[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <CAEEQ3w=ZyzamCKtYA76EksYiHM=-=mS5jaWwH5mk9nz-T0up4w@mail.gmail.com>
Date: Tue, 10 Jan 2023 16:58:36 +0800
From: 运辉崔 <cuiyunhui@...edance.com>
To: Steven Rostedt <rostedt@...dmis.org>
Cc: Eric Dumazet <edumazet@...gle.com>, mhiramat@...nel.org,
davem@...emloft.net, kuba@...nel.org, pabeni@...hat.com,
kuniyu@...zon.com, xiyou.wangcong@...il.com,
duanxiongchun@...edance.com, linux-kernel@...r.kernel.org,
linux-trace-kernel@...r.kernel.org, netdev@...r.kernel.org,
dust.li@...ux.alibaba.com
Subject: Re: [External] Re: [PATCH v4] sock: add tracepoint for send recv length
On Mon, Jan 9, 2023 at 11:39 PM Steven Rostedt <rostedt@...dmis.org> wrote:
>
> But looking at this tracepoint again, I see a issue that can help with the
> dereferencing.
>
> Why is family and protocol passed in?
>
> + trace_sock_send_length(sock->sk, sock->sk->sk_family,
> + sock->sk->sk_protocol, ret, 0);
>
>
> Where the TP_fast_assign() is:
>
> + TP_fast_assign(
> + __entry->sk = sk;
> + __entry->family = sk->sk_family;
> + __entry->protocol = sk->sk_protocol;
> + __entry->length = ret > 0 ? ret : 0;
> + __entry->error = ret < 0 ? ret : 0;
> + __entry->flags = flags;
> + ),
>
> The family and protocol is taken from the sk, and not the parameters. I bet
> dropping those would help.
>
> -- Steve
>
Many thanks to Eric and Steven for your discussions and suggestions, I
will update on v5.
Thanks,
Yunhui
Powered by blists - more mailing lists