DPDK patches and discussions
 help / color / mirror / Atom feed
From: "Nélio Laranjeiro" <nelio.laranjeiro@6wind.com>
To: Yongseok Koh <yskoh@mellanox.com>
Cc: shahafs@mellanox.com, dev@dpdk.org
Subject: Re: [dpdk-dev] [PATCH] net/mlx5: fix possible endless loop when clearing flow flags
Date: Tue, 24 Jul 2018 08:57:42 +0200	[thread overview]
Message-ID: <20180724065742.xynuw5ddr5pnesju@laranjeiro-vm.dev.6wind.com> (raw)
In-Reply-To: <20180723182744.1179-1-yskoh@mellanox.com>

On Mon, Jul 23, 2018 at 11:27:44AM -0700, Yongseok Koh wrote:
> If one of (*priv->rxqs)[] is null, the for loop can iterate infinitely as
> idx can't be increased.
> 
> Fixes: cd24d526395e ("net/mlx5: add mark/flag flow action")
> Cc: Nelio Laranjeiro <nelio.laranjeiro@6wind.com>
> 
> Signed-off-by: Yongseok Koh <yskoh@mellanox.com>
> ---
>  drivers/net/mlx5/mlx5_flow.c | 8 +++-----
>  1 file changed, 3 insertions(+), 5 deletions(-)
> 
> diff --git a/drivers/net/mlx5/mlx5_flow.c b/drivers/net/mlx5/mlx5_flow.c
> index 32854198b..c156f01eb 100644
> --- a/drivers/net/mlx5/mlx5_flow.c
> +++ b/drivers/net/mlx5/mlx5_flow.c
> @@ -2762,22 +2762,20 @@ mlx5_flow_rxq_flags_clear(struct rte_eth_dev *dev)
>  {
>  	struct priv *priv = dev->data->dev_private;
>  	unsigned int i;
> -	unsigned int idx;
>  
> -	for (idx = 0, i = 0; idx != priv->rxqs_n; ++i) {
> +	for (i = 0; i != priv->rxqs_n; ++i) {
>  		struct mlx5_rxq_ctrl *rxq_ctrl;
>  		unsigned int j;
>  
> -		if (!(*priv->rxqs)[idx])
> +		if (!(*priv->rxqs)[i])
>  			continue;
> -		rxq_ctrl = container_of((*priv->rxqs)[idx],
> +		rxq_ctrl = container_of((*priv->rxqs)[i],
>  					struct mlx5_rxq_ctrl, rxq);
>  		rxq_ctrl->flow_mark_n = 0;
>  		rxq_ctrl->rxq.mark = 0;
>  		for (j = 0; j != MLX5_FLOW_TUNNEL; ++j)
>  			rxq_ctrl->flow_tunnels_n[j] = 0;
>  		rxq_ctrl->rxq.tunnel = 0;
> -		++idx;
>  	}
>  }
>  
> -- 
> 2.11.0
 
This patch is wrong, (*priv->rxqs)[i] may un-initialised by the
application, the number of queues says how are in used, it does not mean
they are contiguous in the rxqs arrays and this due to the DPDK API
which configure the number of queues with rte_eth_dev_configure()
whereas queues are instantiated with rte_eth_rx_queue_setup() which
takes an position in the array as parameter.

Indeed this code is wrong, idx should always increase whereas i should
only increase if the (*priv->rxqs)[idx] is non null.

-- 
Nélio Laranjeiro
6WIND

  parent reply	other threads:[~2018-07-24  6:57 UTC|newest]

Thread overview: 6+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2018-07-23 18:27 Yongseok Koh
2018-07-24  5:50 ` Ori Kam
2018-07-24  6:57 ` Nélio Laranjeiro [this message]
2018-07-24 21:47   ` Yongseok Koh
2018-07-25  6:58     ` Nélio Laranjeiro
2018-07-25 13:10       ` Shahaf Shuler

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20180724065742.xynuw5ddr5pnesju@laranjeiro-vm.dev.6wind.com \
    --to=nelio.laranjeiro@6wind.com \
    --cc=dev@dpdk.org \
    --cc=shahafs@mellanox.com \
    --cc=yskoh@mellanox.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).