net/mlx5: DR, Use HW specific logic API when writing STE

STEv0 format and STEv1 HW format are different, each has a
different order:
STEv0: CTRL 32B, TAG 16B, BITMASK 16B
STEv1: CTRL 32B, BITMASK 16B, TAG 16B

To make this transparent to upper layers we introduce a
new ste_ctx function to format the STE prior to writing it.
Signed-off-by: Erez Shitrit <erezsh@nvidia.com>
Signed-off-by: Alex Vesker <valex@nvidia.com>
Signed-off-by: Yevgeny Kliteynik <kliteyn@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>

drivers/net/ethernet/mellanox/mlx5/core/steering/dr_rule.c

@@ -106,10 +106,6 @@ dr_rule_handle_one_ste_in_update_list(struct mlx5dr_ste_send_info *ste_info,
 	int ret;
 
 	list_del(&ste_info->send_list);
-	ret = mlx5dr_send_postsend_ste(dmn, ste_info->ste, ste_info->data,
-				       ste_info->size, ste_info->offset);
-	if (ret)
-		goto out;
 
 	/* Copy data to ste, only reduced size or control, the last 16B (mask)
 	 * is already written to the hw.
@@ -119,6 +115,11 @@ dr_rule_handle_one_ste_in_update_list(struct mlx5dr_ste_send_info *ste_info,
 	else
 		memcpy(ste_info->ste->hw_ste, ste_info->data, DR_STE_SIZE_REDUCED);
 
+	ret = mlx5dr_send_postsend_ste(dmn, ste_info->ste, ste_info->data,
+				       ste_info->size, ste_info->offset);
+	if (ret)
+		goto out;
+
 out:
 	kfree(ste_info);
 	return ret;

drivers/net/ethernet/mellanox/mlx5/core/steering/dr_send.c

@@ -431,6 +431,8 @@ int mlx5dr_send_postsend_ste(struct mlx5dr_domain *dmn, struct mlx5dr_ste *ste,
 {
 	struct postsend_info send_info = {};
 
+	mlx5dr_ste_prepare_for_postsend(dmn->ste_ctx, data, size);
+
 	send_info.write.addr = (uintptr_t)data;
 	send_info.write.length = size;
 	send_info.write.lkey = 0;
@@ -457,6 +459,8 @@ int mlx5dr_send_postsend_htbl(struct mlx5dr_domain *dmn,
 	if (ret)
 		return ret;
 
+	mlx5dr_ste_prepare_for_postsend(dmn->ste_ctx, formatted_ste, DR_STE_SIZE);
+
 	/* Send the data iteration times */
 	for (i = 0; i < iterations; i++) {
 		u32 ste_index = i * (byte_size / DR_STE_SIZE);
@@ -480,6 +484,10 @@ int mlx5dr_send_postsend_htbl(struct mlx5dr_domain *dmn,
 				/* Copy bit_mask */
 				memcpy(data + ste_off + DR_STE_SIZE_REDUCED,
 				       mask, DR_STE_SIZE_MASK);
+				/* Only when we have mask we need to re-arrange the STE */
+				mlx5dr_ste_prepare_for_postsend(dmn->ste_ctx,
+								data + (j * DR_STE_SIZE),
+								DR_STE_SIZE);
 			}
 		}
 
@@ -509,6 +517,7 @@ int mlx5dr_send_postsend_formatted_htbl(struct mlx5dr_domain *dmn,
 	u32 byte_size = htbl->chunk->byte_size;
 	int iterations;
 	int num_stes;
+	u8 *copy_dst;
 	u8 *data;
 	int ret;
 	int i;
@@ -518,20 +527,22 @@ int mlx5dr_send_postsend_formatted_htbl(struct mlx5dr_domain *dmn,
 	if (ret)
 		return ret;
 
-	for (i = 0; i < num_stes; i++) {
-		u8 *copy_dst;
-
-		/* Copy the same ste on the data buffer */
-		copy_dst = data + i * DR_STE_SIZE;
-		memcpy(copy_dst, ste_init_data, DR_STE_SIZE);
-
-		if (update_hw_ste) {
-			/* Copy the reduced ste to hash table ste_arr */
+	if (update_hw_ste) {
+		/* Copy the reduced STE to hash table ste_arr */
+		for (i = 0; i < num_stes; i++) {
 			copy_dst = htbl->hw_ste_arr + i * DR_STE_SIZE_REDUCED;
 			memcpy(copy_dst, ste_init_data, DR_STE_SIZE_REDUCED);
 		}
 	}
 
+	mlx5dr_ste_prepare_for_postsend(dmn->ste_ctx, ste_init_data, DR_STE_SIZE);
+
+	/* Copy the same STE on the data buffer */
+	for (i = 0; i < num_stes; i++) {
+		copy_dst = data + i * DR_STE_SIZE;
+		memcpy(copy_dst, ste_init_data, DR_STE_SIZE);
+	}
+
 	/* Send the data iteration times */
 	for (i = 0; i < iterations; i++) {
 		u8 ste_index = i * (byte_size / DR_STE_SIZE);

drivers/net/ethernet/mellanox/mlx5/core/steering/dr_ste.c

@@ -356,6 +356,13 @@ void mlx5dr_ste_set_hit_addr_by_next_htbl(struct mlx5dr_ste_ctx *ste_ctx,
 	ste_ctx->set_hit_addr(hw_ste, chunk->icm_addr, chunk->num_of_entries);
 }
 
+void mlx5dr_ste_prepare_for_postsend(struct mlx5dr_ste_ctx *ste_ctx,
+				     u8 *hw_ste_p, u32 ste_size)
+{
+	if (ste_ctx->prepare_for_postsend)
+		ste_ctx->prepare_for_postsend(hw_ste_p, ste_size);
+}
+
 /* Init one ste as a pattern for ste data array */
 void mlx5dr_ste_set_formatted_ste(struct mlx5dr_ste_ctx *ste_ctx,
 				  u16 gvmi,

drivers/net/ethernet/mellanox/mlx5/core/steering/dr_ste.h

@@ -160,6 +160,9 @@ struct mlx5dr_ste_ctx {
 					u8 *hw_action,
 					u32 hw_action_sz,
 					u16 *used_hw_action_num);
+
+	/* Send */
+	void (*prepare_for_postsend)(u8 *hw_ste_p, u32 ste_size);
 };
 
 extern struct mlx5dr_ste_ctx ste_ctx_v0;

drivers/net/ethernet/mellanox/mlx5/core/steering/dr_ste_v1.c

@@ -324,6 +324,26 @@ static void dr_ste_v1_init(u8 *hw_ste_p, u16 lu_type,
 	MLX5_SET(ste_match_bwc_v1, hw_ste_p, miss_address_63_48, gvmi);
 }
 
+static void dr_ste_v1_prepare_for_postsend(u8 *hw_ste_p,
+					   u32 ste_size)
+{
+	u8 *tag = hw_ste_p + DR_STE_SIZE_CTRL;
+	u8 *mask = tag + DR_STE_SIZE_TAG;
+	u8 tmp_tag[DR_STE_SIZE_TAG] = {};
+
+	if (ste_size == DR_STE_SIZE_CTRL)
+		return;
+
+	WARN_ON(ste_size != DR_STE_SIZE);
+
+	/* Backup tag */
+	memcpy(tmp_tag, tag, DR_STE_SIZE_TAG);
+
+	/* Swap mask and tag  both are the same size */
+	memcpy(tag, mask, DR_STE_SIZE_MASK);
+	memcpy(mask, tmp_tag, DR_STE_SIZE_TAG);
+}
+
 static void dr_ste_v1_set_rx_flow_tag(u8 *s_action, u32 flow_tag)
 {
 	MLX5_SET(ste_single_action_flow_tag_v1, s_action, action_id,
@@ -1608,4 +1628,6 @@ struct mlx5dr_ste_ctx ste_ctx_v1 = {
 	.set_action_add			= &dr_ste_v1_set_action_add,
 	.set_action_copy		= &dr_ste_v1_set_action_copy,
 	.set_action_decap_l3_list	= &dr_ste_v1_set_action_decap_l3_list,
+	/* Send */
+	.prepare_for_postsend		= &dr_ste_v1_prepare_for_postsend,
 };

drivers/net/ethernet/mellanox/mlx5/core/steering/dr_types.h

@@ -1072,6 +1072,9 @@ struct mlx5dr_icm_chunk *
 mlx5dr_icm_alloc_chunk(struct mlx5dr_icm_pool *pool,
 		       enum mlx5dr_icm_chunk_size chunk_size);
 void mlx5dr_icm_free_chunk(struct mlx5dr_icm_chunk *chunk);
+
+void mlx5dr_ste_prepare_for_postsend(struct mlx5dr_ste_ctx *ste_ctx,
+				     u8 *hw_ste_p, u32 ste_size);
 int mlx5dr_ste_htbl_init_and_postsend(struct mlx5dr_domain *dmn,
 				      struct mlx5dr_domain_rx_tx *nic_dmn,
 				      struct mlx5dr_ste_htbl *htbl,