Received-SPF: pass (google.com: domain of linux-kernel+bounces-209055-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:45d1:ec00::1 as permitted sender) client-ip=2604:1380:45d1:ec00::1;
Date: Mon, 10 Jun 2024 19:23:48 -0400
From: Willem de Bruijn <willemdebruijn.kernel@gmail.com>
To: Chengen Du <chengen.du@canonical.com>, 
 Willem de Bruijn <willemdebruijn.kernel@gmail.com>
Cc: davem@davemloft.net, 
 edumazet@google.com, 
 kuba@kernel.org, 
 pabeni@redhat.com, 
 kaber@trash.net, 
 netdev@vger.kernel.org, 
 linux-kernel@vger.kernel.org, 
 stable@vger.kernel.org
Message-ID: <66678b0413c20_bf52c2946d@willemb.c.googlers.com.notmuch>
In-Reply-To: <CAPza5qfcXSQNxz2kNVWHqYBGgnFLDa-Ey5b9y5OenZndo2a0Og@mail.gmail.com>
References: <20240608025347.90680-1-chengen.du@canonical.com>
 <CAPza5qfuNhDbhV9mau9RE=cNKMwGtJcx4pmjkoHNwpfysnw5yw@mail.gmail.com>
 <66660ec3f3e22_8dbbb294ed@willemb.c.googlers.com.notmuch>
 <CAPza5qfcXSQNxz2kNVWHqYBGgnFLDa-Ey5b9y5OenZndo2a0Og@mail.gmail.com>
Subject: Re: [PATCH v6] af_packet: Handle outgoing VLAN packets without
 hardware offloading
Precedence: bulk
Mime-Version: 1.0
Content-Type: text/plain;
 charset=utf-8
Content-Transfer-Encoding: quoted-printable

Chengen Du wrote:
> Hi Willem,
> =

> On Mon, Jun 10, 2024 at 4:21=E2=80=AFAM Willem de Bruijn
> <willemdebruijn.kernel@gmail.com> wrote:
> >
> > Chengen Du wrote:
> > > Hi,
> > >
> > > I would like to provide some additional explanations about the patc=
h.
> > >
> > >
> > > On Sat, Jun 8, 2024 at 10:54=E2=80=AFAM Chengen Du <chengen.du@cano=
nical.com> wrote:
> > > >
> > > > The issue initially stems from libpcap. The ethertype will be ove=
rwritten
> > > > as the VLAN TPID if the network interface lacks hardware VLAN off=
loading.
> > > > In the outbound packet path, if hardware VLAN offloading is unava=
ilable,
> > > > the VLAN tag is inserted into the payload but then cleared from t=
he sk_buff
> > > > struct. Consequently, this can lead to a false negative when chec=
king for
> > > > the presence of a VLAN tag, causing the packet sniffing outcome t=
o lack
> > > > VLAN tag information (i.e., TCI-TPID). As a result, the packet ca=
pturing
> > > > tool may be unable to parse packets as expected.
> > > >
> > > > The TCI-TPID is missing because the prb_fill_vlan_info() function=
 does not
> > > > modify the tp_vlan_tci/tp_vlan_tpid values, as the information is=
 in the
> > > > payload and not in the sk_buff struct. The skb_vlan_tag_present()=
 function
> > > > only checks vlan_all in the sk_buff struct. In cooked mode, the L=
2 header
> > > > is stripped, preventing the packet capturing tool from determinin=
g the
> > > > correct TCI-TPID value. Additionally, the protocol in SLL is inco=
rrect,
> > > > which means the packet capturing tool cannot parse the L3 header =
correctly.
> > > >
> > > > Link: https://github.com/the-tcpdump-group/libpcap/issues/1105
> > > > Link: https://lore.kernel.org/netdev/20240520070348.26725-1-cheng=
en.du@canonical.com/T/#u
> > > > Fixes: 393e52e33c6c ("packet: deliver VLAN TCI to userspace")
> > > > Cc: stable@vger.kernel.org
> > > > Signed-off-by: Chengen Du <chengen.du@canonical.com>
> > > > ---
> > > >  net/packet/af_packet.c | 57 ++++++++++++++++++++++++++++++++++++=
++++--
> > > >  1 file changed, 55 insertions(+), 2 deletions(-)
> > > >
> > > > diff --git a/net/packet/af_packet.c b/net/packet/af_packet.c
> > > > index ea3ebc160e25..8cffbe1f912d 100644
> > > > --- a/net/packet/af_packet.c
> > > > +++ b/net/packet/af_packet.c
> > > > @@ -538,6 +538,43 @@ static void *packet_current_frame(struct pac=
ket_sock *po,
> > > >         return packet_lookup_frame(po, rb, rb->head, status);
> > > >  }
> > > >
> > > > +static u16 vlan_get_tci(struct sk_buff *skb)
> > > > +{
> > > > +       struct vlan_hdr vhdr, *vh;
> > > > +       u8 *skb_orig_data =3D skb->data;
> > > > +       int skb_orig_len =3D skb->len;
> > > > +
> > > > +       skb_push(skb, skb->data - skb_mac_header(skb));
> > > > +       vh =3D skb_header_pointer(skb, ETH_HLEN, sizeof(vhdr), &v=
hdr);
> > > > +       if (skb_orig_data !=3D skb->data) {
> > > > +               skb->data =3D skb_orig_data;
> > > > +               skb->len =3D skb_orig_len;
> > > > +       }
> > >
> > >
> > > The reason for not directly using skb_header_pointer(skb,
> > > skb_mac_header(skb) + ETH_HLEN, ...) to get the VLAN header is due =
to
> > > the check logic in skb_header_pointer. In the SOCK_DGRAM and
> > > PACKET_OUTGOING scenarios, the offset can be a negative number, whi=
ch
> > > causes the check logic (i.e., likely(hlen - offset >=3D len)) in
> > > __skb_header_pointer() to not work as expected.
> >
> > The calculation is still correct?
> >
> > I think that this is not the first situation where negative offsets
> > can be given to skb_header_pointer.
> =

> The check will pass even if the offset is negative, but I believe this
> may not be the right approach. In my humble opinion, the expected
> check should be similar to the skb_push check, which ensures that
> after moving forward by the offset bytes, skb->data remains larger
> than or equal to skb->head to avoid accessing out-of-bound data. It
> might be worth considering adding a check in __skb_header_pointer to
> handle negative offsets, as this seems logical. However, this change
> could impact a wider range of code. Please correct me if I am
> mistaken.

Your current approach is fine too.

A negative offset greater than skb_headroom would certainly be a
problem. But in these cases where skb->mac_header is known to be
correct, the offset skb_mac_offset() against skb->data must be
within bounds, even if it may be negative.