The Linux Kernel Mailing List
 help / color / mirror / Atom feed
From: Steffen Klassert <steffen.klassert@secunet.com>
To: Rustam Mustafin <rmustafin@cloudlinux.com>
Cc: Herbert Xu <herbert@gondor.apana.org.au>,
	"David S. Miller" <davem@davemloft.net>,
	Eric Dumazet <edumazet@kernel.org>,
	Jakub Kicinski <kuba@kernel.org>, Paolo Abeni <pabeni@redhat.com>,
	Simon Horman <horms@kernel.org>, <netdev@vger.kernel.org>,
	<ipsec@lists.linux.dev>, <linux-kernel@vger.kernel.org>,
	<stable@vger.kernel.org>
Subject: Re: [PATCH ipsec] xfrm6: gate the local error on the frame's own socket
Date: Fri, 20 Nov 2026 08:56:28 +0100	[thread overview]
Message-ID: <av_9LIb7im4b-HcG@gauss4> (raw)
In-Reply-To: <20261114125552.1319352-1-rmustafin@cloudlinux.com>

This is an AI-generated review of your patch. The human sending this
email has considered the AI review valid, or at least plausible.
---
Author: Rustam Mustafin <rmustafin@cloudlinux.com>

xfrm6: gate the local error on the frame's own socket

This commit changes __xfrm6_output() to check skb_to_full_sk(skb) instead
of the sk argument before calling xfrm_local_error() on an oversized packet.

> diff --git a/net/ipv6/xfrm6_output.c b/net/ipv6/xfrm6_output.c
> index 44b221a09a0c8..54fcf4b430915 100644
> --- a/net/ipv6/xfrm6_output.c
> +++ b/net/ipv6/xfrm6_output.c
> @@ -95,7 +95,10 @@ static int __xfrm6_output(struct net *net, struct sock *sk, struct sk_buff *skb)
>  	} else if (toobig && xfrm6_noneed_fragment(skb)) {
>  		skb->ignore_df = 1;
>  		goto skip_frag;
> -	} else if (!skb->ignore_df && toobig && sk) {
> +	} else if (!skb->ignore_df && toobig && skb_to_full_sk(skb)) {
> +		/* xfrm_local_error() reports to skb->sk; a forwarded frame
> +		 * sent through a UDP tunnel has none, sk is the tunnel's.
> +		 */
>  		xfrm_local_error(skb, mtu);
>  		kfree_skb(skb);
>  		return -EMSGSIZE;

Should the earlier xfrm6_local_dontfrag(sk) check in __xfrm6_output() also
use skb_to_full_sk(skb), as xfrm6_tunnel_check_size() does?

net/ipv6/xfrm6_output.c:__xfrm6_output() {
    ...
	if (toobig && xfrm6_local_dontfrag(sk)) {
		xfrm6_local_rxpmtu(skb, mtu);
		kfree_skb(skb);
		return -EMSGSIZE;
	} else if (toobig && xfrm6_noneed_fragment(skb)) {
		skb->ignore_df = 1;
		goto skip_frag;
	} else if (!skb->ignore_df && toobig && skb_to_full_sk(skb)) {
    ...
}

When a caller passes sk == NULL while skb->sk is a local UDPv6 or RAWv6
socket with IPV6_DONTFRAG set, such as ip6_tnl_xmit() ->
ip6tunnel_xmit(NULL, skb, dev, 0) where skb_scrub_packet() clears
skb->ignore_df, xfrm6_local_dontfrag(sk) evaluates to false on NULL and
this change routes the packet into xfrm_local_error() instead of
xfrm6_local_rxpmtu() (or into ip6_fragment() via
ipvlan_process_v6_outbound() -> ip6_local_out(dev_net(dev), NULL, skb) when
skb->ignore_df remains 1).

Conversely, if a forwarded frame (skb->sk == NULL) is sent through
udp_tunnel6_xmit_skb() on a tunnel socket with IPV6_DONTFRAG set,
xfrm6_local_dontfrag(sk) returns true, xfrm6_local_rxpmtu() does nothing
because skb_to_full_sk(skb) is NULL, and the frame is dropped without
reaching ip6_fragment() to send an ICMPv6 Packet Too Big.

In addition, callers such as vrf_process_v6_outbound() pass raw skb->sk as
sk to dst_output(), so if skb->sk is a TCP_NEW_SYN_RECV or TCP_TIME_WAIT
mini-socket, xfrm6_local_dontfrag(sk) reads sk->sk_protocol and
inet_sk(sk)->inet_flags out-of-bounds past struct tcp6_request_sock or
struct tcp6_timewait_sock.

      reply	other threads:[~2026-11-20  7:56 UTC|newest]

Thread overview: 2+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-11-14 12:55 [PATCH ipsec] xfrm6: gate the local error on the frame's own socket Rustam Mustafin
2026-11-20  7:56 ` Steffen Klassert [this message]

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=av_9LIb7im4b-HcG@gauss4 \
    --to=steffen.klassert@secunet.com \
    --cc=davem@davemloft.net \
    --cc=edumazet@kernel.org \
    --cc=herbert@gondor.apana.org.au \
    --cc=horms@kernel.org \
    --cc=ipsec@lists.linux.dev \
    --cc=kuba@kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=netdev@vger.kernel.org \
    --cc=pabeni@redhat.com \
    --cc=rmustafin@cloudlinux.com \
    --cc=stable@vger.kernel.org \
    /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