); @@ -5713,13 +6509,11 @@ error: mlx5_ipool_get (priv->sh->ipool[MLX5_IPOOL_RSS_SHARED_ACTIONS], rss_desc->shared_rss))->refcnt, 1, __ATOMIC_RELAXED); - mlx5_ipool_free(priv->sh->ipool[MLX5_IPOOL_RTE_FLOW], idx); + mlx5_ipool_free(priv->flows[type], idx); rte_errno = ret; /* Restore rte_errno. */ ret = rte_errno; rte_errno = ret; - wks->flow_idx = wks->flow_nested_idx; - if (wks->flow_nested_idx) - wks->flow_nested_idx = 0; + mlx5_flow_pop_thread_workspace(); error_before_hairpin_split: rte_free(translated_actions); return 0; @@ -5767,10 +6561,9 @@ mlx5_flow_create_esw_table_zero_flow(struct rte_eth_dev *dev) .type = RTE_FLOW_ACTION_TYPE_END, }, }; - struct mlx5_priv *priv = dev->data->dev_private; struct rte_flow_error error; - return (void *)(uintptr_t)flow_list_create(dev, &priv->ctrl_flows, + return (void *)(uintptr_t)flow_list_create(dev, MLX5_FLOW_TYPE_CTL, &attr, &pattern, actions, false, &error); } @@ -5789,14 +6582,14 @@ mlx5_flow_validate(struct rte_eth_dev *dev, struct rte_flow_error *error) { int hairpin_flow; - struct mlx5_translated_shared_action - shared_actions[MLX5_MAX_SHARED_ACTIONS]; - int shared_actions_n = MLX5_MAX_SHARED_ACTIONS; + struct mlx5_translated_action_handle + indir_actions[MLX5_MAX_INDIRECT_ACTIONS]; + int indir_actions_n = MLX5_MAX_INDIRECT_ACTIONS; const struct rte_flow_action *actions; struct rte_flow_action *translated_actions = NULL; - int ret = flow_shared_actions_translate(dev, original_actions, - shared_actions, - &shared_actions_n, + int ret = flow_action_handles_translate(dev, original_actions, + indir_actions, + &indir_actions_n, &translated_actions, error); if (ret) @@ -5822,8 +6615,6 @@ mlx5_flow_create(struct rte_eth_dev *dev, const struct rte_flow_action actions[], struct rte_flow_error *error) { - struct mlx5_priv *priv = dev->data->dev_private; - /* * If the device is not started yet, it is not allowed to created a * flow from application. PMD default flows and traffic control flows @@ -5839,8 +6630,9 @@ mlx5_flow_create(struct rte_eth_dev *dev, return NULL; } - return (void *)(uintptr_t)flow_list_create(dev, &priv->flows, - attr, items, actions, true, error); + return (void *)(uintptr_t)flow_list_create(dev, MLX5_FLOW_TYPE_GEN, + attr, items, actions, + true, error); } /** @@ -5848,24 +6640,19 @@ mlx5_flow_create(struct rte_eth_dev *dev, * * @param dev * Pointer to Ethernet device. - * @param list - * Pointer to the Indexed flow list. If this parameter NULL, - * there is no flow removal from the list. Be noted that as - * flow is add to the indexed list, memory of the indexed - * list points to maybe changed as flow destroyed. * @param[in] flow_idx * Index of flow to destroy. */ static void -flow_list_destroy(struct rte_eth_dev *dev, uint32_t *list, +flow_list_destroy(struct rte_eth_dev *dev, enum mlx5_flow_type type, uint32_t flow_idx) { struct mlx5_priv *priv = dev->data->dev_private; - struct rte_flow *flow = mlx5_ipool_get(priv->sh->ipool - [MLX5_IPOOL_RTE_FLOW], flow_idx); + struct rte_flow *flow = mlx5_ipool_get(priv->flows[type], flow_idx); if (!flow) return; + MLX5_ASSERT(flow->type == type); /* * Update RX queue flags only if port is started, otherwise it is * already clean. @@ -5873,25 +6660,16 @@ flow_list_destroy(struct rte_eth_dev *dev, uint32_t *list, if (dev->data->dev_started) flow_rxq_flags_trim(dev, flow); flow_drv_destroy(dev, flow); - if (list) { - rte_spinlock_lock(&priv->flow_list_lock); - ILIST_REMOVE(priv->sh->ipool[MLX5_IPOOL_RTE_FLOW], list, - flow_idx, flow, next); - rte_spinlock_unlock(&priv->flow_list_lock); - } if (flow->tunnel) { struct mlx5_flow_tunnel *tunnel; - rte_spinlock_lock(&mlx5_tunnel_hub(dev)->sl); tunnel = mlx5_find_tunnel_id(dev, flow->tunnel_id); RTE_VERIFY(tunnel); - LIST_REMOVE(tunnel, chain); - rte_spinlock_unlock(&mlx5_tunnel_hub(dev)->sl); if (!__atomic_sub_fetch(&tunnel->refctn, 1, __ATOMIC_RELAXED)) mlx5_flow_tunnel_free(dev, tunnel); } flow_mreg_del_copy_action(dev, flow); - mlx5_ipool_free(priv->sh->ipool[MLX5_IPOOL_RTE_FLOW], flow_idx); + mlx5_ipool_free(priv->flows[type], flow_idx); } /** @@ -5899,18 +6677,21 @@ flow_list_destroy(struct rte_eth_dev *dev, uint32_t *list, * * @param dev * Pointer to Ethernet device. - * @param list - * Pointer to the Indexed flow list. + * @param type + * Flow type to be flushed. * @param active * If flushing is called avtively. */ void -mlx5_flow_list_flush(struct rte_eth_dev *dev, uint32_t *list, bool active) +mlx5_flow_list_flush(struct rte_eth_dev *dev, enum mlx5_flow_type type, + bool active) { - uint32_t num_flushed = 0; + struct mlx5_priv *priv = dev->data->dev_private; + uint32_t num_flushed = 0, fidx = 1; + struct rte_flow *flow; - while (*list) { - flow_list_destroy(dev, list, *list); + MLX5_IPOOL_FOREACH(priv->flows[type], fidx, flow) { + flow_list_destroy(dev, type, fidx); num_flushed++; } if (active) { @@ -5952,75 +6733,127 @@ mlx5_flow_start_default(struct rte_eth_dev *dev) /** * Release key of thread specific flow workspace data. */ -static void +void flow_release_workspace(void *data) { struct mlx5_flow_workspace *wks = data; + struct mlx5_flow_workspace *next; - if (!wks) - return; - free(wks->rss_desc[0].queue); - free(wks->rss_desc[1].queue); - free(wks); + while (wks) { + next = wks->next; + free(wks->rss_desc.queue); + free(wks); + wks = next; + } } /** - * Initialize key of thread specific flow workspace data. + * Get thread specific current flow workspace. + * + * @return pointer to thread specific flow workspace data, NULL on error. */ -static void -flow_alloc_workspace(void) +struct mlx5_flow_workspace* +mlx5_flow_get_thread_workspace(void) { - if (pthread_key_create(&key_workspace, flow_release_workspace)) - DRV_LOG(ERR, "Can't create flow workspace data thread key."); + struct mlx5_flow_workspace *data; + + data = mlx5_flow_os_get_specific_workspace(); + MLX5_ASSERT(data && data->inuse); + if (!data || !data->inuse) + DRV_LOG(ERR, "flow workspace not initialized."); + return data; } /** - * Get thread specific flow workspace. + * Allocate and init new flow workspace. * - * @return pointer to thread specific flowworkspace data, NULL on error. + * @return pointer to flow workspace data, NULL on error. */ -struct mlx5_flow_workspace* -mlx5_flow_get_thread_workspace(void) +static struct mlx5_flow_workspace* +flow_alloc_thread_workspace(void) { - struct mlx5_flow_workspace *data; + struct mlx5_flow_workspace *data = calloc(1, sizeof(*data)); - if (pthread_once(&key_workspace_init, flow_alloc_workspace)) { - DRV_LOG(ERR, "Failed to init flow workspace data thread key."); - return NULL; - } - data = pthread_getspecific(key_workspace); if (!data) { - data = calloc(1, sizeof(*data)); - if (!data) { - DRV_LOG(ERR, "Failed to allocate flow workspace " - "memory."); - return NULL; - } - data->rss_desc[0].queue = calloc(1, - sizeof(uint16_t) * MLX5_RSSQ_DEFAULT_NUM); - if (!data->rss_desc[0].queue) - goto err; - data->rss_desc[1].queue = calloc(1, - sizeof(uint16_t) * MLX5_RSSQ_DEFAULT_NUM); - if (!data->rss_desc[1].queue) - goto err; - data->rssq_num[0] = MLX5_RSSQ_DEFAULT_NUM; - data->rssq_num[1] = MLX5_RSSQ_DEFAULT_NUM; - if (pthread_setspecific(key_workspace, data)) { - DRV_LOG(ERR, "Failed to set flow workspace to thread."); - goto err; - } + DRV_LOG(ERR, "Failed to allocate flow workspace " + "memory."); + return NULL; } + data->rss_desc.queue = calloc(1, + sizeof(uint16_t) * MLX5_RSSQ_DEFAULT_NUM); + if (!data->rss_desc.queue) + goto err; + data->rssq_num = MLX5_RSSQ_DEFAULT_NUM; return data; err: - if (data->rss_desc[0].queue) - free(data->rss_desc[0].queue); - if (data->rss_desc[1].queue) - free(data->rss_desc[1].queue); + if (data->rss_desc.queue) + free(data->rss_desc.queue); free(data); return NULL; } +/** + * Get new thread specific flow workspace. + * + * If current workspace inuse, create new one and set as current. + * + * @return pointer to thread specific flow workspace data, NULL on error. + */ +static struct mlx5_flow_workspace* +mlx5_flow_push_thread_workspace(void) +{ + struct mlx5_flow_workspace *curr; + struct mlx5_flow_workspace *data; + + curr = mlx5_flow_os_get_specific_workspace(); + if (!curr) { + data = flow_alloc_thread_workspace(); + if (!data) + return NULL; + } else if (!curr->inuse) { + data = curr; + } else if (curr->next) { + data = curr->next; + } else { + data = flow_alloc_thread_workspace(); + if (!data) + return NULL; + curr->next = data; + data->prev = curr; + } + data->inuse = 1; + data->flow_idx = 0; + /* Set as current workspace */ + if (mlx5_flow_os_set_specific_workspace(data)) + DRV_LOG(ERR, "Failed to set flow workspace to thread."); + return data; +} + +/** + * Close current thread specific flow workspace. + * + * If previous workspace available, set it as current. + * + * @return pointer to thread specific flow workspace data, NULL on error. + */ +static void +mlx5_flow_pop_thread_workspace(void) +{ + struct mlx5_flow_workspace *data = mlx5_flow_get_thread_workspace(); + + if (!data) + return; + if (!data->inuse) { + DRV_LOG(ERR, "Failed to close unused flow workspace."); + return; + } + data->inuse = 0; + if (!data->prev) + return; + if (mlx5_flow_os_set_specific_workspace(data->prev)) + DRV_LOG(ERR, "Failed to set flow workspace to thread."); +} + /** * Verify the flow list is empty * @@ -6030,18 +6863,19 @@ err: * @return the number of flows not released. */ int -mlx5_flow_verify(struct rte_eth_dev *dev) +mlx5_flow_verify(struct rte_eth_dev *dev __rte_unused) { struct mlx5_priv *priv = dev->data->dev_private; struct rte_flow *flow; - uint32_t idx; - int ret = 0; + uint32_t idx = 0; + int ret = 0, i; - ILIST_FOREACH(priv->sh->ipool[MLX5_IPOOL_RTE_FLOW], priv->flows, idx, - flow, next) { - DRV_LOG(DEBUG, "port %u flow %p still referenced", - dev->data->port_id, (void *)flow); - ++ret; + for (i = 0; i < MLX5_FLOW_TYPE_MAXI; i++) { + MLX5_IPOOL_FOREACH(priv->flows[i], idx, flow) { + DRV_LOG(DEBUG, "port %u flow %p still referenced", + dev->data->port_id, (void *)flow); + ret++; + } } return ret; } @@ -6061,7 +6895,6 @@ int mlx5_ctrl_flow_source_queue(struct rte_eth_dev *dev, uint32_t queue) { - struct mlx5_priv *priv = dev->data->dev_private; const struct rte_flow_attr attr = { .egress = 1, .priority = 0, @@ -6094,8 +6927,8 @@ mlx5_ctrl_flow_source_queue(struct rte_eth_dev *dev, actions[0].type = RTE_FLOW_ACTION_TYPE_JUMP; actions[0].conf = &jump; actions[1].type = RTE_FLOW_ACTION_TYPE_END; - flow_idx = flow_list_create(dev, &priv->ctrl_flows, - &attr, items, actions, false, &error); + flow_idx = flow_list_create(dev, MLX5_FLOW_TYPE_CTL, + &attr, items, actions, false, &error); if (!flow_idx) { DRV_LOG(DEBUG, "Failed to create ctrl flow: rte_errno(%d)," @@ -6134,7 +6967,7 @@ mlx5_ctrl_flow_vlan(struct rte_eth_dev *dev, struct mlx5_priv *priv = dev->data->dev_private; const struct rte_flow_attr attr = { .ingress = 1, - .priority = MLX5_FLOW_PRIO_RSVD, + .priority = MLX5_FLOW_LOWEST_PRIO_INDICATOR, }; struct rte_flow_item items[] = { { @@ -6184,8 +7017,8 @@ mlx5_ctrl_flow_vlan(struct rte_eth_dev *dev, action_rss.types = 0; for (i = 0; i != priv->reta_idx_n; ++i) queue[i] = (*priv->reta_idx)[i]; - flow_idx = flow_list_create(dev, &priv->ctrl_flows, - &attr, items, actions, false, &error); + flow_idx = flow_list_create(dev, MLX5_FLOW_TYPE_CTL, + &attr, items, actions, false, &error); if (!flow_idx) return -rte_errno; return 0; @@ -6226,7 +7059,6 @@ mlx5_ctrl_flow(struct rte_eth_dev *dev, int mlx5_flow_lacp_miss(struct rte_eth_dev *dev) { - struct mlx5_priv *priv = dev->data->dev_private; /* * The LACP matching is done by only using ether type since using * a multicast dst mac causes kernel to give low priority to this flow. @@ -6260,8 +7092,9 @@ mlx5_flow_lacp_miss(struct rte_eth_dev *dev) }, }; struct rte_flow_error error; - uint32_t flow_idx = flow_list_create(dev, &priv->ctrl_flows, - &attr, items, actions, false, &error); + uint32_t flow_idx = flow_list_create(dev, MLX5_FLOW_TYPE_CTL, + &attr, items, actions, + false, &error); if (!flow_idx) return -rte_errno; @@ -6279,9 +7112,8 @@ mlx5_flow_destroy(struct rte_eth_dev *dev, struct rte_flow *flow, struct rte_flow_error *error __rte_unused) { - struct mlx5_priv *priv = dev->data->dev_private; - - flow_list_destroy(dev, &priv->flows, (uintptr_t)(void *)flow); + flow_list_destroy(dev, MLX5_FLOW_TYPE_GEN, + (uintptr_t)(void *)flow); return 0; } @@ -6295,9 +7127,7 @@ int mlx5_flow_flush(struct rte_eth_dev *dev, struct rte_flow_error *error __rte_unused) { - struct mlx5_priv *priv = dev->data->dev_private; - - mlx5_flow_list_flush(dev, &priv->flows, false); + mlx5_flow_list_flush(dev, MLX5_FLOW_TYPE_GEN, false); return 0; } @@ -6323,9 +7153,9 @@ mlx5_flow_isolate(struct rte_eth_dev *dev, } priv->isolated = !!enable; if (enable) - dev->dev_ops = &mlx5_os_dev_ops_isolate; + dev->dev_ops = &mlx5_dev_ops_isolate; else - dev->dev_ops = &mlx5_os_dev_ops; + dev->dev_ops = &mlx5_dev_ops; dev->rx_descriptor_status = mlx5_rx_descriptor_status; dev->tx_descriptor_status = mlx5_tx_descriptor_status; @@ -6348,8 +7178,7 @@ flow_drv_query(struct rte_eth_dev *dev, { struct mlx5_priv *priv = dev->data->dev_private; const struct mlx5_flow_driver_ops *fops; - struct rte_flow *flow = mlx5_ipool_get(priv->sh->ipool - [MLX5_IPOOL_RTE_FLOW], + struct rte_flow *flow = mlx5_ipool_get(priv->flows[MLX5_FLOW_TYPE_GEN], flow_idx); enum mlx5_flow_drv_type ftype; @@ -6389,62 +7218,60 @@ mlx5_flow_query(struct rte_eth_dev *dev, } /** - * Manage filter operations. + * Get rte_flow callbacks. * * @param dev * Pointer to Ethernet device structure. - * @param filter_type - * Filter type. - * @param filter_op - * Operation to perform. - * @param arg + * @param ops * Pointer to operation-specific structure. * - * @return - * 0 on success, a negative errno value otherwise and rte_errno is set. + * @return 0 */ int -mlx5_dev_filter_ctrl(struct rte_eth_dev *dev, - enum rte_filter_type filter_type, - enum rte_filter_op filter_op, - void *arg) -{ - switch (filter_type) { - case RTE_ETH_FILTER_GENERIC: - if (filter_op != RTE_ETH_FILTER_GET) { - rte_errno = EINVAL; - return -rte_errno; - } - *(const void **)arg = &mlx5_flow_ops; - return 0; - default: - DRV_LOG(ERR, "port %u filter type (%d) not supported", - dev->data->port_id, filter_type); - rte_errno = ENOTSUP; - return -rte_errno; - } +mlx5_flow_ops_get(struct rte_eth_dev *dev __rte_unused, + const struct rte_flow_ops **ops) +{ + *ops = &mlx5_flow_ops; return 0; } /** - * Create the needed meter and suffix tables. + * Validate meter policy actions. + * Dispatcher for action type specific validation. * * @param[in] dev - * Pointer to Ethernet device. - * @param[in] fm - * Pointer to the flow meter. + * Pointer to the Ethernet device structure. + * @param[in] action + * The meter policy action object to validate. + * @param[in] attr + * Attributes of flow to determine steering domain. + * @param[out] is_rss + * Is RSS or not. + * @param[out] domain_bitmap + * Domain bitmap. + * @param[out] is_def_policy + * Is default policy or not. + * @param[out] error + * Perform verbose error reporting if not NULL. Initialized in case of + * error only. * * @return - * Pointer to table set on success, NULL otherwise. + * 0 on success, otherwise negative errno value. */ -struct mlx5_meter_domains_infos * -mlx5_flow_create_mtr_tbls(struct rte_eth_dev *dev, - const struct mlx5_flow_meter *fm) +int +mlx5_flow_validate_mtr_acts(struct rte_eth_dev *dev, + const struct rte_flow_action *actions[RTE_COLORS], + struct rte_flow_attr *attr, + bool *is_rss, + uint8_t *domain_bitmap, + uint8_t *policy_mode, + struct rte_mtr_error *error) { const struct mlx5_flow_driver_ops *fops; fops = flow_get_drv_ops(MLX5_FLOW_TYPE_DV); - return fops->create_mtr_tbls(dev, fm); + return fops->validate_mtr_acts(dev, actions, attr, is_rss, + domain_bitmap, policy_mode, error); } /** @@ -6452,66 +7279,233 @@ mlx5_flow_create_mtr_tbls(struct rte_eth_dev *dev, * * @param[in] dev * Pointer to Ethernet device. - * @param[in] tbl - * Pointer to the meter table set. + * @param[in] mtr_policy + * Meter policy struct. + */ +void +mlx5_flow_destroy_mtr_acts(struct rte_eth_dev *dev, + struct mlx5_flow_meter_policy *mtr_policy) +{ + const struct mlx5_flow_driver_ops *fops; + + fops = flow_get_drv_ops(MLX5_FLOW_TYPE_DV); + fops->destroy_mtr_acts(dev, mtr_policy); +} + +/** + * Create policy action, lock free, + * (mutex should be acquired by caller). + * Dispatcher for action type specific call. + * + * @param[in] dev + * Pointer to the Ethernet device structure. + * @param[in] mtr_policy + * Meter policy struct. + * @param[in] action + * Action specification used to create meter actions. + * @param[out] error + * Perform verbose error reporting if not NULL. Initialized in case of + * error only. * * @return - * 0 on success. + * 0 on success, otherwise negative errno value. */ int -mlx5_flow_destroy_mtr_tbls(struct rte_eth_dev *dev, - struct mlx5_meter_domains_infos *tbls) +mlx5_flow_create_mtr_acts(struct rte_eth_dev *dev, + struct mlx5_flow_meter_policy *mtr_policy, + const struct rte_flow_action *actions[RTE_COLORS], + struct rte_mtr_error *error) +{ + const struct mlx5_flow_driver_ops *fops; + + fops = flow_get_drv_ops(MLX5_FLOW_TYPE_DV); + return fops->create_mtr_acts(dev, mtr_policy, actions, error); +} + +/** + * Create policy rules, lock free, + * (mutex should be acquired by caller). + * Dispatcher for action type specific call. + * + * @param[in] dev + * Pointer to the Ethernet device structure. + * @param[in] mtr_policy + * Meter policy struct. + * + * @return + * 0 on success, -1 otherwise. + */ +int +mlx5_flow_create_policy_rules(struct rte_eth_dev *dev, + struct mlx5_flow_meter_policy *mtr_policy) +{ + const struct mlx5_flow_driver_ops *fops; + + fops = flow_get_drv_ops(MLX5_FLOW_TYPE_DV); + return fops->create_policy_rules(dev, mtr_policy); +} + +/** + * Destroy policy rules, lock free, + * (mutex should be acquired by caller). + * Dispatcher for action type specific call. + * + * @param[in] dev + * Pointer to the Ethernet device structure. + * @param[in] mtr_policy + * Meter policy struct. + */ +void +mlx5_flow_destroy_policy_rules(struct rte_eth_dev *dev, + struct mlx5_flow_meter_policy *mtr_policy) { const struct mlx5_flow_driver_ops *fops; fops = flow_get_drv_ops(MLX5_FLOW_TYPE_DV); - return fops->destroy_mtr_tbls(dev, tbls); + fops->destroy_policy_rules(dev, mtr_policy); } /** - * Create policer rules. + * Destroy the default policy table set. + * + * @param[in] dev + * Pointer to Ethernet device. + */ +void +mlx5_flow_destroy_def_policy(struct rte_eth_dev *dev) +{ + const struct mlx5_flow_driver_ops *fops; + + fops = flow_get_drv_ops(MLX5_FLOW_TYPE_DV); + fops->destroy_def_policy(dev); +} + +/** + * Destroy the default policy table set. * * @param[in] dev * Pointer to Ethernet device. - * @param[in] fm - * Pointer to flow meter structure. - * @param[in] attr - * Pointer to flow attributes. * * @return * 0 on success, -1 otherwise. */ int -mlx5_flow_create_policer_rules(struct rte_eth_dev *dev, - struct mlx5_flow_meter *fm, - const struct rte_flow_attr *attr) +mlx5_flow_create_def_policy(struct rte_eth_dev *dev) { const struct mlx5_flow_driver_ops *fops; fops = flow_get_drv_ops(MLX5_FLOW_TYPE_DV); - return fops->create_policer_rules(dev, fm, attr); + return fops->create_def_policy(dev); } /** - * Destroy policer rules. + * Create the needed meter and suffix tables. * - * @param[in] fm - * Pointer to flow meter structure. - * @param[in] attr - * Pointer to flow attributes. + * @param[in] dev + * Pointer to Ethernet device. * * @return * 0 on success, -1 otherwise. */ int -mlx5_flow_destroy_policer_rules(struct rte_eth_dev *dev, - struct mlx5_flow_meter *fm, - const struct rte_flow_attr *attr) +mlx5_flow_create_mtr_tbls(struct rte_eth_dev *dev, + struct mlx5_flow_meter_info *fm, + uint32_t mtr_idx, + uint8_t domain_bitmap) +{ + const struct mlx5_flow_driver_ops *fops; + + fops = flow_get_drv_ops(MLX5_FLOW_TYPE_DV); + return fops->create_mtr_tbls(dev, fm, mtr_idx, domain_bitmap); +} + +/** + * Destroy the meter table set. + * + * @param[in] dev + * Pointer to Ethernet device. + * @param[in] tbl + * Pointer to the meter table set. + */ +void +mlx5_flow_destroy_mtr_tbls(struct rte_eth_dev *dev, + struct mlx5_flow_meter_info *fm) +{ + const struct mlx5_flow_driver_ops *fops; + + fops = flow_get_drv_ops(MLX5_FLOW_TYPE_DV); + fops->destroy_mtr_tbls(dev, fm); +} + +/** + * Destroy the global meter drop table. + * + * @param[in] dev + * Pointer to Ethernet device. + */ +void +mlx5_flow_destroy_mtr_drop_tbls(struct rte_eth_dev *dev) +{ + const struct mlx5_flow_driver_ops *fops; + + fops = flow_get_drv_ops(MLX5_FLOW_TYPE_DV); + fops->destroy_mtr_drop_tbls(dev); +} + +/** + * Destroy the sub policy table with RX queue. + * + * @param[in] dev + * Pointer to Ethernet device. + * @param[in] mtr_policy + * Pointer to meter policy table. + */ +void +mlx5_flow_destroy_sub_policy_with_rxq(struct rte_eth_dev *dev, + struct mlx5_flow_meter_policy *mtr_policy) +{ + const struct mlx5_flow_driver_ops *fops; + + fops = flow_get_drv_ops(MLX5_FLOW_TYPE_DV); + fops->destroy_sub_policy_with_rxq(dev, mtr_policy); +} + +/** + * Allocate the needed aso flow meter id. + * + * @param[in] dev + * Pointer to Ethernet device. + * + * @return + * Index to aso flow meter on success, NULL otherwise. + */ +uint32_t +mlx5_flow_mtr_alloc(struct rte_eth_dev *dev) +{ + const struct mlx5_flow_driver_ops *fops; + + fops = flow_get_drv_ops(MLX5_FLOW_TYPE_DV); + return fops->create_meter(dev); +} + +/** + * Free the aso flow meter id. + * + * @param[in] dev + * Pointer to Ethernet device. + * @param[in] mtr_idx + * Index to aso flow meter to be free. + * + * @return + * 0 on success. + */ +void +mlx5_flow_mtr_free(struct rte_eth_dev *dev, uint32_t mtr_idx) { const struct mlx5_flow_driver_ops *fops; fops = flow_get_drv_ops(MLX5_FLOW_TYPE_DV); - return fops->destroy_policer_rules(dev, fm, attr); + fops->free_meter(dev, mtr_idx); } /** @@ -6634,26 +7628,23 @@ mlx5_flow_create_counter_stat_mem_mng(struct mlx5_dev_ctx_shared *sh) } mem_mng = (struct mlx5_counter_stats_mem_mng *)(mem + size) - 1; size = sizeof(*raw_data) * MLX5_COUNTERS_PER_POOL * raws_n; - mem_mng->umem = mlx5_glue->devx_umem_reg(sh->ctx, mem, size, + mem_mng->umem = mlx5_os_umem_reg(sh->ctx, mem, size, IBV_ACCESS_LOCAL_WRITE); if (!mem_mng->umem) { rte_errno = errno; mlx5_free(mem); return -rte_errno; } + memset(&mkey_attr, 0, sizeof(mkey_attr)); mkey_attr.addr = (uintptr_t)mem; mkey_attr.size = size; mkey_attr.umem_id = mlx5_os_get_umem_id(mem_mng->umem); mkey_attr.pd = sh->pdn; - mkey_attr.log_entity_size = 0; - mkey_attr.pg_access = 0; - mkey_attr.klm_array = NULL; - mkey_attr.klm_num = 0; mkey_attr.relaxed_ordering_write = sh->cmng.relaxed_ordering_write; mkey_attr.relaxed_ordering_read = sh->cmng.relaxed_ordering_read; mem_mng->dm = mlx5_devx_cmd_mkey_create(sh->ctx, &mkey_attr); if (!mem_mng->dm) { - mlx5_glue->devx_umem_dereg(mem_mng->umem); + mlx5_os_umem_dereg(mem_mng->umem); rte_errno = errno; mlx5_free(mem); return -rte_errno; @@ -6900,104 +7891,13 @@ mlx5_flow_async_pool_query_handle(struct mlx5_dev_ctx_shared *sh, sh->cmng.pending_queries--; } -static const struct mlx5_flow_tbl_data_entry * -tunnel_mark_decode(struct rte_eth_dev *dev, uint32_t mark) -{ - struct mlx5_priv *priv = dev->data->dev_private; - struct mlx5_dev_ctx_shared *sh = priv->sh; - struct mlx5_hlist_entry *he; - union tunnel_offload_mark mbits = { .val = mark }; - union mlx5_flow_tbl_key table_key = { - { - .table_id = tunnel_id_to_flow_tbl(mbits.table_id), - .dummy = 0, - .domain = !!mbits.transfer, - .direction = 0, - } - }; - he = mlx5_hlist_lookup(sh->flow_tbls, table_key.v64, NULL); - return he ? - container_of(he, struct mlx5_flow_tbl_data_entry, entry) : NULL; -} - -static void -mlx5_flow_tunnel_grp2tbl_remove_cb(struct mlx5_hlist *list, - struct mlx5_hlist_entry *entry) -{ - struct mlx5_dev_ctx_shared *sh = list->ctx; - struct tunnel_tbl_entry *tte = container_of(entry, typeof(*tte), hash); - - mlx5_ipool_free(sh->ipool[MLX5_IPOOL_TNL_TBL_ID], - tunnel_flow_tbl_to_id(tte->flow_table)); - mlx5_free(tte); -} - -static struct mlx5_hlist_entry * -mlx5_flow_tunnel_grp2tbl_create_cb(struct mlx5_hlist *list, - uint64_t key __rte_unused, - void *ctx __rte_unused) -{ - struct mlx5_dev_ctx_shared *sh = list->ctx; - struct tunnel_tbl_entry *tte; - - tte = mlx5_malloc(MLX5_MEM_SYS | MLX5_MEM_ZERO, - sizeof(*tte), 0, - SOCKET_ID_ANY); - if (!tte) - goto err; - mlx5_ipool_malloc(sh->ipool[MLX5_IPOOL_TNL_TBL_ID], - &tte->flow_table); - if (tte->flow_table >= MLX5_MAX_TABLES) { - DRV_LOG(ERR, "Tunnel TBL ID %d exceed max limit.", - tte->flow_table); - mlx5_ipool_free(sh->ipool[MLX5_IPOOL_TNL_TBL_ID], - tte->flow_table); - goto err; - } else if (!tte->flow_table) { - goto err; - } - tte->flow_table = tunnel_id_to_flow_tbl(tte->flow_table); - return &tte->hash; -err: - if (tte) - mlx5_free(tte); - return NULL; -} - -static uint32_t -tunnel_flow_group_to_flow_table(struct rte_eth_dev *dev, - const struct mlx5_flow_tunnel *tunnel, - uint32_t group, uint32_t *table, - struct rte_flow_error *error) -{ - struct mlx5_hlist_entry *he; - struct tunnel_tbl_entry *tte; - union tunnel_tbl_key key = { - .tunnel_id = tunnel ? tunnel->tunnel_id : 0, - .group = group - }; - struct mlx5_flow_tunnel_hub *thub = mlx5_tunnel_hub(dev); - struct mlx5_hlist *group_hash; - - group_hash = tunnel ? tunnel->groups : thub->groups; - he = mlx5_hlist_register(group_hash, key.val, NULL); - if (!he) - return rte_flow_error_set(error, EINVAL, - RTE_FLOW_ERROR_TYPE_ATTR_GROUP, - NULL, - "tunnel group index not supported"); - tte = container_of(he, typeof(*tte), hash); - *table = tte->flow_table; - DRV_LOG(DEBUG, "port %u tunnel %u group=%#x table=%#x", - dev->data->port_id, key.tunnel_id, group, *table); - return 0; -} - static int flow_group_to_table(uint32_t port_id, uint32_t group, uint32_t *table, - struct flow_grp_info grp_info, struct rte_flow_error *error) + const struct flow_grp_info *grp_info, + struct rte_flow_error *error) { - if (grp_info.transfer && grp_info.external && grp_info.fdb_def_rule) { + if (grp_info->transfer && grp_info->external && + grp_info->fdb_def_rule) { if (group == UINT32_MAX) return rte_flow_error_set (error, EINVAL, @@ -7054,25 +7954,25 @@ int mlx5_flow_group_to_table(struct rte_eth_dev *dev, const struct mlx5_flow_tunnel *tunnel, uint32_t group, uint32_t *table, - struct flow_grp_info grp_info, + const struct flow_grp_info *grp_info, struct rte_flow_error *error) { int ret; bool standard_translation; - if (!grp_info.skip_scale && grp_info.external && + if (!grp_info->skip_scale && grp_info->external && group < MLX5_MAX_TABLES_EXTERNAL) group *= MLX5_FLOW_TABLE_FACTOR; if (is_tunnel_offload_active(dev)) { - standard_translation = !grp_info.external || - grp_info.std_tbl_fix; + standard_translation = !grp_info->external || + grp_info->std_tbl_fix; } else { standard_translation = true; } DRV_LOG(DEBUG, - "port %u group=%#x transfer=%d external=%d fdb_def_rule=%d translate=%s", - dev->data->port_id, group, grp_info.transfer, - grp_info.external, grp_info.fdb_def_rule, + "port %u group=%u transfer=%d external=%d fdb_def_rule=%d translate=%s", + dev->data->port_id, group, grp_info->transfer, + grp_info->external, grp_info->fdb_def_rule, standard_translation ? "STANDARD" : "TUNNEL"); if (standard_translation) ret = flow_group_to_table(dev->data->port_id, group, table, @@ -7110,7 +8010,7 @@ mlx5_flow_discover_mreg_c(struct rte_eth_dev *dev) for (idx = REG_C_2; idx <= REG_C_7; ++idx) { struct rte_flow_attr attr = { .group = MLX5_FLOW_MREG_CP_TABLE_GROUP, - .priority = MLX5_FLOW_PRIO_RSVD, + .priority = MLX5_FLOW_LOWEST_PRIO_INDICATOR, .ingress = 1, }; struct rte_flow_item items[] = { @@ -7144,49 +8044,239 @@ mlx5_flow_discover_mreg_c(struct rte_eth_dev *dev) if (!config->dv_flow_en) break; /* Create internal flow, validation skips copy action. */ - flow_idx = flow_list_create(dev, NULL, &attr, items, - actions, false, &error); - flow = mlx5_ipool_get(priv->sh->ipool[MLX5_IPOOL_RTE_FLOW], + flow_idx = flow_list_create(dev, MLX5_FLOW_TYPE_GEN, &attr, + items, actions, false, &error); + flow = mlx5_ipool_get(priv->flows[MLX5_FLOW_TYPE_GEN], flow_idx); if (!flow) continue; - if (dev->data->dev_started || !flow_drv_apply(dev, flow, NULL)) - config->flow_mreg_c[n++] = idx; - flow_list_destroy(dev, NULL, flow_idx); + config->flow_mreg_c[n++] = idx; + flow_list_destroy(dev, MLX5_FLOW_TYPE_GEN, flow_idx); } for (; n < MLX5_MREG_C_NUM; ++n) config->flow_mreg_c[n] = REG_NON; return 0; } -/** - * Dump flow raw hw data to file - * - * @param[in] dev - * The pointer to Ethernet device. - * @param[in] file - * A pointer to a file for output. - * @param[out] error - * Perform verbose error reporting if not NULL. PMDs initialize this - * structure in case of error only. - * @return - * 0 on success, a nagative value otherwise. - */ int -mlx5_flow_dev_dump(struct rte_eth_dev *dev, - FILE *file, - struct rte_flow_error *error __rte_unused) -{ - struct mlx5_priv *priv = dev->data->dev_private; - struct mlx5_dev_ctx_shared *sh = priv->sh; - - if (!priv->config.dv_flow_en) { - if (fputs("device dv flow disabled\n", file) <= 0) - return -errno; - return -ENOTSUP; +save_dump_file(const uint8_t *data, uint32_t size, + uint32_t type, uint32_t id, void *arg, FILE *file) +{ + char line[BUF_SIZE]; + uint32_t out = 0; + uint32_t k; + uint32_t actions_num; + struct rte_flow_query_count *count; + + memset(line, 0, BUF_SIZE); + switch (type) { + case DR_DUMP_REC_TYPE_PMD_MODIFY_HDR: + actions_num = *(uint32_t *)(arg); + out += snprintf(line + out, BUF_SIZE - out, "%d,0x%x,%d,", + type, id, actions_num); + break; + case DR_DUMP_REC_TYPE_PMD_PKT_REFORMAT: + out += snprintf(line + out, BUF_SIZE - out, "%d,0x%x,", + type, id); + break; + case DR_DUMP_REC_TYPE_PMD_COUNTER: + count = (struct rte_flow_query_count *)arg; + fprintf(file, "%d,0x%x,%" PRIu64 ",%" PRIu64 "\n", type, + id, count->hits, count->bytes); + return 0; + default: + return -1; + } + + for (k = 0; k < size; k++) { + /* Make sure we do not overrun the line buffer length. */ + if (out >= BUF_SIZE - 4) { + line[out] = '\0'; + break; + } + out += snprintf(line + out, BUF_SIZE - out, "%02x", + (data[k]) & 0xff); + } + fprintf(file, "%s\n", line); + return 0; +} + +int +mlx5_flow_query_counter(struct rte_eth_dev *dev, struct rte_flow *flow, + struct rte_flow_query_count *count, struct rte_flow_error *error) +{ + struct rte_flow_action action[2]; + enum mlx5_flow_drv_type ftype; + const struct mlx5_flow_driver_ops *fops; + + if (!flow) { + return rte_flow_error_set(error, ENOENT, + RTE_FLOW_ERROR_TYPE_UNSPECIFIED, + NULL, + "invalid flow handle"); + } + action[0].type = RTE_FLOW_ACTION_TYPE_COUNT; + action[1].type = RTE_FLOW_ACTION_TYPE_END; + if (flow->counter) { + memset(count, 0, sizeof(struct rte_flow_query_count)); + ftype = (enum mlx5_flow_drv_type)(flow->drv_type); + MLX5_ASSERT(ftype > MLX5_FLOW_TYPE_MIN && + ftype < MLX5_FLOW_TYPE_MAX); + fops = flow_get_drv_ops(ftype); + return fops->query(dev, flow, action, count, error); + } + return -1; +} + +#ifdef HAVE_IBV_FLOW_DV_SUPPORT +/** + * Dump flow ipool data to file + * + * @param[in] dev + * The pointer to Ethernet device. + * @param[in] file + * A pointer to a file for output. + * @param[out] error + * Perform verbose error reporting if not NULL. PMDs initialize this + * structure in case of error only. + * @return + * 0 on success, a negative value otherwise. + */ +int +mlx5_flow_dev_dump_ipool(struct rte_eth_dev *dev, + struct rte_flow *flow, FILE *file, + struct rte_flow_error *error) +{ + struct mlx5_priv *priv = dev->data->dev_private; + struct mlx5_flow_dv_modify_hdr_resource *modify_hdr; + struct mlx5_flow_dv_encap_decap_resource *encap_decap; + uint32_t handle_idx; + struct mlx5_flow_handle *dh; + struct rte_flow_query_count count; + uint32_t actions_num; + const uint8_t *data; + size_t size; + uint32_t id; + uint32_t type; + + if (!flow) { + return rte_flow_error_set(error, ENOENT, + RTE_FLOW_ERROR_TYPE_UNSPECIFIED, + NULL, + "invalid flow handle"); + } + handle_idx = flow->dev_handles; + while (handle_idx) { + dh = mlx5_ipool_get(priv->sh->ipool + [MLX5_IPOOL_MLX5_FLOW], handle_idx); + if (!dh) + continue; + handle_idx = dh->next.next; + id = (uint32_t)(uintptr_t)dh->drv_flow; + + /* query counter */ + type = DR_DUMP_REC_TYPE_PMD_COUNTER; + if (!mlx5_flow_query_counter(dev, flow, &count, error)) + save_dump_file(NULL, 0, type, + id, (void *)&count, file); + + /* Get modify_hdr and encap_decap buf from ipools. */ + encap_decap = NULL; + modify_hdr = dh->dvh.modify_hdr; + + if (dh->dvh.rix_encap_decap) { + encap_decap = mlx5_ipool_get(priv->sh->ipool + [MLX5_IPOOL_DECAP_ENCAP], + dh->dvh.rix_encap_decap); + } + if (modify_hdr) { + data = (const uint8_t *)modify_hdr->actions; + size = (size_t)(modify_hdr->actions_num) * 8; + actions_num = modify_hdr->actions_num; + type = DR_DUMP_REC_TYPE_PMD_MODIFY_HDR; + save_dump_file(data, size, type, id, + (void *)(&actions_num), file); + } + if (encap_decap) { + data = encap_decap->buf; + size = encap_decap->size; + type = DR_DUMP_REC_TYPE_PMD_PKT_REFORMAT; + save_dump_file(data, size, type, + id, NULL, file); + } + } + return 0; +} +#endif + +/** + * Dump flow raw hw data to file + * + * @param[in] dev + * The pointer to Ethernet device. + * @param[in] file + * A pointer to a file for output. + * @param[out] error + * Perform verbose error reporting if not NULL. PMDs initialize this + * structure in case of error only. + * @return + * 0 on success, a nagative value otherwise. + */ +int +mlx5_flow_dev_dump(struct rte_eth_dev *dev, struct rte_flow *flow_idx, + FILE *file, + struct rte_flow_error *error __rte_unused) +{ + struct mlx5_priv *priv = dev->data->dev_private; + struct mlx5_dev_ctx_shared *sh = priv->sh; + uint32_t handle_idx; + int ret; + struct mlx5_flow_handle *dh; + struct rte_flow *flow; +#ifdef HAVE_IBV_FLOW_DV_SUPPORT + uint32_t idx; +#endif + + if (!priv->config.dv_flow_en) { + if (fputs("device dv flow disabled\n", file) <= 0) + return -errno; + return -ENOTSUP; + } + + /* dump all */ + if (!flow_idx) { +#ifdef HAVE_IBV_FLOW_DV_SUPPORT + MLX5_IPOOL_FOREACH(priv->flows[MLX5_FLOW_TYPE_GEN], idx, flow) + mlx5_flow_dev_dump_ipool(dev, flow, file, error); +#endif + return mlx5_devx_cmd_flow_dump(sh->fdb_domain, + sh->rx_domain, + sh->tx_domain, file); } - return mlx5_devx_cmd_flow_dump(sh->fdb_domain, sh->rx_domain, - sh->tx_domain, file); + /* dump one */ + flow = mlx5_ipool_get(priv->flows[MLX5_FLOW_TYPE_GEN], + (uintptr_t)(void *)flow_idx); + if (!flow) + return -ENOENT; + +#ifdef HAVE_IBV_FLOW_DV_SUPPORT + mlx5_flow_dev_dump_ipool(dev, flow, file, error); +#endif + handle_idx = flow->dev_handles; + while (handle_idx) { + dh = mlx5_ipool_get(priv->sh->ipool[MLX5_IPOOL_MLX5_FLOW], + handle_idx); + if (!dh) + return -ENOENT; + if (dh->drv_flow) { + ret = mlx5_devx_cmd_flow_single_dump(dh->drv_flow, + file); + if (ret) + return -ENOENT; + } + handle_idx = dh->next.next; + } + return 0; } /** @@ -7229,12 +8319,12 @@ mlx5_flow_get_aged_flows(struct rte_eth_dev *dev, void **contexts, /* Wrapper for driver action_validate op callback */ static int flow_drv_action_validate(struct rte_eth_dev *dev, - const struct rte_flow_shared_action_conf *conf, + const struct rte_flow_indir_action_conf *conf, const struct rte_flow_action *action, const struct mlx5_flow_driver_ops *fops, struct rte_flow_error *error) { - static const char err_msg[] = "shared action validation unsupported"; + static const char err_msg[] = "indirect action validation unsupported"; if (!fops->action_validate) { DRV_LOG(ERR, "port %u %s.", dev->data->port_id, err_msg); @@ -7250,8 +8340,8 @@ flow_drv_action_validate(struct rte_eth_dev *dev, * * @param dev * Pointer to Ethernet device structure. - * @param[in] action - * Handle for the shared action to be destroyed. + * @param[in] handle + * Handle for the indirect action object to be destroyed. * @param[out] error * Perform verbose error reporting if not NULL. PMDs initialize this * structure in case of error only. @@ -7262,11 +8352,11 @@ flow_drv_action_validate(struct rte_eth_dev *dev, * @note: wrapper for driver action_create op callback. */ static int -mlx5_shared_action_destroy(struct rte_eth_dev *dev, - struct rte_flow_shared_action *action, +mlx5_action_handle_destroy(struct rte_eth_dev *dev, + struct rte_flow_action_handle *handle, struct rte_flow_error *error) { - static const char err_msg[] = "shared action destruction unsupported"; + static const char err_msg[] = "indirect action destruction unsupported"; struct rte_flow_attr attr = { .transfer = 0 }; const struct mlx5_flow_driver_ops *fops = flow_get_drv_ops(flow_get_drv_type(dev, &attr)); @@ -7277,18 +8367,18 @@ mlx5_shared_action_destroy(struct rte_eth_dev *dev, NULL, err_msg); return -rte_errno; } - return fops->action_destroy(dev, action, error); + return fops->action_destroy(dev, handle, error); } /* Wrapper for driver action_destroy op callback */ static int flow_drv_action_update(struct rte_eth_dev *dev, - struct rte_flow_shared_action *action, - const void *action_conf, + struct rte_flow_action_handle *handle, + const void *update, const struct mlx5_flow_driver_ops *fops, struct rte_flow_error *error) { - static const char err_msg[] = "shared action update unsupported"; + static const char err_msg[] = "indirect action update unsupported"; if (!fops->action_update) { DRV_LOG(ERR, "port %u %s.", dev->data->port_id, err_msg); @@ -7296,18 +8386,18 @@ flow_drv_action_update(struct rte_eth_dev *dev, NULL, err_msg); return -rte_errno; } - return fops->action_update(dev, action, action_conf, error); + return fops->action_update(dev, handle, update, error); } /* Wrapper for driver action_destroy op callback */ static int flow_drv_action_query(struct rte_eth_dev *dev, - const struct rte_flow_shared_action *action, + const struct rte_flow_action_handle *handle, void *data, const struct mlx5_flow_driver_ops *fops, struct rte_flow_error *error) { - static const char err_msg[] = "shared action query unsupported"; + static const char err_msg[] = "indirect action query unsupported"; if (!fops->action_query) { DRV_LOG(ERR, "port %u %s.", dev->data->port_id, err_msg); @@ -7315,29 +8405,31 @@ flow_drv_action_query(struct rte_eth_dev *dev, NULL, err_msg); return -rte_errno; } - return fops->action_query(dev, action, data, error); + return fops->action_query(dev, handle, data, error); } /** - * Create shared action for reuse in multiple flow rules. + * Create indirect action for reuse in multiple flow rules. * * @param dev * Pointer to Ethernet device structure. + * @param conf + * Pointer to indirect action object configuration. * @param[in] action - * Action configuration for shared action creation. + * Action configuration for indirect action object creation. * @param[out] error * Perform verbose error reporting if not NULL. PMDs initialize this * structure in case of error only. * @return * A valid handle in case of success, NULL otherwise and rte_errno is set. */ -static struct rte_flow_shared_action * -mlx5_shared_action_create(struct rte_eth_dev *dev, - const struct rte_flow_shared_action_conf *conf, +static struct rte_flow_action_handle * +mlx5_action_handle_create(struct rte_eth_dev *dev, + const struct rte_flow_indir_action_conf *conf, const struct rte_flow_action *action, struct rte_flow_error *error) { - static const char err_msg[] = "shared action creation unsupported"; + static const char err_msg[] = "indirect action creation unsupported"; struct rte_flow_attr attr = { .transfer = 0 }; const struct mlx5_flow_driver_ops *fops = flow_get_drv_ops(flow_get_drv_type(dev, &attr)); @@ -7354,19 +8446,20 @@ mlx5_shared_action_create(struct rte_eth_dev *dev, } /** - * Updates inplace the shared action configuration pointed by *action* handle - * with the configuration provided as *action* argument. - * The update of the shared action configuration effects all flow rules reusing - * the action via handle. + * Updates inplace the indirect action configuration pointed by *handle* + * with the configuration provided as *update* argument. + * The update of the indirect action configuration effects all flow rules + * reusing the action via handle. * * @param dev * Pointer to Ethernet device structure. - * @param[in] shared_action - * Handle for the shared action to be updated. - * @param[in] action + * @param[in] handle + * Handle for the indirect action to be updated. + * @param[in] update * Action specification used to modify the action pointed by handle. - * *action* should be of same type with the action pointed by the *action* - * handle argument, otherwise considered as invalid. + * *update* could be of same type with the action pointed by the *handle* + * handle argument, or some other structures like a wrapper, depending on + * the indirect action type. * @param[out] error * Perform verbose error reporting if not NULL. PMDs initialize this * structure in case of error only. @@ -7375,9 +8468,9 @@ mlx5_shared_action_create(struct rte_eth_dev *dev, * 0 on success, a negative errno value otherwise and rte_errno is set. */ static int -mlx5_shared_action_update(struct rte_eth_dev *dev, - struct rte_flow_shared_action *shared_action, - const struct rte_flow_action *action, +mlx5_action_handle_update(struct rte_eth_dev *dev, + struct rte_flow_action_handle *handle, + const void *update, struct rte_flow_error *error) { struct rte_flow_attr attr = { .transfer = 0 }; @@ -7385,26 +8478,27 @@ mlx5_shared_action_update(struct rte_eth_dev *dev, flow_get_drv_ops(flow_get_drv_type(dev, &attr)); int ret; - ret = flow_drv_action_validate(dev, NULL, action, fops, error); + ret = flow_drv_action_validate(dev, NULL, + (const struct rte_flow_action *)update, fops, error); if (ret) return ret; - return flow_drv_action_update(dev, shared_action, action->conf, fops, + return flow_drv_action_update(dev, handle, update, fops, error); } /** - * Query the shared action by handle. + * Query the indirect action by handle. * * This function allows retrieving action-specific data such as counters. * Data is gathered by special action which may be present/referenced in * more than one flow rule definition. * - * \see RTE_FLOW_ACTION_TYPE_COUNT + * see @RTE_FLOW_ACTION_TYPE_COUNT * * @param dev * Pointer to Ethernet device structure. - * @param[in] action - * Handle for the shared action to query. + * @param[in] handle + * Handle for the indirect action to query. * @param[in, out] data * Pointer to storage for the associated query data type. * @param[out] error @@ -7415,8 +8509,8 @@ mlx5_shared_action_update(struct rte_eth_dev *dev, * 0 on success, a negative errno value otherwise and rte_errno is set. */ static int -mlx5_shared_action_query(struct rte_eth_dev *dev, - const struct rte_flow_shared_action *action, +mlx5_action_handle_query(struct rte_eth_dev *dev, + const struct rte_flow_action_handle *handle, void *data, struct rte_flow_error *error) { @@ -7424,11 +8518,11 @@ mlx5_shared_action_query(struct rte_eth_dev *dev, const struct mlx5_flow_driver_ops *fops = flow_get_drv_ops(flow_get_drv_type(dev, &attr)); - return flow_drv_action_query(dev, action, data, fops, error); + return flow_drv_action_query(dev, handle, data, fops, error); } /** - * Destroy all shared actions. + * Destroy all indirect actions (shared RSS). * * @param dev * Pointer to Ethernet device. @@ -7437,188 +8531,22 @@ mlx5_shared_action_query(struct rte_eth_dev *dev, * 0 on success, a negative errno value otherwise and rte_errno is set. */ int -mlx5_shared_action_flush(struct rte_eth_dev *dev) +mlx5_action_handle_flush(struct rte_eth_dev *dev) { struct rte_flow_error error; struct mlx5_priv *priv = dev->data->dev_private; - struct mlx5_shared_action_rss *action; + struct mlx5_shared_action_rss *shared_rss; int ret = 0; uint32_t idx; ILIST_FOREACH(priv->sh->ipool[MLX5_IPOOL_RSS_SHARED_ACTIONS], - priv->rss_shared_actions, idx, action, next) { - ret |= mlx5_shared_action_destroy(dev, - (struct rte_flow_shared_action *)(uintptr_t)idx, &error); - } - return ret; -} - -static void -mlx5_flow_tunnel_free(struct rte_eth_dev *dev, - struct mlx5_flow_tunnel *tunnel) -{ - struct mlx5_priv *priv = dev->data->dev_private; - - DRV_LOG(DEBUG, "port %u release pmd tunnel id=0x%x", - dev->data->port_id, tunnel->tunnel_id); - RTE_VERIFY(!__atomic_load_n(&tunnel->refctn, __ATOMIC_RELAXED)); - mlx5_ipool_free(priv->sh->ipool[MLX5_IPOOL_TUNNEL_ID], - tunnel->tunnel_id); - mlx5_hlist_destroy(tunnel->groups); - mlx5_free(tunnel); -} - -static struct mlx5_flow_tunnel * -mlx5_find_tunnel_id(struct rte_eth_dev *dev, uint32_t id) -{ - struct mlx5_flow_tunnel_hub *thub = mlx5_tunnel_hub(dev); - struct mlx5_flow_tunnel *tun; - - LIST_FOREACH(tun, &thub->tunnels, chain) { - if (tun->tunnel_id == id) - break; - } - - return tun; -} - -static struct mlx5_flow_tunnel * -mlx5_flow_tunnel_allocate(struct rte_eth_dev *dev, - const struct rte_flow_tunnel *app_tunnel) -{ - struct mlx5_priv *priv = dev->data->dev_private; - struct mlx5_flow_tunnel *tunnel; - uint32_t id; - - mlx5_ipool_malloc(priv->sh->ipool[MLX5_IPOOL_RSS_EXPANTION_FLOW_ID], - &id); - if (id >= MLX5_MAX_TUNNELS) { - mlx5_ipool_free(priv->sh->ipool - [MLX5_IPOOL_RSS_EXPANTION_FLOW_ID], id); - DRV_LOG(ERR, "Tunnel ID %d exceed max limit.", id); - return NULL; - } else if (!id) { - return NULL; - } - /** - * mlx5 flow tunnel is an auxlilary data structure - * It's not part of IO. No need to allocate it from - * huge pages pools dedicated for IO - */ - tunnel = mlx5_malloc(MLX5_MEM_SYS | MLX5_MEM_ZERO, sizeof(*tunnel), - 0, SOCKET_ID_ANY); - if (!tunnel) { - mlx5_ipool_free(priv->sh->ipool - [MLX5_IPOOL_RSS_EXPANTION_FLOW_ID], id); - return NULL; - } - tunnel->groups = mlx5_hlist_create("tunnel groups", 1024, 0, 0, - mlx5_flow_tunnel_grp2tbl_create_cb, - NULL, - mlx5_flow_tunnel_grp2tbl_remove_cb); - if (!tunnel->groups) { - mlx5_ipool_free(priv->sh->ipool - [MLX5_IPOOL_RSS_EXPANTION_FLOW_ID], id); - mlx5_free(tunnel); - return NULL; - } - tunnel->groups->ctx = priv->sh; - /* initiate new PMD tunnel */ - memcpy(&tunnel->app_tunnel, app_tunnel, sizeof(*app_tunnel)); - tunnel->tunnel_id = id; - tunnel->action.type = (typeof(tunnel->action.type)) - MLX5_RTE_FLOW_ACTION_TYPE_TUNNEL_SET; - tunnel->action.conf = tunnel; - tunnel->item.type = (typeof(tunnel->item.type)) - MLX5_RTE_FLOW_ITEM_TYPE_TUNNEL; - tunnel->item.spec = tunnel; - tunnel->item.last = NULL; - tunnel->item.mask = NULL; - - DRV_LOG(DEBUG, "port %u new pmd tunnel id=0x%x", - dev->data->port_id, tunnel->tunnel_id); - - return tunnel; -} - -static int -mlx5_get_flow_tunnel(struct rte_eth_dev *dev, - const struct rte_flow_tunnel *app_tunnel, - struct mlx5_flow_tunnel **tunnel) -{ - int ret; - struct mlx5_flow_tunnel_hub *thub = mlx5_tunnel_hub(dev); - struct mlx5_flow_tunnel *tun; - - rte_spinlock_lock(&thub->sl); - LIST_FOREACH(tun, &thub->tunnels, chain) { - if (!memcmp(app_tunnel, &tun->app_tunnel, - sizeof(*app_tunnel))) { - *tunnel = tun; - ret = 0; - break; - } - } - if (!tun) { - tun = mlx5_flow_tunnel_allocate(dev, app_tunnel); - if (tun) { - LIST_INSERT_HEAD(&thub->tunnels, tun, chain); - *tunnel = tun; - } else { - ret = -ENOMEM; - } + priv->rss_shared_actions, idx, shared_rss, next) { + ret |= mlx5_action_handle_destroy(dev, + (struct rte_flow_action_handle *)(uintptr_t)idx, &error); } - rte_spinlock_unlock(&thub->sl); - if (tun) - __atomic_add_fetch(&tun->refctn, 1, __ATOMIC_RELAXED); - return ret; } -void mlx5_release_tunnel_hub(struct mlx5_dev_ctx_shared *sh, uint16_t port_id) -{ - struct mlx5_flow_tunnel_hub *thub = sh->tunnel_hub; - - if (!thub) - return; - if (!LIST_EMPTY(&thub->tunnels)) - DRV_LOG(WARNING, "port %u tunnels present\n", port_id); - mlx5_hlist_destroy(thub->groups); - mlx5_free(thub); -} - -int mlx5_alloc_tunnel_hub(struct mlx5_dev_ctx_shared *sh) -{ - int err; - struct mlx5_flow_tunnel_hub *thub; - - thub = mlx5_malloc(MLX5_MEM_SYS | MLX5_MEM_ZERO, sizeof(*thub), - 0, SOCKET_ID_ANY); - if (!thub) - return -ENOMEM; - LIST_INIT(&thub->tunnels); - rte_spinlock_init(&thub->sl); - thub->groups = mlx5_hlist_create("flow groups", MLX5_MAX_TABLES, 0, - 0, mlx5_flow_tunnel_grp2tbl_create_cb, - NULL, - mlx5_flow_tunnel_grp2tbl_remove_cb); - if (!thub->groups) { - err = -rte_errno; - goto err; - } - thub->groups->ctx = sh; - sh->tunnel_hub = thub; - - return 0; - -err: - if (thub->groups) - mlx5_hlist_destroy(thub->groups); - if (thub) - mlx5_free(thub); - return err; -} - #ifndef HAVE_MLX5DV_DR #define MLX5_DOMAIN_SYNC_FLOW ((1 << 0) | (1 << 1)) #else @@ -7639,3 +8567,852 @@ int rte_pmd_mlx5_sync_flow(uint16_t port_id, uint32_t domains) ret = -ret; return ret; } + +const struct mlx5_flow_tunnel * +mlx5_get_tof(const struct rte_flow_item *item, + const struct rte_flow_action *action, + enum mlx5_tof_rule_type *rule_type) +{ + for (; item->type != RTE_FLOW_ITEM_TYPE_END; item++) { + if (item->type == (typeof(item->type)) + MLX5_RTE_FLOW_ITEM_TYPE_TUNNEL) { + *rule_type = MLX5_TUNNEL_OFFLOAD_MATCH_RULE; + return flow_items_to_tunnel(item); + } + } + for (; action->conf != RTE_FLOW_ACTION_TYPE_END; action++) { + if (action->type == (typeof(action->type)) + MLX5_RTE_FLOW_ACTION_TYPE_TUNNEL_SET) { + *rule_type = MLX5_TUNNEL_OFFLOAD_SET_RULE; + return flow_actions_to_tunnel(action); + } + } + return NULL; +} + +/** + * tunnel offload functionalilty is defined for DV environment only + */ +#ifdef HAVE_IBV_FLOW_DV_SUPPORT +__extension__ +union tunnel_offload_mark { + uint32_t val; + struct { + uint32_t app_reserve:8; + uint32_t table_id:15; + uint32_t transfer:1; + uint32_t _unused_:8; + }; +}; + +static bool +mlx5_access_tunnel_offload_db + (struct rte_eth_dev *dev, + bool (*match)(struct rte_eth_dev *, + struct mlx5_flow_tunnel *, const void *), + void (*hit)(struct rte_eth_dev *, struct mlx5_flow_tunnel *, void *), + void (*miss)(struct rte_eth_dev *, void *), + void *ctx, bool lock_op); + +static int +flow_tunnel_add_default_miss(struct rte_eth_dev *dev, + struct rte_flow *flow, + const struct rte_flow_attr *attr, + const struct rte_flow_action *app_actions, + uint32_t flow_idx, + const struct mlx5_flow_tunnel *tunnel, + struct tunnel_default_miss_ctx *ctx, + struct rte_flow_error *error) +{ + struct mlx5_priv *priv = dev->data->dev_private; + struct mlx5_flow *dev_flow; + struct rte_flow_attr miss_attr = *attr; + const struct rte_flow_item miss_items[2] = { + { + .type = RTE_FLOW_ITEM_TYPE_ETH, + .spec = NULL, + .last = NULL, + .mask = NULL + }, + { + .type = RTE_FLOW_ITEM_TYPE_END, + .spec = NULL, + .last = NULL, + .mask = NULL + } + }; + union tunnel_offload_mark mark_id; + struct rte_flow_action_mark miss_mark; + struct rte_flow_action miss_actions[3] = { + [0] = { .type = RTE_FLOW_ACTION_TYPE_MARK, .conf = &miss_mark }, + [2] = { .type = RTE_FLOW_ACTION_TYPE_END, .conf = NULL } + }; + const struct rte_flow_action_jump *jump_data; + uint32_t i, flow_table = 0; /* prevent compilation warning */ + struct flow_grp_info grp_info = { + .external = 1, + .transfer = attr->transfer, + .fdb_def_rule = !!priv->fdb_def_rule, + .std_tbl_fix = 0, + }; + int ret; + + if (!attr->transfer) { + uint32_t q_size; + + miss_actions[1].type = RTE_FLOW_ACTION_TYPE_RSS; + q_size = priv->reta_idx_n * sizeof(ctx->queue[0]); + ctx->queue = mlx5_malloc(MLX5_MEM_SYS | MLX5_MEM_ZERO, q_size, + 0, SOCKET_ID_ANY); + if (!ctx->queue) + return rte_flow_error_set + (error, ENOMEM, + RTE_FLOW_ERROR_TYPE_ACTION_CONF, + NULL, "invalid default miss RSS"); + ctx->action_rss.func = RTE_ETH_HASH_FUNCTION_DEFAULT, + ctx->action_rss.level = 0, + ctx->action_rss.types = priv->rss_conf.rss_hf, + ctx->action_rss.key_len = priv->rss_conf.rss_key_len, + ctx->action_rss.queue_num = priv->reta_idx_n, + ctx->action_rss.key = priv->rss_conf.rss_key, + ctx->action_rss.queue = ctx->queue; + if (!priv->reta_idx_n || !priv->rxqs_n) + return rte_flow_error_set + (error, EINVAL, + RTE_FLOW_ERROR_TYPE_ACTION_CONF, + NULL, "invalid port configuration"); + if (!(dev->data->dev_conf.rxmode.mq_mode & ETH_MQ_RX_RSS_FLAG)) + ctx->action_rss.types = 0; + for (i = 0; i != priv->reta_idx_n; ++i) + ctx->queue[i] = (*priv->reta_idx)[i]; + } else { + miss_actions[1].type = RTE_FLOW_ACTION_TYPE_JUMP; + ctx->miss_jump.group = MLX5_TNL_MISS_FDB_JUMP_GRP; + } + miss_actions[1].conf = (typeof(miss_actions[1].conf))ctx->raw; + for (; app_actions->type != RTE_FLOW_ACTION_TYPE_JUMP; app_actions++); + jump_data = app_actions->conf; + miss_attr.priority = MLX5_TNL_MISS_RULE_PRIORITY; + miss_attr.group = jump_data->group; + ret = mlx5_flow_group_to_table(dev, tunnel, jump_data->group, + &flow_table, &grp_info, error); + if (ret) + return rte_flow_error_set(error, EINVAL, + RTE_FLOW_ERROR_TYPE_ACTION_CONF, + NULL, "invalid tunnel id"); + mark_id.app_reserve = 0; + mark_id.table_id = tunnel_flow_tbl_to_id(flow_table); + mark_id.transfer = !!attr->transfer; + mark_id._unused_ = 0; + miss_mark.id = mark_id.val; + dev_flow = flow_drv_prepare(dev, flow, &miss_attr, + miss_items, miss_actions, flow_idx, error); + if (!dev_flow) + return -rte_errno; + dev_flow->flow = flow; + dev_flow->external = true; + dev_flow->tunnel = tunnel; + dev_flow->tof_type = MLX5_TUNNEL_OFFLOAD_MISS_RULE; + /* Subflow object was created, we must include one in the list. */ + SILIST_INSERT(&flow->dev_handles, dev_flow->handle_idx, + dev_flow->handle, next); + DRV_LOG(DEBUG, + "port %u tunnel type=%d id=%u miss rule priority=%u group=%u", + dev->data->port_id, tunnel->app_tunnel.type, + tunnel->tunnel_id, miss_attr.priority, miss_attr.group); + ret = flow_drv_translate(dev, dev_flow, &miss_attr, miss_items, + miss_actions, error); + if (!ret) + ret = flow_mreg_update_copy_table(dev, flow, miss_actions, + error); + + return ret; +} + +static const struct mlx5_flow_tbl_data_entry * +tunnel_mark_decode(struct rte_eth_dev *dev, uint32_t mark) +{ + struct mlx5_priv *priv = dev->data->dev_private; + struct mlx5_dev_ctx_shared *sh = priv->sh; + struct mlx5_list_entry *he; + union tunnel_offload_mark mbits = { .val = mark }; + union mlx5_flow_tbl_key table_key = { + { + .level = tunnel_id_to_flow_tbl(mbits.table_id), + .id = 0, + .reserved = 0, + .dummy = 0, + .is_fdb = !!mbits.transfer, + .is_egress = 0, + } + }; + struct mlx5_flow_cb_ctx ctx = { + .data = &table_key.v64, + }; + + he = mlx5_hlist_lookup(sh->flow_tbls, table_key.v64, &ctx); + return he ? + container_of(he, struct mlx5_flow_tbl_data_entry, entry) : NULL; +} + +static void +mlx5_flow_tunnel_grp2tbl_remove_cb(void *tool_ctx, + struct mlx5_list_entry *entry) +{ + struct mlx5_dev_ctx_shared *sh = tool_ctx; + struct tunnel_tbl_entry *tte = container_of(entry, typeof(*tte), hash); + + mlx5_ipool_free(sh->ipool[MLX5_IPOOL_TNL_TBL_ID], + tunnel_flow_tbl_to_id(tte->flow_table)); + mlx5_free(tte); +} + +static int +mlx5_flow_tunnel_grp2tbl_match_cb(void *tool_ctx __rte_unused, + struct mlx5_list_entry *entry, void *cb_ctx) +{ + struct mlx5_flow_cb_ctx *ctx = cb_ctx; + union tunnel_tbl_key tbl = { + .val = *(uint64_t *)(ctx->data), + }; + struct tunnel_tbl_entry *tte = container_of(entry, typeof(*tte), hash); + + return tbl.tunnel_id != tte->tunnel_id || tbl.group != tte->group; +} + +static struct mlx5_list_entry * +mlx5_flow_tunnel_grp2tbl_create_cb(void *tool_ctx, void *cb_ctx) +{ + struct mlx5_dev_ctx_shared *sh = tool_ctx; + struct mlx5_flow_cb_ctx *ctx = cb_ctx; + struct tunnel_tbl_entry *tte; + union tunnel_tbl_key tbl = { + .val = *(uint64_t *)(ctx->data), + }; + + tte = mlx5_malloc(MLX5_MEM_SYS | MLX5_MEM_ZERO, + sizeof(*tte), 0, + SOCKET_ID_ANY); + if (!tte) + goto err; + mlx5_ipool_malloc(sh->ipool[MLX5_IPOOL_TNL_TBL_ID], + &tte->flow_table); + if (tte->flow_table >= MLX5_MAX_TABLES) { + DRV_LOG(ERR, "Tunnel TBL ID %d exceed max limit.", + tte->flow_table); + mlx5_ipool_free(sh->ipool[MLX5_IPOOL_TNL_TBL_ID], + tte->flow_table); + goto err; + } else if (!tte->flow_table) { + goto err; + } + tte->flow_table = tunnel_id_to_flow_tbl(tte->flow_table); + tte->tunnel_id = tbl.tunnel_id; + tte->group = tbl.group; + return &tte->hash; +err: + if (tte) + mlx5_free(tte); + return NULL; +} + +static struct mlx5_list_entry * +mlx5_flow_tunnel_grp2tbl_clone_cb(void *tool_ctx __rte_unused, + struct mlx5_list_entry *oentry, + void *cb_ctx __rte_unused) +{ + struct tunnel_tbl_entry *tte = mlx5_malloc(MLX5_MEM_SYS, sizeof(*tte), + 0, SOCKET_ID_ANY); + + if (!tte) + return NULL; + memcpy(tte, oentry, sizeof(*tte)); + return &tte->hash; +} + +static void +mlx5_flow_tunnel_grp2tbl_clone_free_cb(void *tool_ctx __rte_unused, + struct mlx5_list_entry *entry) +{ + struct tunnel_tbl_entry *tte = container_of(entry, typeof(*tte), hash); + + mlx5_free(tte); +} + +static uint32_t +tunnel_flow_group_to_flow_table(struct rte_eth_dev *dev, + const struct mlx5_flow_tunnel *tunnel, + uint32_t group, uint32_t *table, + struct rte_flow_error *error) +{ + struct mlx5_list_entry *he; + struct tunnel_tbl_entry *tte; + union tunnel_tbl_key key = { + .tunnel_id = tunnel ? tunnel->tunnel_id : 0, + .group = group + }; + struct mlx5_flow_tunnel_hub *thub = mlx5_tunnel_hub(dev); + struct mlx5_hlist *group_hash; + struct mlx5_flow_cb_ctx ctx = { + .data = &key.val, + }; + + group_hash = tunnel ? tunnel->groups : thub->groups; + he = mlx5_hlist_register(group_hash, key.val, &ctx); + if (!he) + return rte_flow_error_set(error, EINVAL, + RTE_FLOW_ERROR_TYPE_ATTR_GROUP, + NULL, + "tunnel group index not supported"); + tte = container_of(he, typeof(*tte), hash); + *table = tte->flow_table; + DRV_LOG(DEBUG, "port %u tunnel %u group=%#x table=%#x", + dev->data->port_id, key.tunnel_id, group, *table); + return 0; +} + +static void +mlx5_flow_tunnel_free(struct rte_eth_dev *dev, + struct mlx5_flow_tunnel *tunnel) +{ + struct mlx5_priv *priv = dev->data->dev_private; + struct mlx5_indexed_pool *ipool; + + DRV_LOG(DEBUG, "port %u release pmd tunnel id=0x%x", + dev->data->port_id, tunnel->tunnel_id); + LIST_REMOVE(tunnel, chain); + mlx5_hlist_destroy(tunnel->groups); + ipool = priv->sh->ipool[MLX5_IPOOL_TUNNEL_ID]; + mlx5_ipool_free(ipool, tunnel->tunnel_id); +} + +static bool +mlx5_access_tunnel_offload_db + (struct rte_eth_dev *dev, + bool (*match)(struct rte_eth_dev *, + struct mlx5_flow_tunnel *, const void *), + void (*hit)(struct rte_eth_dev *, struct mlx5_flow_tunnel *, void *), + void (*miss)(struct rte_eth_dev *, void *), + void *ctx, bool lock_op) +{ + bool verdict = false; + struct mlx5_flow_tunnel_hub *thub = mlx5_tunnel_hub(dev); + struct mlx5_flow_tunnel *tunnel; + + rte_spinlock_lock(&thub->sl); + LIST_FOREACH(tunnel, &thub->tunnels, chain) { + verdict = match(dev, tunnel, (const void *)ctx); + if (verdict) + break; + } + if (!lock_op) + rte_spinlock_unlock(&thub->sl); + if (verdict && hit) + hit(dev, tunnel, ctx); + if (!verdict && miss) + miss(dev, ctx); + if (lock_op) + rte_spinlock_unlock(&thub->sl); + + return verdict; +} + +struct tunnel_db_find_tunnel_id_ctx { + uint32_t tunnel_id; + struct mlx5_flow_tunnel *tunnel; +}; + +static bool +find_tunnel_id_match(struct rte_eth_dev *dev, + struct mlx5_flow_tunnel *tunnel, const void *x) +{ + const struct tunnel_db_find_tunnel_id_ctx *ctx = x; + + RTE_SET_USED(dev); + return tunnel->tunnel_id == ctx->tunnel_id; +} + +static void +find_tunnel_id_hit(struct rte_eth_dev *dev, + struct mlx5_flow_tunnel *tunnel, void *x) +{ + struct tunnel_db_find_tunnel_id_ctx *ctx = x; + RTE_SET_USED(dev); + ctx->tunnel = tunnel; +} + +static struct mlx5_flow_tunnel * +mlx5_find_tunnel_id(struct rte_eth_dev *dev, uint32_t id) +{ + struct tunnel_db_find_tunnel_id_ctx ctx = { + .tunnel_id = id, + }; + + mlx5_access_tunnel_offload_db(dev, find_tunnel_id_match, + find_tunnel_id_hit, NULL, &ctx, true); + + return ctx.tunnel; +} + +static struct mlx5_flow_tunnel * +mlx5_flow_tunnel_allocate(struct rte_eth_dev *dev, + const struct rte_flow_tunnel *app_tunnel) +{ + struct mlx5_priv *priv = dev->data->dev_private; + struct mlx5_indexed_pool *ipool; + struct mlx5_flow_tunnel *tunnel; + uint32_t id; + + ipool = priv->sh->ipool[MLX5_IPOOL_TUNNEL_ID]; + tunnel = mlx5_ipool_zmalloc(ipool, &id); + if (!tunnel) + return NULL; + if (id >= MLX5_MAX_TUNNELS) { + mlx5_ipool_free(ipool, id); + DRV_LOG(ERR, "Tunnel ID %d exceed max limit.", id); + return NULL; + } + tunnel->groups = mlx5_hlist_create("tunnel groups", 64, false, true, + priv->sh, + mlx5_flow_tunnel_grp2tbl_create_cb, + mlx5_flow_tunnel_grp2tbl_match_cb, + mlx5_flow_tunnel_grp2tbl_remove_cb, + mlx5_flow_tunnel_grp2tbl_clone_cb, + mlx5_flow_tunnel_grp2tbl_clone_free_cb); + if (!tunnel->groups) { + mlx5_ipool_free(ipool, id); + return NULL; + } + /* initiate new PMD tunnel */ + memcpy(&tunnel->app_tunnel, app_tunnel, sizeof(*app_tunnel)); + tunnel->tunnel_id = id; + tunnel->action.type = (typeof(tunnel->action.type)) + MLX5_RTE_FLOW_ACTION_TYPE_TUNNEL_SET; + tunnel->action.conf = tunnel; + tunnel->item.type = (typeof(tunnel->item.type)) + MLX5_RTE_FLOW_ITEM_TYPE_TUNNEL; + tunnel->item.spec = tunnel; + tunnel->item.last = NULL; + tunnel->item.mask = NULL; + + DRV_LOG(DEBUG, "port %u new pmd tunnel id=0x%x", + dev->data->port_id, tunnel->tunnel_id); + + return tunnel; +} + +struct tunnel_db_get_tunnel_ctx { + const struct rte_flow_tunnel *app_tunnel; + struct mlx5_flow_tunnel *tunnel; +}; + +static bool get_tunnel_match(struct rte_eth_dev *dev, + struct mlx5_flow_tunnel *tunnel, const void *x) +{ + const struct tunnel_db_get_tunnel_ctx *ctx = x; + + RTE_SET_USED(dev); + return !memcmp(ctx->app_tunnel, &tunnel->app_tunnel, + sizeof(*ctx->app_tunnel)); +} + +static void get_tunnel_hit(struct rte_eth_dev *dev, + struct mlx5_flow_tunnel *tunnel, void *x) +{ + /* called under tunnel spinlock protection */ + struct tunnel_db_get_tunnel_ctx *ctx = x; + + RTE_SET_USED(dev); + tunnel->refctn++; + ctx->tunnel = tunnel; +} + +static void get_tunnel_miss(struct rte_eth_dev *dev, void *x) +{ + /* called under tunnel spinlock protection */ + struct mlx5_flow_tunnel_hub *thub = mlx5_tunnel_hub(dev); + struct tunnel_db_get_tunnel_ctx *ctx = x; + + rte_spinlock_unlock(&thub->sl); + ctx->tunnel = mlx5_flow_tunnel_allocate(dev, ctx->app_tunnel); + rte_spinlock_lock(&thub->sl); + if (ctx->tunnel) { + ctx->tunnel->refctn = 1; + LIST_INSERT_HEAD(&thub->tunnels, ctx->tunnel, chain); + } +} + + +static int +mlx5_get_flow_tunnel(struct rte_eth_dev *dev, + const struct rte_flow_tunnel *app_tunnel, + struct mlx5_flow_tunnel **tunnel) +{ + struct tunnel_db_get_tunnel_ctx ctx = { + .app_tunnel = app_tunnel, + }; + + mlx5_access_tunnel_offload_db(dev, get_tunnel_match, get_tunnel_hit, + get_tunnel_miss, &ctx, true); + *tunnel = ctx.tunnel; + return ctx.tunnel ? 0 : -ENOMEM; +} + +void mlx5_release_tunnel_hub(struct mlx5_dev_ctx_shared *sh, uint16_t port_id) +{ + struct mlx5_flow_tunnel_hub *thub = sh->tunnel_hub; + + if (!thub) + return; + if (!LIST_EMPTY(&thub->tunnels)) + DRV_LOG(WARNING, "port %u tunnels present", port_id); + mlx5_hlist_destroy(thub->groups); + mlx5_free(thub); +} + +int mlx5_alloc_tunnel_hub(struct mlx5_dev_ctx_shared *sh) +{ + int err; + struct mlx5_flow_tunnel_hub *thub; + + thub = mlx5_malloc(MLX5_MEM_SYS | MLX5_MEM_ZERO, sizeof(*thub), + 0, SOCKET_ID_ANY); + if (!thub) + return -ENOMEM; + LIST_INIT(&thub->tunnels); + rte_spinlock_init(&thub->sl); + thub->groups = mlx5_hlist_create("flow groups", 64, + false, true, sh, + mlx5_flow_tunnel_grp2tbl_create_cb, + mlx5_flow_tunnel_grp2tbl_match_cb, + mlx5_flow_tunnel_grp2tbl_remove_cb, + mlx5_flow_tunnel_grp2tbl_clone_cb, + mlx5_flow_tunnel_grp2tbl_clone_free_cb); + if (!thub->groups) { + err = -rte_errno; + goto err; + } + sh->tunnel_hub = thub; + + return 0; + +err: + if (thub->groups) + mlx5_hlist_destroy(thub->groups); + if (thub) + mlx5_free(thub); + return err; +} + +static inline bool +mlx5_flow_tunnel_validate(struct rte_eth_dev *dev, + struct rte_flow_tunnel *tunnel, + const char *err_msg) +{ + err_msg = NULL; + if (!is_tunnel_offload_active(dev)) { + err_msg = "tunnel offload was not activated"; + goto out; + } else if (!tunnel) { + err_msg = "no application tunnel"; + goto out; + } + + switch (tunnel->type) { + default: + err_msg = "unsupported tunnel type"; + goto out; + case RTE_FLOW_ITEM_TYPE_VXLAN: + break; + } + +out: + return !err_msg; +} + +static int +mlx5_flow_tunnel_decap_set(struct rte_eth_dev *dev, + struct rte_flow_tunnel *app_tunnel, + struct rte_flow_action **actions, + uint32_t *num_of_actions, + struct rte_flow_error *error) +{ + int ret; + struct mlx5_flow_tunnel *tunnel; + const char *err_msg = NULL; + bool verdict = mlx5_flow_tunnel_validate(dev, app_tunnel, err_msg); + + if (!verdict) + return rte_flow_error_set(error, EINVAL, + RTE_FLOW_ERROR_TYPE_ACTION_CONF, NULL, + err_msg); + ret = mlx5_get_flow_tunnel(dev, app_tunnel, &tunnel); + if (ret < 0) { + return rte_flow_error_set(error, ret, + RTE_FLOW_ERROR_TYPE_ACTION_CONF, NULL, + "failed to initialize pmd tunnel"); + } + *actions = &tunnel->action; + *num_of_actions = 1; + return 0; +} + +static int +mlx5_flow_tunnel_match(struct rte_eth_dev *dev, + struct rte_flow_tunnel *app_tunnel, + struct rte_flow_item **items, + uint32_t *num_of_items, + struct rte_flow_error *error) +{ + int ret; + struct mlx5_flow_tunnel *tunnel; + const char *err_msg = NULL; + bool verdict = mlx5_flow_tunnel_validate(dev, app_tunnel, err_msg); + + if (!verdict) + return rte_flow_error_set(error, EINVAL, + RTE_FLOW_ERROR_TYPE_HANDLE, NULL, + err_msg); + ret = mlx5_get_flow_tunnel(dev, app_tunnel, &tunnel); + if (ret < 0) { + return rte_flow_error_set(error, ret, + RTE_FLOW_ERROR_TYPE_HANDLE, NULL, + "failed to initialize pmd tunnel"); + } + *items = &tunnel->item; + *num_of_items = 1; + return 0; +} + +struct tunnel_db_element_release_ctx { + struct rte_flow_item *items; + struct rte_flow_action *actions; + uint32_t num_elements; + struct rte_flow_error *error; + int ret; +}; + +static bool +tunnel_element_release_match(struct rte_eth_dev *dev, + struct mlx5_flow_tunnel *tunnel, const void *x) +{ + const struct tunnel_db_element_release_ctx *ctx = x; + + RTE_SET_USED(dev); + if (ctx->num_elements != 1) + return false; + else if (ctx->items) + return ctx->items == &tunnel->item; + else if (ctx->actions) + return ctx->actions == &tunnel->action; + + return false; +} + +static void +tunnel_element_release_hit(struct rte_eth_dev *dev, + struct mlx5_flow_tunnel *tunnel, void *x) +{ + struct tunnel_db_element_release_ctx *ctx = x; + ctx->ret = 0; + if (!__atomic_sub_fetch(&tunnel->refctn, 1, __ATOMIC_RELAXED)) + mlx5_flow_tunnel_free(dev, tunnel); +} + +static void +tunnel_element_release_miss(struct rte_eth_dev *dev, void *x) +{ + struct tunnel_db_element_release_ctx *ctx = x; + RTE_SET_USED(dev); + ctx->ret = rte_flow_error_set(ctx->error, EINVAL, + RTE_FLOW_ERROR_TYPE_HANDLE, NULL, + "invalid argument"); +} + +static int +mlx5_flow_tunnel_item_release(struct rte_eth_dev *dev, + struct rte_flow_item *pmd_items, + uint32_t num_items, struct rte_flow_error *err) +{ + struct tunnel_db_element_release_ctx ctx = { + .items = pmd_items, + .actions = NULL, + .num_elements = num_items, + .error = err, + }; + + mlx5_access_tunnel_offload_db(dev, tunnel_element_release_match, + tunnel_element_release_hit, + tunnel_element_release_miss, &ctx, false); + + return ctx.ret; +} + +static int +mlx5_flow_tunnel_action_release(struct rte_eth_dev *dev, + struct rte_flow_action *pmd_actions, + uint32_t num_actions, struct rte_flow_error *err) +{ + struct tunnel_db_element_release_ctx ctx = { + .items = NULL, + .actions = pmd_actions, + .num_elements = num_actions, + .error = err, + }; + + mlx5_access_tunnel_offload_db(dev, tunnel_element_release_match, + tunnel_element_release_hit, + tunnel_element_release_miss, &ctx, false); + + return ctx.ret; +} + +static int +mlx5_flow_tunnel_get_restore_info(struct rte_eth_dev *dev, + struct rte_mbuf *m, + struct rte_flow_restore_info *info, + struct rte_flow_error *err) +{ + uint64_t ol_flags = m->ol_flags; + const struct mlx5_flow_tbl_data_entry *tble; + const uint64_t mask = PKT_RX_FDIR | PKT_RX_FDIR_ID; + + if (!is_tunnel_offload_active(dev)) { + info->flags = 0; + return 0; + } + + if ((ol_flags & mask) != mask) + goto err; + tble = tunnel_mark_decode(dev, m->hash.fdir.hi); + if (!tble) { + DRV_LOG(DEBUG, "port %u invalid miss tunnel mark %#x", + dev->data->port_id, m->hash.fdir.hi); + goto err; + } + MLX5_ASSERT(tble->tunnel); + memcpy(&info->tunnel, &tble->tunnel->app_tunnel, sizeof(info->tunnel)); + info->group_id = tble->group_id; + info->flags = RTE_FLOW_RESTORE_INFO_TUNNEL | + RTE_FLOW_RESTORE_INFO_GROUP_ID | + RTE_FLOW_RESTORE_INFO_ENCAPSULATED; + + return 0; + +err: + return rte_flow_error_set(err, EINVAL, + RTE_FLOW_ERROR_TYPE_UNSPECIFIED, NULL, + "failed to get restore info"); +} + +#else /* HAVE_IBV_FLOW_DV_SUPPORT */ +static int +mlx5_flow_tunnel_decap_set(__rte_unused struct rte_eth_dev *dev, + __rte_unused struct rte_flow_tunnel *app_tunnel, + __rte_unused struct rte_flow_action **actions, + __rte_unused uint32_t *num_of_actions, + __rte_unused struct rte_flow_error *error) +{ + return -ENOTSUP; +} + +static int +mlx5_flow_tunnel_match(__rte_unused struct rte_eth_dev *dev, + __rte_unused struct rte_flow_tunnel *app_tunnel, + __rte_unused struct rte_flow_item **items, + __rte_unused uint32_t *num_of_items, + __rte_unused struct rte_flow_error *error) +{ + return -ENOTSUP; +} + +static int +mlx5_flow_tunnel_item_release(__rte_unused struct rte_eth_dev *dev, + __rte_unused struct rte_flow_item *pmd_items, + __rte_unused uint32_t num_items, + __rte_unused struct rte_flow_error *err) +{ + return -ENOTSUP; +} + +static int +mlx5_flow_tunnel_action_release(__rte_unused struct rte_eth_dev *dev, + __rte_unused struct rte_flow_action *pmd_action, + __rte_unused uint32_t num_actions, + __rte_unused struct rte_flow_error *err) +{ + return -ENOTSUP; +} + +static int +mlx5_flow_tunnel_get_restore_info(__rte_unused struct rte_eth_dev *dev, + __rte_unused struct rte_mbuf *m, + __rte_unused struct rte_flow_restore_info *i, + __rte_unused struct rte_flow_error *err) +{ + return -ENOTSUP; +} + +static int +flow_tunnel_add_default_miss(__rte_unused struct rte_eth_dev *dev, + __rte_unused struct rte_flow *flow, + __rte_unused const struct rte_flow_attr *attr, + __rte_unused const struct rte_flow_action *actions, + __rte_unused uint32_t flow_idx, + __rte_unused const struct mlx5_flow_tunnel *tunnel, + __rte_unused struct tunnel_default_miss_ctx *ctx, + __rte_unused struct rte_flow_error *error) +{ + return -ENOTSUP; +} + +static struct mlx5_flow_tunnel * +mlx5_find_tunnel_id(__rte_unused struct rte_eth_dev *dev, + __rte_unused uint32_t id) +{ + return NULL; +} + +static void +mlx5_flow_tunnel_free(__rte_unused struct rte_eth_dev *dev, + __rte_unused struct mlx5_flow_tunnel *tunnel) +{ +} + +static uint32_t +tunnel_flow_group_to_flow_table(__rte_unused struct rte_eth_dev *dev, + __rte_unused const struct mlx5_flow_tunnel *t, + __rte_unused uint32_t group, + __rte_unused uint32_t *table, + struct rte_flow_error *error) +{ + return rte_flow_error_set(error, ENOTSUP, + RTE_FLOW_ERROR_TYPE_UNSPECIFIED, NULL, + "tunnel offload requires DV support"); +} + +void +mlx5_release_tunnel_hub(__rte_unused struct mlx5_dev_ctx_shared *sh, + __rte_unused uint16_t port_id) +{ +} +#endif /* HAVE_IBV_FLOW_DV_SUPPORT */ + +static void +mlx5_dbg__print_pattern(const struct rte_flow_item *item) +{ + int ret; + struct rte_flow_error error; + + for (; item->type != RTE_FLOW_ITEM_TYPE_END; item++) { + char *item_name; + ret = rte_flow_conv(RTE_FLOW_CONV_OP_ITEM_NAME_PTR, &item_name, + sizeof(item_name), + (void *)(uintptr_t)item->type, &error); + if (ret > 0) + printf("%s ", item_name); + else + printf("%d\n", (int)item->type); + } + printf("END\n"); +}