Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
MIME-Version: 1.0
References: <20230806075216.13378-1-me@manjusaka.me>
In-Reply-To: <20230806075216.13378-1-me@manjusaka.me>
From:   Eric Dumazet <edumazet@google.com>
Date:   Mon, 7 Aug 2023 14:34:24 +0200
Message-ID: <CANn89i+bMh-xU7PCxf_O5N+vy=83S+V=23mAAmbCuhjuP5Ob9g@mail.gmail.com>
Subject: Re: [PATCH] [RFC PATCH] tcp event: add new tcp:tcp_cwnd_restart event
To:     Manjusaka <me@manjusaka.me>
Cc:     mhiramat@kernel.org, rostedt@goodmis.org, davem@davemloft.net,
        dsahern@kernel.org, kuba@kernel.org, pabeni@redhat.com,
        netdev@vger.kernel.org, linux-kernel@vger.kernel.org,
        linux-trace-kernel@vger.kernel.org, bpf@vger.kernel.org
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

On Sun, Aug 6, 2023 at 9:52=E2=80=AFAM Manjusaka <me@manjusaka.me> wrote:
>
> The tcp_cwnd_restart function would be called if the user has enabled the
> tcp_slow_start_after_idle configuration and would be triggered when the
> connection is idle (like LONG RTO etc.). I think it would be great to
> add a new trace event named 'tcp:tcp_cwnd_reset'; it would help people
> to monitor the TCP state in a complicated network environment(like
> overlay/underlay SDN in Kubernetes, etc)
>
> Signed-off-by: Manjusaka <me@manjusaka.me>
> ---
>  include/trace/events/tcp.h | 7 +++++++
>  net/ipv4/tcp_output.c      | 1 +
>  2 files changed, 8 insertions(+)
>
> diff --git a/include/trace/events/tcp.h b/include/trace/events/tcp.h
> index bf06db8d2046..fa44191cc609 100644
> --- a/include/trace/events/tcp.h
> +++ b/include/trace/events/tcp.h
> @@ -187,6 +187,13 @@ DEFINE_EVENT(tcp_event_sk, tcp_rcv_space_adjust,
>         TP_ARGS(sk)
>  );
>
> +DEFINE_EVENT(tcp_event_sk, tcp_cwnd_restart,
> +
> +       TP_PROTO(struct sock *sk),
> +
> +       TP_ARGS(sk)
> +);
> +
>  TRACE_EVENT(tcp_retransmit_synack,
>
>         TP_PROTO(const struct sock *sk, const struct request_sock *req),
> diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c
> index 51d8638d4b4c..e902fa74303d 100644
> --- a/net/ipv4/tcp_output.c
> +++ b/net/ipv4/tcp_output.c
> @@ -141,6 +141,7 @@ static __u16 tcp_advertise_mss(struct sock *sk)
>   */
>  void tcp_cwnd_restart(struct sock *sk, s32 delta)
>  {
> +       trace_tcp_cwnd_restart(sk);

Do not include code before variable declarations.

>         struct tcp_sock *tp =3D tcp_sk(sk);
>         u32 restart_cwnd =3D tcp_init_cwnd(tp, __sk_dst_get(sk));
>         u32 cwnd =3D tcp_snd_cwnd(tp);
> --
> 2.34.1
>

I would rather add a trace in tcp_ca_event(), this would be more generic ?