mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: Ian Rogers <irogers@google.com>
To: Peter Zijlstra <peterz@infradead.org>,
	Ingo Molnar <mingo@redhat.com>,
	 Arnaldo Carvalho de Melo <acme@kernel.org>,
	Namhyung Kim <namhyung@kernel.org>,
	 Mark Rutland <mark.rutland@arm.com>,
	 Alexander Shishkin <alexander.shishkin@linux.intel.com>,
	Jiri Olsa <jolsa@kernel.org>,  Ian Rogers <irogers@google.com>,
	Adrian Hunter <adrian.hunter@intel.com>,
	 Kan Liang <kan.liang@linux.intel.com>,
	James Clark <james.clark@linaro.org>,
	 Xu Yang <xu.yang_2@nxp.com>, Howard Chu <howardchu95@gmail.com>,
	 "Dr. David Alan Gilbert" <linux@treblig.org>,
	Levi Yun <yeoreum.yun@arm.com>, Andi Kleen <ak@linux.intel.com>,
	 Thomas Richter <tmricht@linux.ibm.com>,
	Weilin Wang <weilin.wang@intel.com>,
	 Tiezhu Yang <yangtiezhu@loongson.cn>,
	Gautam Menghani <gautam@linux.ibm.com>,
	 linux-perf-users@vger.kernel.org, linux-kernel@vger.kernel.org
Subject: [PATCH v1 11/12] perf python: Correct pyrf_evsel__read for tool PMUs
Date: Wed,  9 Jul 2025 14:40:28 -0700	[thread overview]
Message-ID: <20250709214029.1769089-12-irogers@google.com> (raw)
In-Reply-To: <20250709214029.1769089-1-irogers@google.com>

Tool PMUs assume that stat's process_counter_values is being used to
read the counters. Specifically they hold onto old values in
evsel->prev_raw_counts and give the cumulative count based off of this
value. Update pyrf_evsel__read to allocate counts and prev_raw_counts,
use evsel__read_counter rather than perf_evsel__read so tool PMUs are
read from not just perf_event_open events, make the returned
pyrf_counts_values contain the delta value rather than the cumulative
value.

Fixes: 739621f65702 ("perf python: Add evsel read method")
Signed-off-by: Ian Rogers <irogers@google.com>
---
 tools/perf/util/python.c | 47 +++++++++++++++++++++++++++++++++++++---
 1 file changed, 44 insertions(+), 3 deletions(-)

diff --git a/tools/perf/util/python.c b/tools/perf/util/python.c
index 4a3c2b4dd79f..f689560192f4 100644
--- a/tools/perf/util/python.c
+++ b/tools/perf/util/python.c
@@ -10,6 +10,7 @@
 #endif
 #include <perf/mmap.h>
 #include "callchain.h"
+#include "counts.h"
 #include "evlist.h"
 #include "evsel.h"
 #include "event.h"
@@ -889,12 +890,38 @@ static PyObject *pyrf_evsel__threads(struct pyrf_evsel *pevsel)
 	return (PyObject *)pthread_map;
 }
 
+/*
+ * Ensure evsel's counts and prev_raw_counts are allocated, the latter
+ * used by tool PMUs to compute the cumulative count as expected by
+ * stat's process_counter_values.
+ */
+static int evsel__ensure_counts(struct evsel *evsel)
+{
+	int nthreads, ncpus;
+
+	if (evsel->counts != NULL)
+		return 0;
+
+	nthreads = perf_thread_map__nr(evsel->core.threads);
+	ncpus = perf_cpu_map__nr(evsel->core.cpus);
+
+	evsel->counts = perf_counts__new(ncpus, nthreads);
+	if (evsel->counts == NULL)
+		return -ENOMEM;
+
+	evsel->prev_raw_counts = perf_counts__new(ncpus, nthreads);
+	if (evsel->prev_raw_counts == NULL)
+		return -ENOMEM;
+
+	return 0;
+}
+
 static PyObject *pyrf_evsel__read(struct pyrf_evsel *pevsel,
 				  PyObject *args, PyObject *kwargs)
 {
 	struct evsel *evsel = &pevsel->evsel;
 	int cpu = 0, cpu_idx, thread = 0, thread_idx;
-	struct perf_counts_values counts;
+	struct perf_counts_values *old_count, *new_count;
 	struct pyrf_counts_values *count_values = PyObject_New(struct pyrf_counts_values,
 							       &pyrf_counts_values__type);
 
@@ -915,8 +942,22 @@ static PyObject *pyrf_evsel__read(struct pyrf_evsel *pevsel,
 			     thread);
 		return NULL;
 	}
-	perf_evsel__read(&(evsel->core), cpu_idx, thread_idx, &counts);
-	count_values->values = counts;
+
+	if (evsel__ensure_counts(evsel))
+		return PyErr_NoMemory();
+
+	/* Set up pointers to the old and newly read counter values. */
+	old_count = perf_counts(evsel->prev_raw_counts, cpu_idx, thread_idx);
+	new_count = perf_counts(evsel->counts, cpu_idx, thread_idx);
+	/* Update the value in evsel->counts. */
+	evsel__read_counter(evsel, cpu_idx, thread_idx);
+	/* Copy the value and turn it into the delta from old_count. */
+	count_values->values = *new_count;
+	count_values->values.val -= old_count->val;
+	count_values->values.ena -= old_count->ena;
+	count_values->values.run -= old_count->run;
+	/* Save the new count over the old_count for the next read. */
+	*old_count = *new_count;
 	return (PyObject *)count_values;
 }
 
-- 
2.50.0.727.gbf7dc18ff4-goog


  parent reply	other threads:[~2025-07-09 21:40 UTC|newest]

Thread overview: 13+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-07-09 21:40 [PATCH v1 00/12] Python motivated fixes and cleanup Ian Rogers
2025-07-09 21:40 ` [PATCH v1 01/12] perf hwmon_pmu: Avoid shortening hwmon PMU name Ian Rogers
2025-07-09 21:40 ` [PATCH v1 02/12] perf parse-events: Minor tidy up of event_type helper Ian Rogers
2025-07-09 21:40 ` [PATCH v1 03/12] perf spark: Fix includes and add SPDX Ian Rogers
2025-07-09 21:40 ` [PATCH v1 04/12] perf pmu: Tolerate failure to read the type for wellknown PMUs Ian Rogers
2025-07-09 21:40 ` [PATCH v1 05/12] perf metricgroup: Factor out for-each function and move out printing Ian Rogers
2025-07-09 21:40 ` [PATCH v1 06/12] perf stat: Move metric list from config to evlist Ian Rogers
2025-07-09 21:40 ` [PATCH v1 07/12] perf expr: Accumulate rather than replace in the context counts Ian Rogers
2025-07-09 21:40 ` [PATCH v1 08/12] perf jevents: If the long_desc and desc are identical then drop the long_desc Ian Rogers
2025-07-09 21:40 ` [PATCH v1 09/12] perf python: In str(evsel) use the evsel__pmu_name helper Ian Rogers
2025-07-09 21:40 ` [PATCH v1 10/12] perf python: Fix thread check in pyrf_evsel__read Ian Rogers
2025-07-09 21:40 ` Ian Rogers [this message]
2025-07-09 21:40 ` [PATCH v1 12/12] perf python: Improve leader copying from evlist Ian Rogers

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20250709214029.1769089-12-irogers@google.com \
    --to=irogers@google.com \
    --cc=acme@kernel.org \
    --cc=adrian.hunter@intel.com \
    --cc=ak@linux.intel.com \
    --cc=alexander.shishkin@linux.intel.com \
    --cc=gautam@linux.ibm.com \
    --cc=howardchu95@gmail.com \
    --cc=james.clark@linaro.org \
    --cc=jolsa@kernel.org \
    --cc=kan.liang@linux.intel.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-perf-users@vger.kernel.org \
    --cc=linux@treblig.org \
    --cc=mark.rutland@arm.com \
    --cc=mingo@redhat.com \
    --cc=namhyung@kernel.org \
    --cc=peterz@infradead.org \
    --cc=tmricht@linux.ibm.com \
    --cc=weilin.wang@intel.com \
    --cc=xu.yang_2@nxp.com \
    --cc=yangtiezhu@loongson.cn \
    --cc=yeoreum.yun@arm.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox

all inboxes | Powered by JetHome®