new file mode 100644
@@ -0,0 +1,528 @@
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright (c) 2022 NVIDIA Corporation & Affiliates
+ */
+
+#include "mlx5dr_internal.h"
+
+static void mlx5dr_rule_skip(struct mlx5dr_matcher *matcher,
+ const struct rte_flow_item *items,
+ bool *skip_rx, bool *skip_tx)
+{
+ struct mlx5dr_match_template *mt = matcher->mt[0];
+ const struct flow_hw_port_info *vport;
+ const struct rte_flow_item_ethdev *v;
+
+ /* Flow_src is the 1st priority */
+ if (matcher->attr.optimize_flow_src) {
+ *skip_tx = matcher->attr.optimize_flow_src == MLX5DR_MATCHER_FLOW_SRC_WIRE;
+ *skip_rx = matcher->attr.optimize_flow_src == MLX5DR_MATCHER_FLOW_SRC_VPORT;
+ return;
+ }
+
+ /* By default FDB rules are added to both RX and TX */
+ *skip_rx = false;
+ *skip_tx = false;
+
+ if (mt->item_flags & MLX5_FLOW_ITEM_REPRESENTED_PORT) {
+ v = items[mt->vport_item_id].spec;
+ vport = flow_hw_conv_port_id(v->port_id);
+ if (unlikely(!vport)) {
+ DR_LOG(ERR, "Fail to map port ID %d, ignoring", v->port_id);
+ return;
+ }
+
+ if (!vport->is_wire)
+ /* Match vport ID is not WIRE -> Skip RX */
+ *skip_rx = true;
+ else
+ /* Match vport ID is WIRE -> Skip TX */
+ *skip_tx = true;
+ }
+}
+
+static void mlx5dr_rule_init_dep_wqe(struct mlx5dr_send_ring_dep_wqe *dep_wqe,
+ struct mlx5dr_rule *rule,
+ const struct rte_flow_item *items,
+ void *user_data)
+{
+ struct mlx5dr_matcher *matcher = rule->matcher;
+ struct mlx5dr_table *tbl = matcher->tbl;
+ bool skip_rx, skip_tx;
+
+ dep_wqe->rule = rule;
+ dep_wqe->user_data = user_data;
+
+ switch (tbl->type) {
+ case MLX5DR_TABLE_TYPE_NIC_RX:
+ case MLX5DR_TABLE_TYPE_NIC_TX:
+ dep_wqe->rtc_0 = matcher->match_ste.rtc_0->id;
+ dep_wqe->retry_rtc_0 = matcher->col_matcher ?
+ matcher->col_matcher->match_ste.rtc_0->id : 0;
+ dep_wqe->rtc_1 = 0;
+ dep_wqe->retry_rtc_1 = 0;
+ break;
+
+ case MLX5DR_TABLE_TYPE_FDB:
+ mlx5dr_rule_skip(matcher, items, &skip_rx, &skip_tx);
+
+ if (!skip_rx) {
+ dep_wqe->rtc_0 = matcher->match_ste.rtc_0->id;
+ dep_wqe->retry_rtc_0 = matcher->col_matcher ?
+ matcher->col_matcher->match_ste.rtc_0->id : 0;
+ } else {
+ dep_wqe->rtc_0 = 0;
+ dep_wqe->retry_rtc_0 = 0;
+ }
+
+ if (!skip_tx) {
+ dep_wqe->rtc_1 = matcher->match_ste.rtc_1->id;
+ dep_wqe->retry_rtc_1 = matcher->col_matcher ?
+ matcher->col_matcher->match_ste.rtc_1->id : 0;
+ } else {
+ dep_wqe->rtc_1 = 0;
+ dep_wqe->retry_rtc_1 = 0;
+ }
+
+ break;
+
+ default:
+ assert(false);
+ break;
+ }
+}
+
+static void mlx5dr_rule_gen_comp(struct mlx5dr_send_engine *queue,
+ struct mlx5dr_rule *rule,
+ bool err,
+ void *user_data,
+ enum mlx5dr_rule_status rule_status_on_succ)
+{
+ enum rte_flow_op_status comp_status;
+
+ if (!err) {
+ comp_status = RTE_FLOW_OP_SUCCESS;
+ rule->status = rule_status_on_succ;
+ } else {
+ comp_status = RTE_FLOW_OP_ERROR;
+ rule->status = MLX5DR_RULE_STATUS_FAILED;
+ }
+
+ mlx5dr_send_engine_inc_rule(queue);
+ mlx5dr_send_engine_gen_comp(queue, user_data, comp_status);
+}
+
+static int mlx5dr_rule_alloc_action_ste(struct mlx5dr_rule *rule,
+ struct mlx5dr_rule_attr *attr)
+{
+ struct mlx5dr_matcher *matcher = rule->matcher;
+ int ret;
+
+ /* Use rule_idx for locking optimzation, otherwise allocate from pool */
+ if (matcher->attr.optimize_using_rule_idx) {
+ rule->action_ste_idx = attr->rule_idx * matcher->action_ste.max_stes;
+ } else {
+ struct mlx5dr_pool_chunk ste = {0};
+
+ ste.order = rte_log2_u32(matcher->action_ste.max_stes);
+ ret = mlx5dr_pool_chunk_alloc(matcher->action_ste.pool, &ste);
+ if (ret) {
+ DR_LOG(ERR, "Failed to allocate STE for rule actions");
+ return ret;
+ }
+ rule->action_ste_idx = ste.offset;
+ }
+ return 0;
+}
+
+void mlx5dr_rule_free_action_ste_idx(struct mlx5dr_rule *rule)
+{
+ struct mlx5dr_matcher *matcher = rule->matcher;
+
+ if (rule->action_ste_idx > -1 && !matcher->attr.optimize_using_rule_idx) {
+ struct mlx5dr_pool_chunk ste = {0};
+
+ /* This release is safe only when the rule match part was deleted */
+ ste.order = rte_log2_u32(matcher->action_ste.max_stes);
+ ste.offset = rule->action_ste_idx;
+ mlx5dr_pool_chunk_free(matcher->action_ste.pool, &ste);
+ }
+}
+
+static void mlx5dr_rule_create_init(struct mlx5dr_rule *rule,
+ struct mlx5dr_send_ste_attr *ste_attr,
+ struct mlx5dr_actions_apply_data *apply)
+{
+ struct mlx5dr_matcher *matcher = rule->matcher;
+ struct mlx5dr_table *tbl = matcher->tbl;
+ struct mlx5dr_context *ctx = tbl->ctx;
+
+ /* Init rule before reuse */
+ rule->rtc_0 = 0;
+ rule->rtc_1 = 0;
+ rule->pending_wqes = 0;
+ rule->action_ste_idx = -1;
+ rule->status = MLX5DR_RULE_STATUS_CREATING;
+
+ /* Init default send STE attributes */
+ ste_attr->gta_opcode = MLX5DR_WQE_GTA_OP_ACTIVATE;
+ ste_attr->send_attr.opmod = MLX5DR_WQE_GTA_OPMOD_STE;
+ ste_attr->send_attr.opcode = MLX5DR_WQE_OPCODE_TBL_ACCESS;
+ ste_attr->send_attr.len = MLX5DR_WQE_SZ_GTA_CTRL + MLX5DR_WQE_SZ_GTA_DATA;
+
+ /* Init default action apply */
+ apply->tbl_type = tbl->type;
+ apply->common_res = &ctx->common_res[tbl->type];
+ apply->jump_to_action_stc = matcher->action_ste.stc.offset;
+ apply->require_dep = 0;
+}
+
+static int mlx5dr_rule_create_hws(struct mlx5dr_rule *rule,
+ struct mlx5dr_rule_attr *attr,
+ uint8_t mt_idx,
+ const struct rte_flow_item items[],
+ uint8_t at_idx,
+ struct mlx5dr_rule_action rule_actions[])
+{
+ struct mlx5dr_action_template *at = rule->matcher->at[at_idx];
+ struct mlx5dr_match_template *mt = rule->matcher->mt[mt_idx];
+ bool is_jumbo = mlx5dr_definer_is_jumbo(mt->definer);
+ struct mlx5dr_matcher *matcher = rule->matcher;
+ struct mlx5dr_context *ctx = matcher->tbl->ctx;
+ struct mlx5dr_send_ste_attr ste_attr = {0};
+ struct mlx5dr_send_ring_dep_wqe *dep_wqe;
+ struct mlx5dr_actions_wqe_setter *setter;
+ struct mlx5dr_actions_apply_data apply;
+ struct mlx5dr_send_engine *queue;
+ uint8_t total_stes, action_stes;
+ int i, ret;
+
+ queue = &ctx->send_queue[attr->queue_id];
+ if (unlikely(mlx5dr_send_engine_err(queue))) {
+ rte_errno = EIO;
+ return rte_errno;
+ }
+
+ mlx5dr_rule_create_init(rule, &ste_attr, &apply);
+
+ /* Allocate dependent match WQE since rule might have dependent writes.
+ * The queued dependent WQE can be later aborted or kept as a dependency.
+ * dep_wqe buffers (ctrl, data) are also reused for all STE writes.
+ */
+ dep_wqe = mlx5dr_send_add_new_dep_wqe(queue);
+ mlx5dr_rule_init_dep_wqe(dep_wqe, rule, items, attr->user_data);
+
+ ste_attr.wqe_ctrl = &dep_wqe->wqe_ctrl;
+ ste_attr.wqe_data = &dep_wqe->wqe_data;
+ apply.wqe_ctrl = &dep_wqe->wqe_ctrl;
+ apply.wqe_data = (uint32_t *)&dep_wqe->wqe_data;
+ apply.rule_action = rule_actions;
+ apply.queue = queue;
+
+ setter = &at->setters[at->num_of_action_stes];
+ total_stes = at->num_of_action_stes + (is_jumbo && !at->only_term);
+ action_stes = total_stes - 1;
+
+ if (action_stes) {
+ /* Allocate action STEs for complex rules */
+ ret = mlx5dr_rule_alloc_action_ste(rule, attr);
+ if (ret) {
+ DR_LOG(ERR, "Failed to allocate action memory %d", ret);
+ mlx5dr_send_abort_new_dep_wqe(queue);
+ return ret;
+ }
+ /* Skip RX/TX based on the dep_wqe init */
+ ste_attr.rtc_0 = dep_wqe->rtc_0 ? matcher->action_ste.rtc_0->id : 0;
+ ste_attr.rtc_1 = dep_wqe->rtc_1 ? matcher->action_ste.rtc_1->id : 0;
+ /* Action STEs are written to a specific index last to first */
+ ste_attr.direct_index = rule->action_ste_idx + action_stes;
+ apply.next_direct_idx = ste_attr.direct_index;
+ } else {
+ apply.next_direct_idx = 0;
+ }
+
+ for (i = total_stes; i-- > 0;) {
+ mlx5dr_action_apply_setter(&apply, setter--, !i && is_jumbo);
+
+ if (i == 0) {
+ /* Handle last match STE */
+ mlx5dr_definer_create_tag(items, mt->fc, mt->fc_sz,
+ (uint8_t *)dep_wqe->wqe_data.action);
+
+ /* Rule has dependent WQEs, match dep_wqe is queued */
+ if (action_stes || apply.require_dep)
+ break;
+
+ /* Rule has no dependencies, abort dep_wqe and send WQE now */
+ mlx5dr_send_abort_new_dep_wqe(queue);
+ ste_attr.wqe_tag_is_jumbo = is_jumbo;
+ ste_attr.send_attr.notify_hw = !attr->burst;
+ ste_attr.send_attr.user_data = dep_wqe->user_data;
+ ste_attr.send_attr.rule = dep_wqe->rule;
+ ste_attr.direct_index = 0;
+ ste_attr.rtc_0 = dep_wqe->rtc_0;
+ ste_attr.rtc_1 = dep_wqe->rtc_1;
+ ste_attr.used_id_rtc_0 = &rule->rtc_0;
+ ste_attr.used_id_rtc_1 = &rule->rtc_1;
+ ste_attr.retry_rtc_0 = dep_wqe->retry_rtc_0;
+ ste_attr.retry_rtc_1 = dep_wqe->retry_rtc_1;
+ } else {
+ apply.next_direct_idx = --ste_attr.direct_index;
+ }
+
+ mlx5dr_send_ste(queue, &ste_attr);
+ }
+
+ /* Backup TAG on the rule for deletion */
+ if (is_jumbo)
+ memcpy(rule->tag.jumbo, dep_wqe->wqe_data.action, MLX5DR_JUMBO_TAG_SZ);
+ else
+ memcpy(rule->tag.match, dep_wqe->wqe_data.tag, MLX5DR_MATCH_TAG_SZ);
+
+ mlx5dr_send_engine_inc_rule(queue);
+
+ /* Send dependent WQEs */
+ if (!attr->burst)
+ mlx5dr_send_all_dep_wqe(queue);
+
+ return 0;
+}
+
+static void mlx5dr_rule_destroy_failed_hws(struct mlx5dr_rule *rule,
+ struct mlx5dr_rule_attr *attr)
+{
+ struct mlx5dr_context *ctx = rule->matcher->tbl->ctx;
+ struct mlx5dr_send_engine *queue;
+
+ queue = &ctx->send_queue[attr->queue_id];
+
+ mlx5dr_rule_gen_comp(queue, rule, false,
+ attr->user_data, MLX5DR_RULE_STATUS_DELETED);
+
+ /* Rule failed now we can safely release action STEs */
+ mlx5dr_rule_free_action_ste_idx(rule);
+
+ /* If a rule that was indicated as burst (need to trigger HW) has failed
+ * insertion we won't ring the HW as nothing is being written to the WQ.
+ * In such case update the last WQE and ring the HW with that work
+ */
+ if (attr->burst)
+ return;
+
+ mlx5dr_send_all_dep_wqe(queue);
+ mlx5dr_send_engine_flush_queue(queue);
+}
+
+static int mlx5dr_rule_destroy_hws(struct mlx5dr_rule *rule,
+ struct mlx5dr_rule_attr *attr)
+{
+ struct mlx5dr_context *ctx = rule->matcher->tbl->ctx;
+ struct mlx5dr_matcher *matcher = rule->matcher;
+ struct mlx5dr_wqe_gta_ctrl_seg wqe_ctrl = {0};
+ struct mlx5dr_send_ste_attr ste_attr = {0};
+ struct mlx5dr_send_engine *queue;
+
+ queue = &ctx->send_queue[attr->queue_id];
+
+ /* Rule is not completed yet */
+ if (rule->status == MLX5DR_RULE_STATUS_CREATING) {
+ rte_errno = EBUSY;
+ return rte_errno;
+ }
+
+ /* Rule failed and doesn't require cleanup */
+ if (rule->status == MLX5DR_RULE_STATUS_FAILED) {
+ mlx5dr_rule_destroy_failed_hws(rule, attr);
+ return 0;
+ }
+
+ if (unlikely(mlx5dr_send_engine_err(queue))) {
+ mlx5dr_rule_destroy_failed_hws(rule, attr);
+ return 0;
+ }
+
+ mlx5dr_send_engine_inc_rule(queue);
+
+ /* Send dependent WQE */
+ if (!attr->burst)
+ mlx5dr_send_all_dep_wqe(queue);
+
+ rule->status = MLX5DR_RULE_STATUS_DELETING;
+
+ ste_attr.send_attr.opmod = MLX5DR_WQE_GTA_OPMOD_STE;
+ ste_attr.send_attr.opcode = MLX5DR_WQE_OPCODE_TBL_ACCESS;
+ ste_attr.send_attr.len = MLX5DR_WQE_SZ_GTA_CTRL + MLX5DR_WQE_SZ_GTA_DATA;
+
+ ste_attr.send_attr.rule = rule;
+ ste_attr.send_attr.notify_hw = !attr->burst;
+ ste_attr.send_attr.user_data = attr->user_data;
+
+ ste_attr.rtc_0 = rule->rtc_0;
+ ste_attr.rtc_1 = rule->rtc_1;
+ ste_attr.used_id_rtc_0 = &rule->rtc_0;
+ ste_attr.used_id_rtc_1 = &rule->rtc_1;
+ ste_attr.wqe_ctrl = &wqe_ctrl;
+ ste_attr.wqe_tag = &rule->tag;
+ ste_attr.wqe_tag_is_jumbo = mlx5dr_definer_is_jumbo(matcher->mt[0]->definer);
+ ste_attr.gta_opcode = MLX5DR_WQE_GTA_OP_DEACTIVATE;
+
+ mlx5dr_send_ste(queue, &ste_attr);
+
+ return 0;
+}
+
+static int mlx5dr_rule_create_root(struct mlx5dr_rule *rule,
+ struct mlx5dr_rule_attr *rule_attr,
+ const struct rte_flow_item items[],
+ uint8_t at_idx,
+ struct mlx5dr_rule_action rule_actions[])
+{
+ struct mlx5dv_flow_matcher *dv_matcher = rule->matcher->dv_matcher;
+ uint8_t num_actions = rule->matcher->at[at_idx]->num_actions;
+ struct mlx5dr_context *ctx = rule->matcher->tbl->ctx;
+ struct mlx5dv_flow_match_parameters *value;
+ struct mlx5_flow_attr flow_attr = {0};
+ struct mlx5dv_flow_action_attr *attr;
+ struct rte_flow_error error;
+ uint8_t match_criteria;
+ int ret;
+
+ attr = simple_calloc(num_actions, sizeof(*attr));
+ if (!attr) {
+ rte_errno = ENOMEM;
+ return rte_errno;
+ }
+
+ value = simple_calloc(1, MLX5_ST_SZ_BYTES(fte_match_param) +
+ offsetof(struct mlx5dv_flow_match_parameters, match_buf));
+ if (!value) {
+ rte_errno = ENOMEM;
+ goto free_attr;
+ }
+
+ flow_attr.tbl_type = rule->matcher->tbl->type;
+
+ ret = flow_dv_translate_items_hws(items, &flow_attr, value->match_buf,
+ MLX5_SET_MATCHER_HS_V, NULL,
+ &match_criteria,
+ &error);
+ if (ret) {
+ DR_LOG(ERR, "Failed to convert items to PRM [%s]", error.message);
+ goto free_value;
+ }
+
+ /* Convert actions to verb action attr */
+ ret = mlx5dr_action_root_build_attr(rule_actions, num_actions, attr);
+ if (ret)
+ goto free_value;
+
+ /* Create verb flow */
+ value->match_sz = MLX5_ST_SZ_BYTES(fte_match_param);
+ rule->flow = mlx5_glue->dv_create_flow_root(dv_matcher,
+ value,
+ num_actions,
+ attr);
+
+ mlx5dr_rule_gen_comp(&ctx->send_queue[rule_attr->queue_id], rule, !rule->flow,
+ rule_attr->user_data, MLX5DR_RULE_STATUS_CREATED);
+
+ simple_free(value);
+ simple_free(attr);
+
+ return 0;
+
+free_value:
+ simple_free(value);
+free_attr:
+ simple_free(attr);
+
+ return -rte_errno;
+}
+
+static int mlx5dr_rule_destroy_root(struct mlx5dr_rule *rule,
+ struct mlx5dr_rule_attr *attr)
+{
+ struct mlx5dr_context *ctx = rule->matcher->tbl->ctx;
+ int err = 0;
+
+ if (rule->flow)
+ err = ibv_destroy_flow(rule->flow);
+
+ mlx5dr_rule_gen_comp(&ctx->send_queue[attr->queue_id], rule, err,
+ attr->user_data, MLX5DR_RULE_STATUS_DELETED);
+
+ return 0;
+}
+
+int mlx5dr_rule_create(struct mlx5dr_matcher *matcher,
+ uint8_t mt_idx,
+ const struct rte_flow_item items[],
+ uint8_t at_idx,
+ struct mlx5dr_rule_action rule_actions[],
+ struct mlx5dr_rule_attr *attr,
+ struct mlx5dr_rule *rule_handle)
+{
+ struct mlx5dr_context *ctx;
+ int ret;
+
+ rule_handle->matcher = matcher;
+ ctx = matcher->tbl->ctx;
+
+ if (unlikely(!attr->user_data)) {
+ rte_errno = EINVAL;
+ return -rte_errno;
+ }
+
+ /* Check if there is room in queue */
+ if (unlikely(mlx5dr_send_engine_full(&ctx->send_queue[attr->queue_id]))) {
+ rte_errno = EBUSY;
+ return -rte_errno;
+ }
+
+ assert(matcher->num_of_mt >= mt_idx);
+ assert(matcher->num_of_at >= at_idx);
+
+ if (unlikely(mlx5dr_table_is_root(matcher->tbl)))
+ ret = mlx5dr_rule_create_root(rule_handle,
+ attr,
+ items,
+ at_idx,
+ rule_actions);
+ else
+ ret = mlx5dr_rule_create_hws(rule_handle,
+ attr,
+ mt_idx,
+ items,
+ at_idx,
+ rule_actions);
+ return -ret;
+}
+
+int mlx5dr_rule_destroy(struct mlx5dr_rule *rule,
+ struct mlx5dr_rule_attr *attr)
+{
+ struct mlx5dr_context *ctx = rule->matcher->tbl->ctx;
+ int ret;
+
+ if (unlikely(!attr->user_data)) {
+ rte_errno = EINVAL;
+ return -rte_errno;
+ }
+
+ /* Check if there is room in queue */
+ if (unlikely(mlx5dr_send_engine_full(&ctx->send_queue[attr->queue_id]))) {
+ rte_errno = EBUSY;
+ return -rte_errno;
+ }
+
+ if (unlikely(mlx5dr_table_is_root(rule->matcher->tbl)))
+ ret = mlx5dr_rule_destroy_root(rule, attr);
+ else
+ ret = mlx5dr_rule_destroy_hws(rule, attr);
+
+ return -ret;
+}
+
+size_t mlx5dr_rule_get_handle_size(void)
+{
+ return sizeof(struct mlx5dr_rule);
+}
new file mode 100644
@@ -0,0 +1,50 @@
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright (c) 2022 NVIDIA Corporation & Affiliates
+ */
+
+#ifndef MLX5DR_RULE_H_
+#define MLX5DR_RULE_H_
+
+enum {
+ MLX5DR_STE_CTRL_SZ = 20,
+ MLX5DR_ACTIONS_SZ = 12,
+ MLX5DR_MATCH_TAG_SZ = 32,
+ MLX5DR_JUMBO_TAG_SZ = 44,
+};
+
+enum mlx5dr_rule_status {
+ MLX5DR_RULE_STATUS_UNKNOWN,
+ MLX5DR_RULE_STATUS_CREATING,
+ MLX5DR_RULE_STATUS_CREATED,
+ MLX5DR_RULE_STATUS_DELETING,
+ MLX5DR_RULE_STATUS_DELETED,
+ MLX5DR_RULE_STATUS_FAILING,
+ MLX5DR_RULE_STATUS_FAILED,
+};
+
+struct mlx5dr_rule_match_tag {
+ union {
+ uint8_t jumbo[MLX5DR_JUMBO_TAG_SZ];
+ struct {
+ uint8_t reserved[MLX5DR_ACTIONS_SZ];
+ uint8_t match[MLX5DR_MATCH_TAG_SZ];
+ };
+ };
+};
+
+struct mlx5dr_rule {
+ struct mlx5dr_matcher *matcher;
+ union {
+ struct mlx5dr_rule_match_tag tag;
+ struct ibv_flow *flow;
+ };
+ uint32_t rtc_0; /* The RTC into which the STE was inserted */
+ uint32_t rtc_1; /* The RTC into which the STE was inserted */
+ int action_ste_idx; /* Action STE pool ID */
+ uint8_t status; /* enum mlx5dr_rule_status */
+ uint8_t pending_wqes;
+};
+
+void mlx5dr_rule_free_action_ste_idx(struct mlx5dr_rule *rule);
+
+#endif /* MLX5DR_RULE_H_ */