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: <20230726121618.19198-1-zegao@tencent.com> <CAP-5=fVJ_RL9Md92R4nWBKvZcF9iovL07-=dJ7bqN3LWwufVGw@mail.gmail.com>
In-Reply-To: <CAP-5=fVJ_RL9Md92R4nWBKvZcF9iovL07-=dJ7bqN3LWwufVGw@mail.gmail.com>
From:   Ze Gao <zegao2021@gmail.com>
Date:   Mon, 31 Jul 2023 10:09:23 +0800
Message-ID: <CAD8CoPBC-OT+EWp_Gfqg31p8S91hbReR1nEnVWBzsUmtmKtZKw@mail.gmail.com>
Subject: Re: [RFC PATCH v2 0/3] report task state in symbolic chars from sched tracepoint
To:     Ian Rogers <irogers@google.com>
Cc:     Peter Zijlstra <peterz@infradead.org>,
        Steven Rostedt <rostedt@goodmis.org>,
        Namhyung Kim <namhyung@kernel.org>,
        Adrian Hunter <adrian.hunter@intel.com>,
        Alexander Shishkin <alexander.shishkin@linux.intel.com>,
        Arnaldo Carvalho de Melo <acme@kernel.org>,
        Ingo Molnar <mingo@redhat.com>, Jiri Olsa <jolsa@kernel.org>,
        Mark Rutland <mark.rutland@arm.com>,
        Masami Hiramatsu <mhiramat@kernel.org>,
        linux-kernel@vger.kernel.org, linux-perf-users@vger.kernel.org,
        linux-trace-kernel@vger.kernel.org,
        linux-trace-devel@vger.kernel.org, Ze Gao <zegao@tencent.com>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

Thanks Ian,

In regard to ABI, symbolic chars are much more stable and I think
we can benefit from this in the long run.

Regards,
Ze

On Sat, Jul 29, 2023 at 1:29=E2=80=AFAM Ian Rogers <irogers@google.com> wro=
te:
>
> On Wed, Jul 26, 2023 at 5:16=E2=80=AFAM Ze Gao <zegao2021@gmail.com> wrot=
e:
> >
> >
> > This is the 2nd attempt to fix the report task state issue in sched
> > tracepint, here is the first version:
> >
> > https://lore.kernel.org/linux-trace-kernel/20230725072254.32045-1-zegao=
@tencent.com
> >
> > Against v1, add a new var to report task state in symbolic char instead
> > of replacing the old one and to not to break anything.
> >
> > --
> >
> > In the status quo, we should see three different outcomes of the report=
ed
> > sched-out task state from perf-script, perf-sched-timehist, and Tp_prin=
tk
> > of tracepoint sched_switch.  And it's not hard to figure out that the
> > former two are built upon the third one, and the reason why we see this
> > inconsistency is that the former two does not catch up with the interna=
l
> > change of reported task state definitions as the kernel evolves.
> >
> > IMHO, exporting internal representations of task state in the tracepoin=
t
> > sched_switch is not a good practice and not encouraged at all, which ca=
n
> > easily break userspace tools that relies on it. Especially when tracepo=
ints
> > are massively used in many observability tools nowadays due to its stab=
le
> > nature, which makes them no longer used for debug only purpose and we
> > should be careful to decide what ought to be reported to userspace and =
what
> > ought not.
> >
> > Therefore, to fix the issues mentioned above for good, instead of choos=
ing
> > I proposed to add a new variable to report task state in sched_switch w=
ith
> > a symbolic character along with the old hardcoded value, and save the
> > further processing of userspace tools and spare them from knowing
> > implementation details in the kernel.
> >
> > After this patch seires, we report 'RSDTtXZPI' the same as in procfs, p=
lus
> > a 'p' which denotes PREEMP_ACTIVE and is used for sched_switch tracepoi=
nt only.
> >
> > Reviews welcome!
>
> Thanks Ze,
>
> I think this is worthwhile cleanup and makes the code overall simpler.
> I don't know if others have strong opinions, I don't often work in
> this code, but I think the patches are worth landing this.
>
> Acked-by: Ian Rogers <irogers@google.com>
>
> Thanks,
> Ian
>
> > Regards,
> >
> > Ze
> >
> > Ze Gao (2):
> >   sched, tracing: add to report task state in symbolic chars
> >   perf sched: use the new prev_state_char instead in tracepoint
> >     sched_switch
> >
> >  include/trace/events/sched.h | 60 +++++++++++++++++++++---------------
> >  tools/perf/builtin-sched.c   | 57 ++++++----------------------------
> >  2 files changed, 45 insertions(+), 72 deletions(-)
> >
> > Ze Gao (1):
> >   libtraceevent: use the new prev_state_char instead in tracepoint
> >     sched_switch
> >
> >  plugins/plugin_sched_switch.c | 29 ++++-------------------------
> >  1 file changed, 4 insertions(+), 25 deletions(-)
> >
> > --
> > 2.40.1
> >