Re: [PATCH v3 bpf-next 4/9] tcp: Add unknown_opt arg to tcp_parse_options

From: Martin KaFai Lau <kafai@fb.com>
To: Eric Dumazet <edumazet@google.com>
Cc: bpf <bpf@vger.kernel.org>, Alexei Starovoitov <ast@kernel.org>,
	Daniel Borkmann <daniel@iogearbox.net>,
	kernel-team <kernel-team@fb.com>, Lawrence Brakmo <brakmo@fb.com>,
	Neal Cardwell <ncardwell@google.com>,
	netdev <netdev@vger.kernel.org>,
	Yuchung Cheng <ycheng@google.com>
Subject: Re: [PATCH v3 bpf-next 4/9] tcp: Add unknown_opt arg to tcp_parse_options
Date: Fri, 31 Jul 2020 10:37:34 -0700	[thread overview]
Message-ID: <20200731173734.2eqvydx4cb5lllty@kafai-mbp.dhcp.thefacebook.com> (raw)
In-Reply-To: <CANn89i+f4se896OPGx6dPKZuObeJR2gaTExqoAHmDK=r7cTmaw@mail.gmail.com>

On Fri, Jul 31, 2020 at 09:12:10AM -0700, Eric Dumazet wrote:
> On Thu, Jul 30, 2020 at 1:58 PM Martin KaFai Lau <kafai@fb.com> wrote:
> >
> > In the latter patch, the bpf prog only wants to be called to handle
> > a header option if that particular header option cannot be handled by
> > the kernel.  This unknown option could be written by the peer's bpf-prog.
> > It could also be a new standard option that the running kernel does not
> > support it while a bpf-prog can handle it.
> >
> > In a latter patch, the bpf prog will be called from tcp_validate_incoming()
> > if there is unknown option and a flag is set in tp->bpf_sock_ops_cb_flags.
> >
> > Instead of using skb->cb[] in an earlier attempt, this patch
> > adds an optional arg "bool *unknown_opt" to tcp_parse_options().
> > The bool will be set to true if it has encountered an option
> > that the kernel does not recognize.
> >
> > Signed-off-by: Martin KaFai Lau <kafai@fb.com>
> > ---
> >  drivers/infiniband/hw/cxgb4/cm.c |  2 +-
> >  include/net/tcp.h                |  3 ++-
> >  net/ipv4/syncookies.c            |  2 +-
> >  net/ipv4/tcp_input.c             | 40 +++++++++++++++++++++-----------
> >  net/ipv4/tcp_minisocks.c         |  4 ++--
> >  net/ipv6/syncookies.c            |  2 +-
> >  6 files changed, 34 insertions(+), 19 deletions(-)
> >
> > diff --git a/drivers/infiniband/hw/cxgb4/cm.c b/drivers/infiniband/hw/cxgb4/cm.c
> > index 30e08bcc9afb..dedca6576bb9 100644
> > --- a/drivers/infiniband/hw/cxgb4/cm.c
> > +++ b/drivers/infiniband/hw/cxgb4/cm.c
> > @@ -3949,7 +3949,7 @@ static void build_cpl_pass_accept_req(struct sk_buff *skb, int stid , u8 tos)
> >          */
> >         memset(&tmp_opt, 0, sizeof(tmp_opt));
> >         tcp_clear_options(&tmp_opt);
> > -       tcp_parse_options(&init_net, skb, &tmp_opt, 0, NULL);
> > +       tcp_parse_options(&init_net, skb, &tmp_opt, 0, NULL, NULL);
> >
> >         req = __skb_push(skb, sizeof(*req));
> >         memset(req, 0, sizeof(*req));
> > diff --git a/include/net/tcp.h b/include/net/tcp.h
> > index 895e7aabf136..d49d8f1c961a 100644
> > --- a/include/net/tcp.h
> > +++ b/include/net/tcp.h
> > @@ -413,7 +413,8 @@ int tcp_mmap(struct file *file, struct socket *sock,
> >  #endif
> >  void tcp_parse_options(const struct net *net, const struct sk_buff *skb,
> >                        struct tcp_options_received *opt_rx,
> > -                      int estab, struct tcp_fastopen_cookie *foc);
> > +                      int estab, struct tcp_fastopen_cookie *foc,
> > +                      bool *unknown_opt);
> >  const u8 *tcp_parse_md5sig_option(const struct tcphdr *th);
> >
> 
> Instead of changing signatures of many functions (and make future
> stable backports challenging)
> how about adding a field into 'struct tcp_options_received' ?
Sounds good.  There is a one byte hole in 'struct tcp_options_received',
so it won't matter much even there is "rx_opt" in "struct tcp_sock".