Received-SPF: pass (google.com: domain of linux-kernel+bounces-76065-linux.lists.archive=gmail.com@vger.kernel.org designates 147.75.48.161 as permitted sender) client-ip=147.75.48.161;
Precedence: bulk
MIME-Version: 1.0
References: <00000000000043b1310611e388aa@google.com> <20240222040046.2568269-1-lizhi.xu@windriver.com>
In-Reply-To: <20240222040046.2568269-1-lizhi.xu@windriver.com>
From: Eric Dumazet <edumazet@google.com>
Date: Thu, 22 Feb 2024 09:11:14 +0100
Message-ID: <CANn89iKJ=-V7xS2PqpkwrqF=kg7FJdTqYgVEsnvR5hie24aqZg@mail.gmail.com>
Subject: Re: [PATCH net-next] net/mpls: fix WARNING in mpls_gso_segment
To: Lizhi Xu <lizhi.xu@windriver.com>
Cc: syzbot+99d15fcdb0132a1e1a82@syzkaller.appspotmail.com, davem@davemloft.net, 
	dsahern@kernel.org, fw@strlen.de, horms@kernel.org, kuba@kernel.org, 
	linux-kernel@vger.kernel.org, netdev@vger.kernel.org, pabeni@redhat.com, 
	syzkaller-bugs@googlegroups.com
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

On Thu, Feb 22, 2024 at 5:00=E2=80=AFAM Lizhi Xu <lizhi.xu@windriver.com> w=
rote:
>
> When the network header pointer is greater than the inner network header,=
 the
> difference between the two can cause mpls_hlen overflow.
>
> Reported-and-tested-by: syzbot+99d15fcdb0132a1e1a82@syzkaller.appspotmail=
com
> Signed-off-by: Lizhi Xu <lizhi.xu@windriver.com>
> ---
>  net/mpls/mpls_gso.c | 4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/net/mpls/mpls_gso.c b/net/mpls/mpls_gso.c
> index 533d082f0701..2ab24b2fd90f 100644
> --- a/net/mpls/mpls_gso.c
> +++ b/net/mpls/mpls_gso.c
> @@ -25,11 +25,11 @@ static struct sk_buff *mpls_gso_segment(struct sk_buf=
f *skb,
>         netdev_features_t mpls_features;
>         u16 mac_len =3D skb->mac_len;
>         __be16 mpls_protocol;
> -       unsigned int mpls_hlen;
> +       int mpls_hlen;
>
>         skb_reset_network_header(skb);
>         mpls_hlen =3D skb_inner_network_header(skb) - skb_network_header(=
skb);
> -       if (unlikely(!mpls_hlen || mpls_hlen % MPLS_HLEN))
> +       if (unlikely(mpls_hlen <=3D 0 || mpls_hlen % MPLS_HLEN))
>                 goto out;
>         if (unlikely(!pskb_may_pull(skb, mpls_hlen)))
>                 goto out;
> --
> 2.43.0
>

I think Florian posted this patch, right ?

We must add a Fixes: tag

Also we should ask ourselves :
Why are we even looking at skb_inner_network_header(skb) if this was not se=
t ?

Lets not hide a real bug, we need to understand the root cause.