Enqueue generated counter IDs on a ring in bulk.
Generate them and store in an array before putting them
on a ring all at once. That bring better cache access
and speeds up the mlx5_hws_cnt_pool_create() function.
Signed-off-by: Alexander Kozyrev <akozy...@nvidia.com>
---
drivers/net/mlx5/mlx5_hws_cnt.c | 19 ++++++++++++++-----
1 file changed, 14 insertions(+), 5 deletions(-)
diff --git a/drivers/net/mlx5/mlx5_hws_cnt.c b/drivers/net/mlx5/mlx5_hws_cnt.c
index 7b5e7310af..2d2faa2c65 100644
--- a/drivers/net/mlx5/mlx5_hws_cnt.c
+++ b/drivers/net/mlx5/mlx5_hws_cnt.c
@@ -22,12 +22,17 @@
#define HWS_CNT_CACHE_THRESHOLD_DEFAULT 254
#define HWS_CNT_ALLOC_FACTOR_DEFAULT 20
-static void
+static int
__hws_cnt_id_load(struct mlx5_hws_cnt_pool *cpool)
{
uint32_t cnt_num = mlx5_hws_cnt_pool_get_size(cpool);
uint32_t iidx;
+ cnt_id_t *cnt_arr = NULL;
+ cnt_arr = mlx5_malloc(MLX5_MEM_ANY | MLX5_MEM_ZERO,
+ cnt_num * sizeof(cnt_id_t), 0, SOCKET_ID_ANY);
+ if (cnt_arr == NULL)
+ return -ENOMEM;
/*
* Counter ID order is important for tracking the max number of in used
* counter for querying, which means counter internal index order must
@@ -38,10 +43,12 @@ __hws_cnt_id_load(struct mlx5_hws_cnt_pool *cpool)
*/
for (iidx = 0; iidx < cnt_num; iidx++) {
cnt_id_t cnt_id = mlx5_hws_cnt_id_gen(cpool, iidx);
-
- rte_ring_enqueue_elem(cpool->free_list, &cnt_id,
- sizeof(cnt_id));
+ cnt_arr[iidx] = cnt_id;
}
+ rte_ring_enqueue_bulk_elem(cpool->free_list, cnt_arr,
+ sizeof(cnt_id_t), cnt_num, NULL);
+ mlx5_free(cnt_arr);
+ return 0;
}
static void
@@ -745,7 +752,9 @@ mlx5_hws_cnt_pool_create(struct rte_eth_dev *dev,
cpool->raw_mng = mlx5_hws_cnt_raw_data_alloc(priv->sh, sz, error);
if (cpool->raw_mng == NULL)
goto error;
- __hws_cnt_id_load(cpool);
+ ret = __hws_cnt_id_load(cpool);
+ if (ret != 0)
+ goto error;
/*
* Bump query gen right after pool create so the
* pre-loaded counters can be used directly
--
2.43.5
