Linux RDMA and InfiniBand development
 help / color / mirror / Atom feed
From: sashiko-bot@kernel.org
To: "Artem Dinaburg" <artem@trailofbits.com>
Cc: linux-rdma@vger.kernel.org
Subject: Re: [PATCH 6.1.y] net/mlx5e: TC, Fix internal port memory leak
Date: Fri, 25 Sep 2026 17:35:12 +0000	[thread overview]
Message-ID: <20260925173513.273EF1F000FF@smtp.kernel.org> (raw)
In-Reply-To: <20260924173505.87763-1-artem@trailofbits.com>

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [High] mlx5e: double free of int_port in flow post actions backport
--

commit 31e3e836a049949a2f6ef2d83fb49017f4be37e3
Author: Jianbo Liu <jianbol@nvidia.com>

net/mlx5e: TC, Fix internal port memory leak

This commit attempts to fix a memory leak of the internal port object when
flow rules forward packets from an internal port over a tunnel. It moves the
cleanup into free_flow_post_acts() to release references on cloned
post-action attributes.

> diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c
> index 05888942ef276..d2f226a09a0c7 100644
> --- a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c
> +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c
[ ... ]
> @@ -3713,6 +3714,16 @@ free_flow_post_acts(struct mlx5e_tc_flow *flow)
>  				mlx5_modify_header_dealloc(flow->priv->mdev, attr->modify_hdr);
>  		}
>  
> +		if (mlx5e_is_eswitch_flow(flow)) {
> +			esw_attr = attr->esw_attr;
> +			if (esw_attr->int_port)
> +				mlx5e_tc_int_port_put(mlx5e_get_int_port_priv(flow->priv),
> +						      esw_attr->int_port);
> +			if (esw_attr->dest_int_port)
> +				mlx5e_tc_int_port_put(mlx5e_get_int_port_priv(flow->priv),
> +						      esw_attr->dest_int_port);
                                                      ^^^^^^^^^^^^^^^^^^^^^^^
[Severity: High]
Can this cause a double free of the mlx5e_tc_int_port object?

When parsing a multi-table action, mlx5e_clone_flow_attr_for_post_act()
performs a shallow copy of the primary attribute:

mlx5e_clone_flow_attr_for_post_act() {
	...
	memcpy(attr2, attr, attr_sz);
	...
}

This duplicates the dest_int_port and int_port pointers into the cloned
attribute without acquiring new references for them.

During flow deletion, mlx5e_tc_del_fdb_flow() already correctly puts the
primary attribute's dest_int_port, dropping the refcount and freeing the
internal port object:

mlx5e_tc_del_fdb_flow() {
	...
	if (esw_attr->dest_int_port)
		mlx5e_tc_int_port_put(mlx5e_get_int_port_priv(priv),
				      esw_attr->dest_int_port);
	...
}

Since the cloned attributes never took their own references, won't putting
the pointers again here in free_flow_post_acts() result in a use-after-free
and double free of the already-freed mlx5e_tc_int_port object?

> +		}
> +
>  		list_del(&attr->list);
>  		kvfree(attr->parse_attr);
>  		kfree(attr);

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260924173505.87763-1-artem@trailofbits.com?part=1

  reply	other threads:[~2026-09-25 17:35 UTC|newest]

Thread overview: 3+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-09-24 17:35 [PATCH 6.1.y] net/mlx5e: TC, Fix internal port memory leak Artem Dinaburg
2026-09-25 17:35 ` sashiko-bot [this message]
2026-09-25 18:48 ` Sasha Levin

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=20260925173513.273EF1F000FF@smtp.kernel.org \
    --to=sashiko-bot@kernel.org \
    --cc=artem@trailofbits.com \
    --cc=linux-rdma@vger.kernel.org \
    --cc=sashiko-reviews@lists.linux.dev \
    /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