packages/kernel
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
kernel/patches/0140-perf-metricgroup-Support-multiple-events-for-metricg.patch
hongrongxuan a7fbfa09c7 Perf tool: perf tool add Hip09 json support
2023-11-10 14:47:39 +08:00

317 lines
9.0 KiB
Diff
Raw Permalink Blame History

From c650ff6651b9ee0f846f73275d580f062197ec78 Mon Sep 17 00:00:00 2001
From: Jin Yao <yao.jin@linux.intel.com>
Date: Wed, 28 Aug 2019 13:59:32 +0800
Subject: [PATCH 023/201] perf metricgroup: Support multiple events for
metricgroup
mainline inclusion
from mainline-v5.4-rc1
commit f01642e4912bb80a01d693f4cc6fb0897207a090
category: feature
bugzilla: https://gitee.com/openeuler/kernel/issues/I8C0CX
Reference: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=f01642e4912bb80a01d693f4cc6fb0897207a090
------------------------------------------------------------------------
Some uncore metrics don't work as expected. For example, on
cascadelakex:
root@lkp-csl-2sp2:~# perf stat -M UNC_M_PMM_BANDWIDTH.TOTAL -a -- sleep 1
Performance counter stats for 'system wide':
1841092 unc_m_pmm_rpq_inserts
3680816 unc_m_pmm_wpq_inserts
1.001775055 seconds time elapsed
root@lkp-csl-2sp2:~# perf stat -M UNC_M_PMM_READ_LATENCY -a -- sleep 1
Performance counter stats for 'system wide':
860649746 unc_m_pmm_rpq_occupancy.all
1840557 unc_m_pmm_rpq_inserts
12790627455 unc_m_clockticks
1.001773348 seconds time elapsed
No metrics 'UNC_M_PMM_BANDWIDTH.TOTAL' or 'UNC_M_PMM_READ_LATENCY' are
reported.
The issue is, the case of an alias expanding to mulitple events is not
supported, typically the uncore events. (see comments in
find_evsel_group()).
For UNC_M_PMM_BANDWIDTH.TOTAL in above example, the expanded event group
is '{unc_m_pmm_rpq_inserts,unc_m_pmm_wpq_inserts}:W', but the actual
events passed to find_evsel_group are:
unc_m_pmm_rpq_inserts
unc_m_pmm_rpq_inserts
unc_m_pmm_rpq_inserts
unc_m_pmm_rpq_inserts
unc_m_pmm_rpq_inserts
unc_m_pmm_rpq_inserts
unc_m_pmm_wpq_inserts
unc_m_pmm_wpq_inserts
unc_m_pmm_wpq_inserts
unc_m_pmm_wpq_inserts
unc_m_pmm_wpq_inserts
unc_m_pmm_wpq_inserts
For this multiple events case, it's not supported well.
This patch introduces a new field 'metric_leader' in struct evsel. The
first event is considered as a metric leader. For the rest of same
events, they point to the first event via it's metric_leader field in
struct evsel.
This design is for adding the counting results of all same events to the
first event in group (the metric_leader).
With this patch,
root@lkp-csl-2sp2:~# perf stat -M UNC_M_PMM_BANDWIDTH.TOTAL -a -- sleep 1
Performance counter stats for 'system wide':
1842108 unc_m_pmm_rpq_inserts # 337.2 MB/sec UNC_M_PMM_BANDWIDTH.TOTAL
3682209 unc_m_pmm_wpq_inserts
1.001819706 seconds time elapsed
root@lkp-csl-2sp2:~# perf stat -M UNC_M_PMM_READ_LATENCY -a -- sleep 1
Performance counter stats for 'system wide':
861970685 unc_m_pmm_rpq_occupancy.all # 219.4 ns UNC_M_PMM_READ_LATENCY
1842772 unc_m_pmm_rpq_inserts
12790196356 unc_m_clockticks
1.001749103 seconds time elapsed
Now we can see the correct metrics 'UNC_M_PMM_BANDWIDTH.TOTAL' and
'UNC_M_PMM_READ_LATENCY'.
Signed-off-by: Jin Yao <yao.jin@linux.intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Andi Kleen <ak@linux.intel.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: http://lore.kernel.org/lkml/20190828055932.8269-5-yao.jin@linux.intel.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Signed-off-by: hongrongxuan <hongrongxuan@huawei.com>
Conflicts:
tools/perf/util/metricgroup.c
---
tools/perf/util/evsel.h | 1 +
tools/perf/util/metricgroup.c | 90 ++++++++++++++++++-----------------
tools/perf/util/stat-shadow.c | 27 +++++++++--
3 files changed, 71 insertions(+), 47 deletions(-)
diff --git a/tools/perf/util/evsel.h b/tools/perf/util/evsel.h
index dab3930e2a8c..be63c73ca69c 100644
--- a/tools/perf/util/evsel.h
+++ b/tools/perf/util/evsel.h
@@ -149,6 +149,7 @@ struct perf_evsel {
const char * metric_expr;
const char * metric_name;
struct perf_evsel **metric_events;
+ struct perf_evsel *metric_leader;
bool collect_stat;
bool weak_group;
const char *pmu_name;
diff --git a/tools/perf/util/metricgroup.c b/tools/perf/util/metricgroup.c
index d01f16448bab..375866dab75b 100644
--- a/tools/perf/util/metricgroup.c
+++ b/tools/perf/util/metricgroup.c
@@ -96,57 +96,61 @@ struct egroup {
const char *metric_unit;
};
-static bool record_evsel(int *ind, struct perf_evsel **start,
- int idnum,
- struct perf_evsel **metric_events,
- struct perf_evsel *ev)
-{
- metric_events[*ind] = ev;
- if (*ind == 0)
- *start = ev;
- if (++*ind == idnum) {
- metric_events[*ind] = NULL;
- return true;
- }
- return false;
-}
-
static struct perf_evsel *find_evsel_group(struct perf_evlist *perf_evlist,
- const char **ids,
- int idnum,
- struct perf_evsel **metric_events)
+ const char **ids,
+ int idnum,
+ struct perf_evsel **metric_events)
{
- struct perf_evsel *ev, *start = NULL;
- int ind = 0;
+ struct perf_evsel *ev;
+ int i = 0;
+ bool leader_found;
evlist__for_each_entry (perf_evlist, ev) {
- if (ev->collect_stat)
- continue;
- if (!strcmp(ev->name, ids[ind])) {
- if (record_evsel(&ind, &start, idnum,
- metric_events, ev))
- return start;
+ if (!strcmp(ev->name, ids[i])) {
+ if (!metric_events[i])
+ metric_events[i] = ev;
} else {
- /*
- * We saw some other event that is not
- * in our list of events. Discard
- * the whole match and start again.
- */
- ind = 0;
- start = NULL;
- if (!strcmp(ev->name, ids[ind])) {
- if (record_evsel(&ind, &start, idnum,
- metric_events, ev))
- return start;
+ if (++i == idnum) {
+ /* Discard the whole match and start again */
+ i = 0;
+ memset(metric_events, 0,
+ sizeof(struct evsel *) * idnum);
+ continue;
+ }
+
+ if (!strcmp(ev->name, ids[i]))
+ metric_events[i] = ev;
+ else {
+ /* Discard the whole match and start again */
+ i = 0;
+ memset(metric_events, 0,
+ sizeof(struct evsel *) * idnum);
+ continue;
}
}
}
- /*
- * This can happen when an alias expands to multiple
- * events, like for uncore events.
- * We don't support this case for now.
- */
- return NULL;
+
+ if (i != idnum - 1) {
+ /* Not whole match */
+ return NULL;
+ }
+
+ metric_events[idnum] = NULL;
+
+ for (i = 0; i < idnum; i++) {
+ leader_found = false;
+ evlist__for_each_entry(perf_evlist, ev) {
+ if (!leader_found && (ev == metric_events[i]))
+ leader_found = true;
+
+ if (leader_found &&
+ !strcmp(ev->name, metric_events[i]->name)) {
+ ev->metric_leader = metric_events[i];
+ }
+ }
+ }
+
+ return metric_events[0];
}
static int metricgroup__setup_events(struct list_head *groups,
diff --git a/tools/perf/util/stat-shadow.c b/tools/perf/util/stat-shadow.c
index f687ca257e5d..a274f04cbeee 100644
--- a/tools/perf/util/stat-shadow.c
+++ b/tools/perf/util/stat-shadow.c
@@ -29,6 +29,8 @@ struct saved_value {
int cpu;
struct runtime_stat *stat;
struct stats stats;
+ u64 metric_total;
+ int metric_other;
};
static int saved_value_cmp(struct rb_node *rb_node, const void *entry)
@@ -210,6 +212,7 @@ void perf_stat__update_shadow_stats(struct perf_evsel *counter, u64 count,
{
int ctx = evsel_context(counter);
u64 count_ns = count;
+ struct saved_value *v;
count *= counter->scale;
@@ -264,9 +267,15 @@ void perf_stat__update_shadow_stats(struct perf_evsel *counter, u64 count,
update_runtime_stat(st, STAT_APERF, ctx, cpu, count);
if (counter->collect_stat) {
- struct saved_value *v = saved_value_lookup(counter, cpu, true,
- STAT_NONE, 0, st);
+ v = saved_value_lookup(counter, cpu, true, STAT_NONE, 0, st);
update_stats(&v->stats, count);
+ if (counter->metric_leader)
+ v->metric_total += count;
+ } else if (counter->metric_leader) {
+ v = saved_value_lookup(counter->metric_leader,
+ cpu, true, STAT_NONE, 0, st);
+ v->metric_total += count;
+ v->metric_other++;
}
}
@@ -727,10 +736,10 @@ static void generic_metric(struct perf_stat_config *config,
char *n, *pn;
expr__ctx_init(&pctx);
- expr__add_id(&pctx, name, avg);
for (i = 0; metric_events[i]; i++) {
struct saved_value *v;
struct stats *stats;
+ u64 metric_total = 0;
if (!strcmp(metric_events[i]->name, "duration_time")) {
stats = &walltime_nsecs_stats;
@@ -742,6 +751,9 @@ static void generic_metric(struct perf_stat_config *config,
break;
stats = &v->stats;
scale = 1.0;
+
+ if (v->metric_other)
+ metric_total = v->metric_total;
}
n = strdup(metric_events[i]->name);
@@ -755,8 +767,15 @@ static void generic_metric(struct perf_stat_config *config,
pn = strchr(n, ' ');
if (pn)
*pn = 0;
- expr__add_id(&pctx, n, avg_stats(stats)*scale);
+
+ if (metric_total)
+ expr__add_id(&pctx, n, metric_total);
+ else
+ expr__add_id(&pctx, n, avg_stats(stats)*scale);
}
+
+ expr__add_id(&pctx, name, avg);
+
if (!metric_events[i]) {
const char *p = metric_expr;
--
2.27.0
Reference in New Issue View Git Blame Copy Permalink