MIME-Version: 1.0
References: <20211123230632.1503854-1-zenczykowski@gmail.com>
In-Reply-To: <20211123230632.1503854-1-zenczykowski@gmail.com>
From:   Song Liu <song@kernel.org>
Date:   Fri, 26 Nov 2021 18:32:10 -0800
Message-ID: <CAPhsuW7AYkW5-5Labsj9ziFQzcEaqcZU0xCqTfhqJhNJf8kuxA@mail.gmail.com>
Subject: Re: [PATCH bpf-next] net-bpf: bpf_skb_change_proto() - add support
 for ipv6 fragments
To:     =?UTF-8?Q?Maciej_=C5=BBenczykowski?= <zenczykowski@gmail.com>
Cc:     =?UTF-8?Q?Maciej_=C5=BBenczykowski?= <maze@google.com>,
        Alexei Starovoitov <ast@kernel.org>,
        Daniel Borkmann <daniel@iogearbox.net>,
        Linux Network Development Mailing List 
        <netdev@vger.kernel.org>,
        Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
        BPF Mailing List <bpf@vger.kernel.org>,
        "David S . Miller" <davem@davemloft.net>,
        Lorenzo Colitti <lorenzo@google.com>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

On Tue, Nov 23, 2021 at 3:06 PM Maciej =C5=BBenczykowski
<zenczykowski@gmail.com> wrote:
>
> From: Maciej =C5=BBenczykowski <maze@google.com>
>
> IPv4 fragments (20 byte IPv4 header) need to be translated to/from
> IPv6 fragments (40 byte IPv6 header with additional 8 byte IPv6
> fragmentation header).
>
> This allows this to be done by adding an extra flag BPF_F_IPV6_FRAGMENT
> to bpf_skb_change_proto().
>
> I think this is already technically achievable via the use of
> bpf_skb_adjust_room() which was added in v4.12 commit 2be7e212d541,
> but this is far easier to use and eliminates the need to call two
> helper functions, so it's also faster.
>
> Cc: Lorenzo Colitti <lorenzo@google.com>
> Signed-off-by: Maciej =C5=BBenczykowski <maze@google.com>

Please add a selftest to exercise the new flag.

Thanks,
Song