From mboxrd@z Thu Jan  1 00:00:00 1970
Content-Type: multipart/mixed; boundary="===============8494494874679415398=="
MIME-Version: 1.0
From: Sergey Senozhatsky <sergey.senozhatsky at gmail.com>
Subject: [Powertop] [PATCH] perf_bundle: introduce fixup_sample_trace_cpu()
Date: Tue, 08 Oct 2013 17:55:34 +0300
Message-ID: <20131008145534.GC2452@swordfish.minsk.epam.com>
In-Reply-To: CANHzE=3B1HATt=6=On+NWchhT_JhP2wDAz7D5M3=Vt6xXeit8A@mail.gmail.com
To: powertop@lists.01.org
List-ID: <powertop@lists.01.org>

--===============8494494874679415398==
Content-Type: text/plain; charset="utf-8"
MIME-Version: 1.0
Content-Transfer-Encoding: quoted-printable

Could you please test this patch?


struct perf_sample's trace cpu is a raw_smp_processor_id()
(as requsted by PERF_SAMPLE_CPU) by the time of perf_event_output()
call, which may differ from original struct perf_event's cpu.
thus we need to fix sample->trace.cpu via fixup_sample_trace_cpu()
call.

debugging output has demonstrated that some events (namely cpu_frequency)
were routed incorrectly (sample->trace.cpu !=3D pevent_get_field_val("cpu_i=
d")):
fix [cpu_frequency] cpu_nr from 1 to 0
fix [cpu_frequency] cpu_nr from 1 to 0
fix [cpu_frequency] cpu_nr from 1 to 2
fix [cpu_frequency] cpu_nr from 1 to 2
fix [cpu_frequency] cpu_nr from 1 to 3
fix [cpu_frequency] cpu_nr from 1 to 3
fix [cpu_frequency] cpu_nr from 1 to 0
fix [cpu_frequency] cpu_nr from 1 to 0
fix [cpu_frequency] cpu_nr from 1 to 2
fix [cpu_frequency] cpu_nr from 1 to 2
fix [cpu_frequency] cpu_nr from 1 to 0
fix [cpu_frequency] cpu_nr from 1 to 0
fix [cpu_frequency] cpu_nr from 1 to 0

Thanks for reporting and tracking this down to (not in any particular
order): Shaojie Sun, Jon Medhurst, Sanjay Rawat, SunShaoJie.

Cc: Shaojie Sun <shaojie.sun(a)linaro.org>
Cc: Jon Medhurst <jon.medhurst(a)linaro.org>
Cc: Sanjay Rawat <sanjay.rawat(a)linaro.org>
Cc: SunShaoJie <sunshaojie(a)huawei.com>
Signed-off-by: Sergey Senozhatsky <sergey.senozhatsky(a)gmail.com>

---

diff --git a/src/perf/perf_bundle.cpp b/src/perf/perf_bundle.cpp
index 8d480e8..b0e982b 100644
--- a/src/perf/perf_bundle.cpp
+++ b/src/perf/perf_bundle.cpp
@@ -285,6 +285,31 @@ static bool event_sort_function (void *i, void *j)
 	return (timestamp(I)<timestamp(J));
 }
 =

+/*
+ * sample's PERF_SAMPLE_CPU cpu nr is a raw_smp_processor_id() by the
+ * time of perf_event_output(), which may differ from struct perf_event
+ * cpu, thus we need to fix sample->trace.cpu.
+ */
+static void fixup_sample_trace_cpu(struct perf_sample *sample)
+{
+	struct event_format *event;
+	struct pevent_record rec;
+	unsigned long long cpu_nr;
+	int type;
+	int ret;
+
+	rec.data =3D &sample->data;
+	type =3D pevent_data_type(perf_event::pevent, &rec);
+	event =3D pevent_find_event(perf_event::pevent, type);
+	if (!event)
+		return;
+	/** don't touch trace if event does not contain cpu_id field*/
+	ret =3D pevent_get_field_val(NULL, event, "cpu_id", &rec, &cpu_nr, 0);
+	if (ret < 0)
+		return;
+	sample->trace.cpu =3D cpu_nr;
+}
+
 void perf_bundle::process(void)
 {
 	unsigned int i;
@@ -309,6 +334,7 @@ void perf_bundle::process(void)
 		if (sample->header.type !=3D PERF_RECORD_SAMPLE)
 			continue;
 =

+		fixup_sample_trace_cpu(sample);
 		handle_trace_point(&sample->data, sample->trace.cpu, sample->trace.time);
 	}
 }


--===============8494494874679415398==--