MIME-Version: 1.0
References: <20211129231830.1117781-1-namhyung@kernel.org> <20211129231830.1117781-5-namhyung@kernel.org>
 <011B6FDF-15F0-411F-98F2-155E95BE7047@linux.vnet.ibm.com>
In-Reply-To: <011B6FDF-15F0-411F-98F2-155E95BE7047@linux.vnet.ibm.com>
From:   Namhyung Kim <namhyung@kernel.org>
Date:   Mon, 6 Dec 2021 09:16:18 -0800
Message-ID: <CAM9d7cgaYC9BxeA3dRtb8CP+9BnFQ7pXgG9X+ku_CZ43i=ZqGg@mail.gmail.com>
Subject: Re: [PATCH 4/5] perf ftrace: Add -b/--use-bpf option for latency subcommand
To:     Athira Rajeev <atrajeev@linux.vnet.ibm.com>
Cc:     Arnaldo Carvalho de Melo <acme@kernel.org>,
        Jiri Olsa <jolsa@redhat.com>, Ingo Molnar <mingo@kernel.org>,
        Peter Zijlstra <peterz@infradead.org>,
        LKML <linux-kernel@vger.kernel.org>,
        Andi Kleen <ak@linux.intel.com>,
        Ian Rogers <irogers@google.com>,
        Stephane Eranian <eranian@google.com>,
        Song Liu <songliubraving@fb.com>,
        Changbin Du <changbin.du@gmail.com>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: 8BIT
Precedence: bulk

Hello Athira,

On Sun, Dec 5, 2021 at 6:24 PM Athira Rajeev
<atrajeev@linux.vnet.ibm.com> wrote:
>
>
>
> > On 30-Nov-2021, at 4:48 AM, Namhyung Kim <namhyung@kernel.org> wrote:
> >
> > The -b/--use-bpf option is to use BPF to get latency info of kernel
> > functions.  It'd have better performance impact and I observed that
> > latency of same function is smaller than before when using BPF.
> >
> > Signed-off-by: Namhyung Kim <namhyung@kernel.org>
> > ---
[SNIP]
> > @@ -1144,6 +1166,8 @@ int cmd_ftrace(int argc, const char **argv)
> >       const struct option latency_options[] = {
> >       OPT_CALLBACK('T', "trace-funcs", &ftrace.filters, "func",
> >                    "Show latency of given function", parse_filter_func),
> > +     OPT_BOOLEAN('b', "use-bpf", &ftrace.target.use_bpf,
> > +                 "Use BPF to measure function latency"),
>
>
> Hi Namhyung,
>
> Can this be inside BPF_SKEL check, similar to how we have “bpf-prog” and other options in builtin-stat.c ?
>
> #ifdef HAVE_BPF_SKEL
> <<OPT_BOOLEAN for use-bpf>>
> #endif
>
> Otherwise when using “-b” and if perf is not built with BPF_SKEL, we will just return in perf_ftrace__latency_prepare_bpf without any error messages.

Thanks for reporting this.  Yeah, it should report
error messages in such conditions.

I think it'd be better to add an error message
rather than hiding the option.

Thanks,
Namhyung