Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
MIME-Version: 1.0
References: <20210904064044.125549-1-yan2228598786@gmail.com>
In-Reply-To: <20210904064044.125549-1-yan2228598786@gmail.com>
From:   Cong Wang <xiyou.wangcong@gmail.com>
Date:   Sat, 4 Sep 2021 12:04:04 -0700
Message-ID: <CAM_iQpWCRYXc1PAOTehgMztfbmSVF=RqudOjhZhGeP_huaKjZw@mail.gmail.com>
Subject: Re: [PATCH] net: tcp_drop adds `reason` and SNMP parameters for
 tracing v4
To:     Zhongya Yan <yan2228598786@gmail.com>
Cc:     Eric Dumazet <edumazet@google.com>,
        Linux Kernel Network Developers <netdev@vger.kernel.org>,
        LKML <linux-kernel@vger.kernel.org>,
        Jakub Kicinski <kuba@kernel.org>,
        Steven Rostedt <rostedt@goodmis.org>,
        Ingo Molnar <mingo@redhat.com>,
        David Miller <davem@davemloft.net>,
        Hideaki YOSHIFUJI <yoshfuji@linux-ipv6.org>,
        dsahern@kernel.org, hengqi.chen@gmail.com,
        Yonghong Song <yhs@fb.com>,
        Brendan Gregg <brendan.d.gregg@gmail.com>, 2228598786@qq.com
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

On Fri, Sep 3, 2021 at 11:43 PM Zhongya Yan <yan2228598786@gmail.com> wrote:
>
> I used the suggestion from `Brendan Gregg`. In addition to the
> `reason` parameter there is also the `field` parameter pointing
> to `SNMP` to distinguish the `tcp_drop` cause. I know what I
> submitted is not accurate, so I am submitting the current
> patch to get comments and criticism from everyone so that I
> can submit better code and solutions.And of course to make me
> more familiar and understand the `linux` kernel network code.

Any reason why only limit this to TCP? I am pretty sure we are
interested in packet drops across the entire stack.

You can take a look at net/core/drop_monitor.c, it actually has
a big advantage over your solution, which is sending entire
dropped packets to user-space for inspection.

If you can have a solution for all packet drops, not just TCP,
it would make your patch much more useful. Therefore, I'd
suggest you to consider extending the drop monitor.

Thanks.