[PATCH net-next] mlx5: use correct counter

[PATCH net-next] mlx5: use correct counter

[Jonathan Lemon]

mlx5e_grp_q_update_stats seems to be using the wrong counter
for if_down_packets.

Signed-off-by: Jonathan Lemon <jonathan.lemon@gmail.com>
---
 drivers/net/ethernet/mellanox/mlx5/core/en_stats.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_stats.c b/drivers/net/ethernet/mellanox/mlx5/core/en_stats.c
index 6eee3c7d4b06..1d16e03a987d 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_stats.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_stats.c
@@ -363,7 +363,7 @@ static void mlx5e_grp_q_update_stats(struct mlx5e_priv *priv)
 	    !mlx5_core_query_q_counter(priv->mdev, priv->drop_rq_q_counter, 0,
 				       out, sizeof(out)))
 		qcnt->rx_if_down_packets = MLX5_GET(query_q_counter_out, out,
-						    out_of_buffer);
+						    rx_if_down_packets);
 }
 
 #define VNIC_ENV_OFF(c) MLX5_BYTE_OFF(query_vnic_env_out, c)
-- 
2.17.1

Re: [PATCH net-next] mlx5: use correct counter

[Saeed Mahameed]

On Tue, 2019-08-06 at 11:28 -0700, Jonathan Lemon wrote:
> mlx5e_grp_q_update_stats seems to be using the wrong counter
> for if_down_packets.
> 
> Signed-off-by: Jonathan Lemon <jonathan.lemon@gmail.com>
> ---
>  drivers/net/ethernet/mellanox/mlx5/core/en_stats.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_stats.c
> b/drivers/net/ethernet/mellanox/mlx5/core/en_stats.c
> index 6eee3c7d4b06..1d16e03a987d 100644
> --- a/drivers/net/ethernet/mellanox/mlx5/core/en_stats.c
> +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_stats.c
> @@ -363,7 +363,7 @@ static void mlx5e_grp_q_update_stats(struct
> mlx5e_priv *priv)
>  	    !mlx5_core_query_q_counter(priv->mdev, priv-
> >drop_rq_q_counter, 0,
>  				       out, sizeof(out)))
>  		qcnt->rx_if_down_packets =
> MLX5_GET(query_q_counter_out, out,
> -						    out_of_buffer);
> +						    rx_if_down_packets)
> ;

Hi Jonathan, 

This patch in not applicable (won't compile and there is no issue with
current code).

Although it is confusing but the code is correct as is.

1) your patch won't compile since there is no rx_if_down_packets field
in query_q_counter_out hw definition struct: please check
include/linux/mlx5/mlx5_ifc.h
mlx5_ifc_query_q_counter_out_bits
2) the code works as is since when interface is down and port is up,
technically from hw perspective there is "no buffer available" so the
out_of_buffer counter of the drop_rq_q_counter will count packets
dropped due to interface down..

Thanks,
Saeed.

Re: [PATCH net-next] mlx5: use correct counter

[Jonathan Lemon]

On 6 Aug 2019, at 12:03, Saeed Mahameed wrote:

> On Tue, 2019-08-06 at 11:28 -0700, Jonathan Lemon wrote:
>> mlx5e_grp_q_update_stats seems to be using the wrong counter
>> for if_down_packets.
>>
>> Signed-off-by: Jonathan Lemon <jonathan.lemon@gmail.com>
>> ---
>>  drivers/net/ethernet/mellanox/mlx5/core/en_stats.c | 2 +-
>>  1 file changed, 1 insertion(+), 1 deletion(-)
>>
>> diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_stats.c
>> b/drivers/net/ethernet/mellanox/mlx5/core/en_stats.c
>> index 6eee3c7d4b06..1d16e03a987d 100644
>> --- a/drivers/net/ethernet/mellanox/mlx5/core/en_stats.c
>> +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_stats.c
>> @@ -363,7 +363,7 @@ static void mlx5e_grp_q_update_stats(struct
>> mlx5e_priv *priv)
>>  	    !mlx5_core_query_q_counter(priv->mdev, priv-
>>> drop_rq_q_counter, 0,
>>  				       out, sizeof(out)))
>>  		qcnt->rx_if_down_packets =
>> MLX5_GET(query_q_counter_out, out,
>> -						    out_of_buffer);
>> +						    rx_if_down_packets)
>> ;
>
> Hi Jonathan,
>
> This patch in not applicable (won't compile and there is no issue with
> current code).
>
> Although it is confusing but the code is correct as is.
>
> 1) your patch won't compile since there is no rx_if_down_packets field
> in query_q_counter_out hw definition struct: please check
> include/linux/mlx5/mlx5_ifc.h
> mlx5_ifc_query_q_counter_out_bits
> 2) the code works as is since when interface is down and port is up,
> technically from hw perspective there is "no buffer available" so the
> out_of_buffer counter of the drop_rq_q_counter will count packets
> dropped due to interface down..

Hmm, confusing, but okay.  I'll send mail on an issue we're having
separately.
-- 
Jonathan