Re: [dpdk-dev] [PATCH] net/mlx5: fix drop action seg fault

["Nélio Laranjeiro" <nelio.laranjeiro@6wind.com>]

On Wed, Jun 21, 2017 at 11:55:58AM +0000, Shachar Beiser wrote:
> Missing room in flow allocation to store the drop specification.
> Changing flow without storing the change in rte_flow.
> Fixes: 88c77dedfbb0 ("net/mlx5: implement drop action in hardware classifier")
> 
> Signed-off-by: Shachar Beiser <shacharbe@mellanox.com>
> ---
>  drivers/net/mlx5/mlx5_flow.c | 8 +++++---
>  1 file changed, 5 insertions(+), 3 deletions(-)
> 
> diff --git a/drivers/net/mlx5/mlx5_flow.c b/drivers/net/mlx5/mlx5_flow.c
> index 12893c6..cad3d16 100644
> --- a/drivers/net/mlx5/mlx5_flow.c
> +++ b/drivers/net/mlx5/mlx5_flow.c
> @@ -580,6 +580,8 @@ struct mlx5_flow_action {
>  	}
>  	if (action->mark && !flow->ibv_attr && !action->drop)
>  		flow->offset += sizeof(struct ibv_exp_flow_spec_action_tag);
> +	if (!flow->ibv_attr && action->drop)
> +		flow->offset += sizeof(struct ibv_exp_flow_spec_action_drop);
>  	if (!action->queue && !action->drop) {
>  		rte_flow_error_set(error, ENOTSUP, RTE_FLOW_ERROR_TYPE_HANDLE,
>  				   NULL, "no valid action");
> @@ -1011,9 +1013,6 @@ struct mlx5_flow_action {
>  		return NULL;
>  	}
>  	rte_flow->drop = 1;
> -	rte_flow->ibv_attr = flow->ibv_attr;
> -	if (!priv->started)
> -		return rte_flow;
>  #ifdef HAVE_VERBS_IBV_EXP_FLOW_SPEC_ACTION_DROP
>  	drop = (void *)((uintptr_t)flow->ibv_attr + flow->offset);
>  	*drop = (struct ibv_exp_flow_spec_action_drop){
> @@ -1023,6 +1022,9 @@ struct mlx5_flow_action {
>  	++flow->ibv_attr->num_of_specs;
>  	flow->offset += sizeof(struct ibv_exp_flow_spec_action_drop);
>  #endif
> +	rte_flow->ibv_attr = flow->ibv_attr;
> +	if (!priv->started)
> +		return rte_flow;
>  	rte_flow->qp = priv->flow_drop_queue->qp;
>  	rte_flow->ibv_flow = ibv_exp_create_flow(rte_flow->qp,
>  						 rte_flow->ibv_attr);
> -- 
> 1.8.3.1

Acked-by: Nelio Laranjeiro <nelio.laranjeiro@6wind.com>

Ferruh, this patch is fixing an issue of a patch only present in your
master-net branch, the fixline sha1 will be wrong.

Thanks,

-- 
Nélio Laranjeiro
6WIND