According to documentation [1], function rte_atomic16_cmpset() return value is non-zero on success; 0 on failure. In existing code this function is called, and the return value is compared to AGE_CANDIDATE, which is defined as 1. Such comparison is incorrect and can lead to unwanted behavior.
This patch updates the calls to rte_atomic16_cmpset(), to check that the return value is 0 or non-zero. [1] https://doc.dpdk.org/api/rte__atomic_8h.html Fixes: fa2d01c87d2b ("net/mlx5: support flow aging") Cc: sta...@dpdk.org Signed-off-by: Dekel Peled <dek...@nvidia.com> Acked-by: Matan Azrad <ma...@nvidia.com> --- drivers/net/mlx5/mlx5_flow.c | 8 ++------ drivers/net/mlx5/mlx5_flow_dv.c | 9 ++++----- 2 files changed, 6 insertions(+), 11 deletions(-) diff --git a/drivers/net/mlx5/mlx5_flow.c b/drivers/net/mlx5/mlx5_flow.c index 3d38e11..2729629 100644 --- a/drivers/net/mlx5/mlx5_flow.c +++ b/drivers/net/mlx5/mlx5_flow.c @@ -6747,12 +6747,8 @@ struct mlx5_meter_domains_infos * priv = rte_eth_devices[age_param->port_id].data->dev_private; age_info = GET_PORT_AGE_INFO(priv); rte_spinlock_lock(&age_info->aged_sl); - /* If the cpmset fails, release happens. */ - if (rte_atomic16_cmpset((volatile uint16_t *) - &age_param->state, - AGE_CANDIDATE, - AGE_TMOUT) == - AGE_CANDIDATE) { + if (rte_atomic16_cmpset((volatile uint16_t *)&age_param->state, + AGE_CANDIDATE, AGE_TMOUT)) { TAILQ_INSERT_TAIL(&age_info->aged_counters, cnt, next); MLX5_AGE_SET(age_info, MLX5_AGE_EVENT_NEW); } diff --git a/drivers/net/mlx5/mlx5_flow_dv.c b/drivers/net/mlx5/mlx5_flow_dv.c index 361c32d..174189a 100644 --- a/drivers/net/mlx5/mlx5_flow_dv.c +++ b/drivers/net/mlx5/mlx5_flow_dv.c @@ -5103,10 +5103,8 @@ struct field_modify_info modify_tcp[] = { age_info = GET_PORT_AGE_INFO(priv); age_param = flow_dv_counter_idx_get_age(dev, counter); - if (rte_atomic16_cmpset((volatile uint16_t *) - &age_param->state, - AGE_CANDIDATE, AGE_FREE) - != AGE_CANDIDATE) { + if (!rte_atomic16_cmpset((volatile uint16_t *)&age_param->state, + AGE_CANDIDATE, AGE_FREE)) { /** * We need the lock even it is age timeout, * since counter may still in process. @@ -5114,9 +5112,10 @@ struct field_modify_info modify_tcp[] = { rte_spinlock_lock(&age_info->aged_sl); TAILQ_REMOVE(&age_info->aged_counters, cnt, next); rte_spinlock_unlock(&age_info->aged_sl); + rte_atomic16_set(&age_param->state, AGE_FREE); } - rte_atomic16_set(&age_param->state, AGE_FREE); } + /** * Release a flow counter. * -- 1.8.3.1
