Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::3:7 as permitted sender) client-ip=2620:137:e000::3:7;
MIME-Version: 1.0
References: <20231126151652.372783-1-syoshida@redhat.com>
In-Reply-To: <20231126151652.372783-1-syoshida@redhat.com>
From:   Eric Dumazet <edumazet@google.com>
Date:   Tue, 28 Nov 2023 16:45:05 +0100
Message-ID: <CANn89iKcstKYWoqUCXHO__7PfVRMFNnN5nRQVCTAADvFbcJRww@mail.gmail.com>
Subject: Re: [PATCH net] ipv4: ip_gre: Handle skb_pull() failure in ipgre_xmit()
To:     Shigeru Yoshida <syoshida@redhat.com>
Cc:     davem@davemloft.net, dsahern@kernel.org, kuba@kernel.org,
        pabeni@redhat.com, netdev@vger.kernel.org,
        linux-kernel@vger.kernel.org
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

On Sun, Nov 26, 2023 at 4:17=E2=80=AFPM Shigeru Yoshida <syoshida@redhat.co=
m> wrote:
>
> In ipgre_xmit(), skb_pull() may fail even if pskb_inet_may_pull() returns
> true. For example, applications can create a malformed packet that causes
> this problem with PF_PACKET.
>
> This patch fixes the problem by dropping skb and returning from the
> function if skb_pull() fails.
>
> Fixes: c54419321455 ("GRE: Refactor GRE tunneling code.")
> Signed-off-by: Shigeru Yoshida <syoshida@redhat.com>
> ---
>  net/ipv4/ip_gre.c | 3 ++-
>  1 file changed, 2 insertions(+), 1 deletion(-)
>
> diff --git a/net/ipv4/ip_gre.c b/net/ipv4/ip_gre.c
> index 22a26d1d29a0..95efa97cb84b 100644
> --- a/net/ipv4/ip_gre.c
> +++ b/net/ipv4/ip_gre.c
> @@ -643,7 +643,8 @@ static netdev_tx_t ipgre_xmit(struct sk_buff *skb,
>                 /* Pull skb since ip_tunnel_xmit() needs skb->data pointi=
ng
>                  * to gre header.
>                  */
> -               skb_pull(skb, tunnel->hlen + sizeof(struct iphdr));
> +               if (!skb_pull(skb, tunnel->hlen + sizeof(struct iphdr)))
> +                       goto free_skb;
>                 skb_reset_mac_header(skb);
>
>                 if (skb->ip_summed =3D=3D CHECKSUM_PARTIAL &&
> --


I have syszbot reports with an actual repro for this one.

I do not think your patch is correct, something should be fixed
earlier (before we hit this point)