Re: [dpdk-dev] [PATCH] net/mlx5: fix use of atomic cmpset for age state

[Raslan Darawsheh]

Hi,

> -----Original Message-----
> From: dev <dev-boun...@dpdk.org> On Behalf Of Dekel Peled
> Sent: Thursday, October 15, 2020 2:44 PM
> To: Matan Azrad <ma...@nvidia.com>; Shahaf Shuler
> <shah...@nvidia.com>; Slava Ovsiienko <viachesl...@nvidia.com>
> Cc: dev@dpdk.org; sta...@dpdk.org
> Subject: [dpdk-dev] [PATCH] net/mlx5: fix use of atomic cmpset for age state
> 
> 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://nam11.safelinks.protection.outlook.com/?url=https%3A%2F%2Fdoc.
> dpdk.org%2Fapi%2Frte__atomic_8h.html&amp;data=04%7C01%7Crasland%
> 40nvidia.com%7Cc92848ff1c4a42b83bfb08d870ffe9d0%7C43083d15727340c1b
> 7db39efd9ccc17a%7C0%7C0%7C637383591730997946%7CUnknown%7CTWFp
> bGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVC
> I6Mn0%3D%7C1000&amp;sdata=l%2Bq5%2F8MO4Qrh8MrJzEGrrQCTzE%2FH
> 5Iw03qqXj5LG2PY%3D&amp;reserved=0
> 
> 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

Patch applied to next-net-mlx,

Kindest regards,
Raslan Darawsheh