2020-03-17 19:02:16 +08:00
|
|
|
// SPDX-License-Identifier: GPL-2.0
|
perf test: Improve pmu event metric testing
Break pmu-events test into 2 and add a test to verify that all pmu
metric expressions simply parse. Try to parse all metric ids/events,
skip/warn if metrics for the current architecture fail to parse. To
support warning for a skip, and an ability for a subtest to describe why
it skips.
Tested on power9, skylakex, haswell, broadwell, westmere, sandybridge and
ivybridge.
May skip/warn on other architectures if metrics are invalid. In
particular s390 is untested, but its expressions are trivial. The
untested architectures with expressions are power8, cascadelakex,
tremontx, skylake, jaketown, ivytown and variants of haswell and
broadwell.
v3. addresses review comments from John Garry <john.garry@huawei.com>,
Jiri Olsa <jolsa@redhat.com> and Arnaldo Carvalho de Melo
<acme@kernel.org>.
v2. changes the commit message as event parsing errors no longer cause
the test to fail.
Committer notes:
Check the return value of strtod() to fix the build in systems where
that function is declared with attribute warn_unused_result.
Signed-off-by: Ian Rogers <irogers@google.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Andi Kleen <ak@linux.intel.com>
Cc: Jin Yao <yao.jin@linux.intel.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: John Garry <john.garry@huawei.com>
Cc: Kajol Jain <kjain@linux.ibm.com>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Leo Yan <leo.yan@linaro.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Paul Clarke <pc@us.ibm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lore.kernel.org/lkml/20200513212933.41273-1-irogers@google.com
[ split from a larger patch ]
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2020-05-14 05:29:33 +08:00
|
|
|
#include "math.h"
|
2020-03-17 19:02:16 +08:00
|
|
|
#include "parse-events.h"
|
|
|
|
#include "pmu.h"
|
|
|
|
#include "tests.h"
|
|
|
|
#include <errno.h>
|
|
|
|
#include <stdio.h>
|
|
|
|
#include <linux/kernel.h>
|
2020-03-17 19:02:19 +08:00
|
|
|
#include <linux/zalloc.h>
|
2020-03-17 19:02:16 +08:00
|
|
|
#include "debug.h"
|
|
|
|
#include "../pmu-events/pmu-events.h"
|
perf test: Improve pmu event metric testing
Break pmu-events test into 2 and add a test to verify that all pmu
metric expressions simply parse. Try to parse all metric ids/events,
skip/warn if metrics for the current architecture fail to parse. To
support warning for a skip, and an ability for a subtest to describe why
it skips.
Tested on power9, skylakex, haswell, broadwell, westmere, sandybridge and
ivybridge.
May skip/warn on other architectures if metrics are invalid. In
particular s390 is untested, but its expressions are trivial. The
untested architectures with expressions are power8, cascadelakex,
tremontx, skylake, jaketown, ivytown and variants of haswell and
broadwell.
v3. addresses review comments from John Garry <john.garry@huawei.com>,
Jiri Olsa <jolsa@redhat.com> and Arnaldo Carvalho de Melo
<acme@kernel.org>.
v2. changes the commit message as event parsing errors no longer cause
the test to fail.
Committer notes:
Check the return value of strtod() to fix the build in systems where
that function is declared with attribute warn_unused_result.
Signed-off-by: Ian Rogers <irogers@google.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Andi Kleen <ak@linux.intel.com>
Cc: Jin Yao <yao.jin@linux.intel.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: John Garry <john.garry@huawei.com>
Cc: Kajol Jain <kjain@linux.ibm.com>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Leo Yan <leo.yan@linaro.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Paul Clarke <pc@us.ibm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lore.kernel.org/lkml/20200513212933.41273-1-irogers@google.com
[ split from a larger patch ]
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2020-05-14 05:29:33 +08:00
|
|
|
#include "util/evlist.h"
|
|
|
|
#include "util/expr.h"
|
|
|
|
#include "util/parse-events.h"
|
2020-03-17 19:02:16 +08:00
|
|
|
|
|
|
|
struct perf_pmu_test_event {
|
|
|
|
struct pmu_event event;
|
2020-03-17 19:02:19 +08:00
|
|
|
|
|
|
|
/* extra events for aliases */
|
|
|
|
const char *alias_str;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Note: For when PublicDescription does not exist in the JSON, we
|
|
|
|
* will have no long_desc in pmu_event.long_desc, but long_desc may
|
|
|
|
* be set in the alias.
|
|
|
|
*/
|
|
|
|
const char *alias_long_desc;
|
2020-03-17 19:02:16 +08:00
|
|
|
};
|
2020-03-17 19:02:19 +08:00
|
|
|
|
2020-03-17 19:02:16 +08:00
|
|
|
static struct perf_pmu_test_event test_cpu_events[] = {
|
|
|
|
{
|
|
|
|
.event = {
|
|
|
|
.name = "bp_l1_btb_correct",
|
|
|
|
.event = "event=0x8a",
|
|
|
|
.desc = "L1 BTB Correction",
|
|
|
|
.topic = "branch",
|
|
|
|
},
|
2020-03-17 19:02:19 +08:00
|
|
|
.alias_str = "event=0x8a",
|
|
|
|
.alias_long_desc = "L1 BTB Correction",
|
2020-03-17 19:02:16 +08:00
|
|
|
},
|
|
|
|
{
|
|
|
|
.event = {
|
|
|
|
.name = "bp_l2_btb_correct",
|
|
|
|
.event = "event=0x8b",
|
|
|
|
.desc = "L2 BTB Correction",
|
|
|
|
.topic = "branch",
|
|
|
|
},
|
2020-03-17 19:02:19 +08:00
|
|
|
.alias_str = "event=0x8b",
|
|
|
|
.alias_long_desc = "L2 BTB Correction",
|
2020-03-17 19:02:16 +08:00
|
|
|
},
|
|
|
|
{
|
|
|
|
.event = {
|
|
|
|
.name = "segment_reg_loads.any",
|
|
|
|
.event = "umask=0x80,period=200000,event=0x6",
|
|
|
|
.desc = "Number of segment register loads",
|
|
|
|
.topic = "other",
|
|
|
|
},
|
2020-03-17 19:02:19 +08:00
|
|
|
.alias_str = "umask=0x80,(null)=0x30d40,event=0x6",
|
|
|
|
.alias_long_desc = "Number of segment register loads",
|
2020-03-17 19:02:16 +08:00
|
|
|
},
|
|
|
|
{
|
|
|
|
.event = {
|
|
|
|
.name = "dispatch_blocked.any",
|
|
|
|
.event = "umask=0x20,period=200000,event=0x9",
|
|
|
|
.desc = "Memory cluster signals to block micro-op dispatch for any reason",
|
|
|
|
.topic = "other",
|
|
|
|
},
|
2020-03-17 19:02:19 +08:00
|
|
|
.alias_str = "umask=0x20,(null)=0x30d40,event=0x9",
|
|
|
|
.alias_long_desc = "Memory cluster signals to block micro-op dispatch for any reason",
|
2020-03-17 19:02:16 +08:00
|
|
|
},
|
|
|
|
{
|
|
|
|
.event = {
|
|
|
|
.name = "eist_trans",
|
|
|
|
.event = "umask=0x0,period=200000,event=0x3a",
|
|
|
|
.desc = "Number of Enhanced Intel SpeedStep(R) Technology (EIST) transitions",
|
|
|
|
.topic = "other",
|
|
|
|
},
|
2020-03-17 19:02:19 +08:00
|
|
|
.alias_str = "umask=0,(null)=0x30d40,event=0x3a",
|
|
|
|
.alias_long_desc = "Number of Enhanced Intel SpeedStep(R) Technology (EIST) transitions",
|
2020-03-17 19:02:16 +08:00
|
|
|
},
|
|
|
|
{ /* sentinel */
|
|
|
|
.event = {
|
|
|
|
.name = NULL,
|
|
|
|
},
|
|
|
|
},
|
|
|
|
};
|
|
|
|
|
|
|
|
static struct perf_pmu_test_event test_uncore_events[] = {
|
|
|
|
{
|
|
|
|
.event = {
|
|
|
|
.name = "uncore_hisi_ddrc.flux_wcmd",
|
|
|
|
.event = "event=0x2",
|
|
|
|
.desc = "DDRC write commands. Unit: hisi_sccl,ddrc ",
|
|
|
|
.topic = "uncore",
|
|
|
|
.long_desc = "DDRC write commands",
|
|
|
|
.pmu = "hisi_sccl,ddrc",
|
|
|
|
},
|
2020-03-17 19:02:19 +08:00
|
|
|
.alias_str = "event=0x2",
|
|
|
|
.alias_long_desc = "DDRC write commands",
|
2020-03-17 19:02:16 +08:00
|
|
|
},
|
|
|
|
{
|
|
|
|
.event = {
|
|
|
|
.name = "unc_cbo_xsnp_response.miss_eviction",
|
|
|
|
.event = "umask=0x81,event=0x22",
|
|
|
|
.desc = "Unit: uncore_cbox A cross-core snoop resulted from L3 Eviction which misses in some processor core",
|
|
|
|
.topic = "uncore",
|
|
|
|
.long_desc = "A cross-core snoop resulted from L3 Eviction which misses in some processor core",
|
|
|
|
.pmu = "uncore_cbox",
|
|
|
|
},
|
2020-03-17 19:02:19 +08:00
|
|
|
.alias_str = "umask=0x81,event=0x22",
|
|
|
|
.alias_long_desc = "A cross-core snoop resulted from L3 Eviction which misses in some processor core",
|
2020-03-17 19:02:16 +08:00
|
|
|
},
|
|
|
|
{ /* sentinel */
|
|
|
|
.event = {
|
|
|
|
.name = NULL,
|
|
|
|
},
|
|
|
|
}
|
|
|
|
};
|
|
|
|
|
|
|
|
const int total_test_events_size = ARRAY_SIZE(test_uncore_events);
|
|
|
|
|
|
|
|
static bool is_same(const char *reference, const char *test)
|
|
|
|
{
|
|
|
|
if (!reference && !test)
|
|
|
|
return true;
|
|
|
|
|
|
|
|
if (reference && !test)
|
|
|
|
return false;
|
|
|
|
|
|
|
|
if (!reference && test)
|
|
|
|
return false;
|
|
|
|
|
|
|
|
return !strcmp(reference, test);
|
|
|
|
}
|
|
|
|
|
|
|
|
static struct pmu_events_map *__test_pmu_get_events_map(void)
|
|
|
|
{
|
|
|
|
struct pmu_events_map *map;
|
|
|
|
|
|
|
|
for (map = &pmu_events_map[0]; map->cpuid; map++) {
|
|
|
|
if (!strcmp(map->cpuid, "testcpu"))
|
|
|
|
return map;
|
|
|
|
}
|
|
|
|
|
|
|
|
pr_err("could not find test events map\n");
|
|
|
|
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Verify generated events from pmu-events.c is as expected */
|
perf test: Improve pmu event metric testing
Break pmu-events test into 2 and add a test to verify that all pmu
metric expressions simply parse. Try to parse all metric ids/events,
skip/warn if metrics for the current architecture fail to parse. To
support warning for a skip, and an ability for a subtest to describe why
it skips.
Tested on power9, skylakex, haswell, broadwell, westmere, sandybridge and
ivybridge.
May skip/warn on other architectures if metrics are invalid. In
particular s390 is untested, but its expressions are trivial. The
untested architectures with expressions are power8, cascadelakex,
tremontx, skylake, jaketown, ivytown and variants of haswell and
broadwell.
v3. addresses review comments from John Garry <john.garry@huawei.com>,
Jiri Olsa <jolsa@redhat.com> and Arnaldo Carvalho de Melo
<acme@kernel.org>.
v2. changes the commit message as event parsing errors no longer cause
the test to fail.
Committer notes:
Check the return value of strtod() to fix the build in systems where
that function is declared with attribute warn_unused_result.
Signed-off-by: Ian Rogers <irogers@google.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Andi Kleen <ak@linux.intel.com>
Cc: Jin Yao <yao.jin@linux.intel.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: John Garry <john.garry@huawei.com>
Cc: Kajol Jain <kjain@linux.ibm.com>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Leo Yan <leo.yan@linaro.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Paul Clarke <pc@us.ibm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lore.kernel.org/lkml/20200513212933.41273-1-irogers@google.com
[ split from a larger patch ]
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2020-05-14 05:29:33 +08:00
|
|
|
static int test_pmu_event_table(void)
|
2020-03-17 19:02:16 +08:00
|
|
|
{
|
|
|
|
struct pmu_events_map *map = __test_pmu_get_events_map();
|
|
|
|
struct pmu_event *table;
|
|
|
|
int map_events = 0, expected_events;
|
|
|
|
|
|
|
|
/* ignore 2x sentinels */
|
|
|
|
expected_events = ARRAY_SIZE(test_cpu_events) +
|
|
|
|
ARRAY_SIZE(test_uncore_events) - 2;
|
|
|
|
|
|
|
|
if (!map)
|
|
|
|
return -1;
|
|
|
|
|
|
|
|
for (table = map->table; table->name; table++) {
|
|
|
|
struct perf_pmu_test_event *test;
|
|
|
|
struct pmu_event *te;
|
|
|
|
bool found = false;
|
|
|
|
|
|
|
|
if (table->pmu)
|
|
|
|
test = &test_uncore_events[0];
|
|
|
|
else
|
|
|
|
test = &test_cpu_events[0];
|
|
|
|
|
|
|
|
te = &test->event;
|
|
|
|
|
|
|
|
for (; te->name; test++, te = &test->event) {
|
|
|
|
if (strcmp(table->name, te->name))
|
|
|
|
continue;
|
|
|
|
found = true;
|
|
|
|
map_events++;
|
|
|
|
|
|
|
|
if (!is_same(table->desc, te->desc)) {
|
|
|
|
pr_debug2("testing event table %s: mismatched desc, %s vs %s\n",
|
|
|
|
table->name, table->desc, te->desc);
|
|
|
|
return -1;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (!is_same(table->topic, te->topic)) {
|
|
|
|
pr_debug2("testing event table %s: mismatched topic, %s vs %s\n",
|
|
|
|
table->name, table->topic,
|
|
|
|
te->topic);
|
|
|
|
return -1;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (!is_same(table->long_desc, te->long_desc)) {
|
|
|
|
pr_debug2("testing event table %s: mismatched long_desc, %s vs %s\n",
|
|
|
|
table->name, table->long_desc,
|
|
|
|
te->long_desc);
|
|
|
|
return -1;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (!is_same(table->unit, te->unit)) {
|
|
|
|
pr_debug2("testing event table %s: mismatched unit, %s vs %s\n",
|
|
|
|
table->name, table->unit,
|
|
|
|
te->unit);
|
|
|
|
return -1;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (!is_same(table->perpkg, te->perpkg)) {
|
|
|
|
pr_debug2("testing event table %s: mismatched perpkg, %s vs %s\n",
|
|
|
|
table->name, table->perpkg,
|
|
|
|
te->perpkg);
|
|
|
|
return -1;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (!is_same(table->metric_expr, te->metric_expr)) {
|
|
|
|
pr_debug2("testing event table %s: mismatched metric_expr, %s vs %s\n",
|
|
|
|
table->name, table->metric_expr,
|
|
|
|
te->metric_expr);
|
|
|
|
return -1;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (!is_same(table->metric_name, te->metric_name)) {
|
|
|
|
pr_debug2("testing event table %s: mismatched metric_name, %s vs %s\n",
|
|
|
|
table->name, table->metric_name,
|
|
|
|
te->metric_name);
|
|
|
|
return -1;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (!is_same(table->deprecated, te->deprecated)) {
|
|
|
|
pr_debug2("testing event table %s: mismatched deprecated, %s vs %s\n",
|
|
|
|
table->name, table->deprecated,
|
|
|
|
te->deprecated);
|
|
|
|
return -1;
|
|
|
|
}
|
|
|
|
|
|
|
|
pr_debug("testing event table %s: pass\n", table->name);
|
|
|
|
}
|
|
|
|
|
|
|
|
if (!found) {
|
|
|
|
pr_err("testing event table: could not find event %s\n",
|
|
|
|
table->name);
|
|
|
|
return -1;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
if (map_events != expected_events) {
|
|
|
|
pr_err("testing event table: found %d, but expected %d\n",
|
|
|
|
map_events, expected_events);
|
|
|
|
return -1;
|
|
|
|
}
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
2020-03-17 19:02:19 +08:00
|
|
|
|
|
|
|
static struct perf_pmu_alias *find_alias(const char *test_event, struct list_head *aliases)
|
|
|
|
{
|
|
|
|
struct perf_pmu_alias *alias;
|
|
|
|
|
|
|
|
list_for_each_entry(alias, aliases, list)
|
|
|
|
if (!strcmp(test_event, alias->name))
|
|
|
|
return alias;
|
|
|
|
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Verify aliases are as expected */
|
|
|
|
static int __test__pmu_event_aliases(char *pmu_name, int *count)
|
|
|
|
{
|
|
|
|
struct perf_pmu_test_event *test;
|
|
|
|
struct pmu_event *te;
|
|
|
|
struct perf_pmu *pmu;
|
|
|
|
LIST_HEAD(aliases);
|
|
|
|
int res = 0;
|
|
|
|
bool use_uncore_table;
|
|
|
|
struct pmu_events_map *map = __test_pmu_get_events_map();
|
|
|
|
|
|
|
|
if (!map)
|
|
|
|
return -1;
|
|
|
|
|
|
|
|
if (is_pmu_core(pmu_name)) {
|
|
|
|
test = &test_cpu_events[0];
|
|
|
|
use_uncore_table = false;
|
|
|
|
} else {
|
|
|
|
test = &test_uncore_events[0];
|
|
|
|
use_uncore_table = true;
|
|
|
|
}
|
|
|
|
|
|
|
|
pmu = zalloc(sizeof(*pmu));
|
|
|
|
if (!pmu)
|
|
|
|
return -1;
|
|
|
|
|
|
|
|
pmu->name = pmu_name;
|
|
|
|
|
|
|
|
pmu_add_cpu_aliases_map(&aliases, pmu, map);
|
|
|
|
|
|
|
|
for (te = &test->event; te->name; test++, te = &test->event) {
|
|
|
|
struct perf_pmu_alias *alias = find_alias(te->name, &aliases);
|
|
|
|
|
|
|
|
if (!alias) {
|
|
|
|
bool uncore_match = pmu_uncore_alias_match(pmu_name,
|
|
|
|
te->pmu);
|
|
|
|
|
|
|
|
if (use_uncore_table && !uncore_match) {
|
|
|
|
pr_debug3("testing aliases PMU %s: skip matching alias %s\n",
|
|
|
|
pmu_name, te->name);
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
|
|
|
|
pr_debug2("testing aliases PMU %s: no alias, alias_table->name=%s\n",
|
|
|
|
pmu_name, te->name);
|
|
|
|
res = -1;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (!is_same(alias->desc, te->desc)) {
|
|
|
|
pr_debug2("testing aliases PMU %s: mismatched desc, %s vs %s\n",
|
|
|
|
pmu_name, alias->desc, te->desc);
|
|
|
|
res = -1;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (!is_same(alias->long_desc, test->alias_long_desc)) {
|
|
|
|
pr_debug2("testing aliases PMU %s: mismatched long_desc, %s vs %s\n",
|
|
|
|
pmu_name, alias->long_desc,
|
|
|
|
test->alias_long_desc);
|
|
|
|
res = -1;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (!is_same(alias->str, test->alias_str)) {
|
|
|
|
pr_debug2("testing aliases PMU %s: mismatched str, %s vs %s\n",
|
|
|
|
pmu_name, alias->str, test->alias_str);
|
|
|
|
res = -1;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (!is_same(alias->topic, te->topic)) {
|
|
|
|
pr_debug2("testing aliases PMU %s: mismatched topic, %s vs %s\n",
|
|
|
|
pmu_name, alias->topic, te->topic);
|
|
|
|
res = -1;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
(*count)++;
|
|
|
|
pr_debug2("testing aliases PMU %s: matched event %s\n",
|
|
|
|
pmu_name, alias->name);
|
|
|
|
}
|
|
|
|
|
|
|
|
free(pmu);
|
|
|
|
return res;
|
|
|
|
}
|
|
|
|
|
perf test: Improve pmu event metric testing
Break pmu-events test into 2 and add a test to verify that all pmu
metric expressions simply parse. Try to parse all metric ids/events,
skip/warn if metrics for the current architecture fail to parse. To
support warning for a skip, and an ability for a subtest to describe why
it skips.
Tested on power9, skylakex, haswell, broadwell, westmere, sandybridge and
ivybridge.
May skip/warn on other architectures if metrics are invalid. In
particular s390 is untested, but its expressions are trivial. The
untested architectures with expressions are power8, cascadelakex,
tremontx, skylake, jaketown, ivytown and variants of haswell and
broadwell.
v3. addresses review comments from John Garry <john.garry@huawei.com>,
Jiri Olsa <jolsa@redhat.com> and Arnaldo Carvalho de Melo
<acme@kernel.org>.
v2. changes the commit message as event parsing errors no longer cause
the test to fail.
Committer notes:
Check the return value of strtod() to fix the build in systems where
that function is declared with attribute warn_unused_result.
Signed-off-by: Ian Rogers <irogers@google.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Andi Kleen <ak@linux.intel.com>
Cc: Jin Yao <yao.jin@linux.intel.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: John Garry <john.garry@huawei.com>
Cc: Kajol Jain <kjain@linux.ibm.com>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Leo Yan <leo.yan@linaro.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Paul Clarke <pc@us.ibm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lore.kernel.org/lkml/20200513212933.41273-1-irogers@google.com
[ split from a larger patch ]
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2020-05-14 05:29:33 +08:00
|
|
|
|
|
|
|
static int test_aliases(void)
|
2020-03-17 19:02:16 +08:00
|
|
|
{
|
2020-03-17 19:02:19 +08:00
|
|
|
struct perf_pmu *pmu = NULL;
|
|
|
|
|
|
|
|
while ((pmu = perf_pmu__scan(pmu)) != NULL) {
|
|
|
|
int count = 0;
|
|
|
|
|
|
|
|
if (list_empty(&pmu->format)) {
|
|
|
|
pr_debug2("skipping testing PMU %s\n", pmu->name);
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (__test__pmu_event_aliases(pmu->name, &count)) {
|
|
|
|
pr_debug("testing PMU %s aliases: failed\n", pmu->name);
|
|
|
|
return -1;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (count == 0)
|
|
|
|
pr_debug3("testing PMU %s aliases: no events to match\n",
|
|
|
|
pmu->name);
|
|
|
|
else
|
|
|
|
pr_debug("testing PMU %s aliases: pass\n", pmu->name);
|
|
|
|
}
|
|
|
|
|
2020-03-17 19:02:16 +08:00
|
|
|
return 0;
|
|
|
|
}
|
perf test: Improve pmu event metric testing
Break pmu-events test into 2 and add a test to verify that all pmu
metric expressions simply parse. Try to parse all metric ids/events,
skip/warn if metrics for the current architecture fail to parse. To
support warning for a skip, and an ability for a subtest to describe why
it skips.
Tested on power9, skylakex, haswell, broadwell, westmere, sandybridge and
ivybridge.
May skip/warn on other architectures if metrics are invalid. In
particular s390 is untested, but its expressions are trivial. The
untested architectures with expressions are power8, cascadelakex,
tremontx, skylake, jaketown, ivytown and variants of haswell and
broadwell.
v3. addresses review comments from John Garry <john.garry@huawei.com>,
Jiri Olsa <jolsa@redhat.com> and Arnaldo Carvalho de Melo
<acme@kernel.org>.
v2. changes the commit message as event parsing errors no longer cause
the test to fail.
Committer notes:
Check the return value of strtod() to fix the build in systems where
that function is declared with attribute warn_unused_result.
Signed-off-by: Ian Rogers <irogers@google.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Andi Kleen <ak@linux.intel.com>
Cc: Jin Yao <yao.jin@linux.intel.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: John Garry <john.garry@huawei.com>
Cc: Kajol Jain <kjain@linux.ibm.com>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Leo Yan <leo.yan@linaro.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Paul Clarke <pc@us.ibm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lore.kernel.org/lkml/20200513212933.41273-1-irogers@google.com
[ split from a larger patch ]
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2020-05-14 05:29:33 +08:00
|
|
|
|
|
|
|
static bool is_number(const char *str)
|
|
|
|
{
|
|
|
|
char *end_ptr;
|
|
|
|
double v;
|
|
|
|
|
|
|
|
errno = 0;
|
|
|
|
v = strtod(str, &end_ptr);
|
|
|
|
(void)v; // We're not interested in this value, only if it is valid
|
|
|
|
return errno == 0 && end_ptr != str;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int check_parse_id(const char *id, bool same_cpu, struct pmu_event *pe)
|
|
|
|
{
|
|
|
|
struct parse_events_error error;
|
|
|
|
struct evlist *evlist;
|
|
|
|
int ret;
|
|
|
|
|
|
|
|
/* Numbers are always valid. */
|
|
|
|
if (is_number(id))
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
evlist = evlist__new();
|
|
|
|
memset(&error, 0, sizeof(error));
|
|
|
|
ret = parse_events(evlist, id, &error);
|
|
|
|
if (ret && same_cpu) {
|
|
|
|
pr_warning("Parse event failed metric '%s' id '%s' expr '%s'\n",
|
|
|
|
pe->metric_name, id, pe->metric_expr);
|
|
|
|
pr_warning("Error string '%s' help '%s'\n", error.str,
|
|
|
|
error.help);
|
|
|
|
} else if (ret) {
|
|
|
|
pr_debug3("Parse event failed, but for an event that may not be supported by this CPU.\nid '%s' metric '%s' expr '%s'\n",
|
|
|
|
id, pe->metric_name, pe->metric_expr);
|
|
|
|
ret = 0;
|
|
|
|
}
|
|
|
|
evlist__delete(evlist);
|
|
|
|
free(error.str);
|
|
|
|
free(error.help);
|
|
|
|
free(error.first_str);
|
|
|
|
free(error.first_help);
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void expr_failure(const char *msg,
|
|
|
|
const struct pmu_events_map *map,
|
|
|
|
const struct pmu_event *pe)
|
|
|
|
{
|
|
|
|
pr_debug("%s for map %s %s %s\n",
|
|
|
|
msg, map->cpuid, map->version, map->type);
|
|
|
|
pr_debug("On metric %s\n", pe->metric_name);
|
|
|
|
pr_debug("On expression %s\n", pe->metric_expr);
|
|
|
|
}
|
|
|
|
|
|
|
|
static int test_parsing(void)
|
|
|
|
{
|
|
|
|
struct pmu_events_map *cpus_map = perf_pmu__find_map(NULL);
|
|
|
|
struct pmu_events_map *map;
|
|
|
|
struct pmu_event *pe;
|
|
|
|
int i, j, k;
|
|
|
|
int ret = 0;
|
|
|
|
struct expr_parse_ctx ctx;
|
|
|
|
double result;
|
|
|
|
|
|
|
|
i = 0;
|
|
|
|
for (;;) {
|
|
|
|
map = &pmu_events_map[i++];
|
|
|
|
if (!map->table)
|
|
|
|
break;
|
|
|
|
j = 0;
|
|
|
|
for (;;) {
|
2020-05-16 06:17:32 +08:00
|
|
|
struct hashmap_entry *cur;
|
|
|
|
size_t bkt;
|
|
|
|
|
perf test: Improve pmu event metric testing
Break pmu-events test into 2 and add a test to verify that all pmu
metric expressions simply parse. Try to parse all metric ids/events,
skip/warn if metrics for the current architecture fail to parse. To
support warning for a skip, and an ability for a subtest to describe why
it skips.
Tested on power9, skylakex, haswell, broadwell, westmere, sandybridge and
ivybridge.
May skip/warn on other architectures if metrics are invalid. In
particular s390 is untested, but its expressions are trivial. The
untested architectures with expressions are power8, cascadelakex,
tremontx, skylake, jaketown, ivytown and variants of haswell and
broadwell.
v3. addresses review comments from John Garry <john.garry@huawei.com>,
Jiri Olsa <jolsa@redhat.com> and Arnaldo Carvalho de Melo
<acme@kernel.org>.
v2. changes the commit message as event parsing errors no longer cause
the test to fail.
Committer notes:
Check the return value of strtod() to fix the build in systems where
that function is declared with attribute warn_unused_result.
Signed-off-by: Ian Rogers <irogers@google.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Andi Kleen <ak@linux.intel.com>
Cc: Jin Yao <yao.jin@linux.intel.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: John Garry <john.garry@huawei.com>
Cc: Kajol Jain <kjain@linux.ibm.com>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Leo Yan <leo.yan@linaro.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Paul Clarke <pc@us.ibm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lore.kernel.org/lkml/20200513212933.41273-1-irogers@google.com
[ split from a larger patch ]
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2020-05-14 05:29:33 +08:00
|
|
|
pe = &map->table[j++];
|
|
|
|
if (!pe->name && !pe->metric_group && !pe->metric_name)
|
|
|
|
break;
|
|
|
|
if (!pe->metric_expr)
|
|
|
|
continue;
|
2020-05-16 06:17:32 +08:00
|
|
|
expr__ctx_init(&ctx);
|
|
|
|
if (expr__find_other(pe->metric_expr, NULL, &ctx, 0)
|
|
|
|
< 0) {
|
perf test: Improve pmu event metric testing
Break pmu-events test into 2 and add a test to verify that all pmu
metric expressions simply parse. Try to parse all metric ids/events,
skip/warn if metrics for the current architecture fail to parse. To
support warning for a skip, and an ability for a subtest to describe why
it skips.
Tested on power9, skylakex, haswell, broadwell, westmere, sandybridge and
ivybridge.
May skip/warn on other architectures if metrics are invalid. In
particular s390 is untested, but its expressions are trivial. The
untested architectures with expressions are power8, cascadelakex,
tremontx, skylake, jaketown, ivytown and variants of haswell and
broadwell.
v3. addresses review comments from John Garry <john.garry@huawei.com>,
Jiri Olsa <jolsa@redhat.com> and Arnaldo Carvalho de Melo
<acme@kernel.org>.
v2. changes the commit message as event parsing errors no longer cause
the test to fail.
Committer notes:
Check the return value of strtod() to fix the build in systems where
that function is declared with attribute warn_unused_result.
Signed-off-by: Ian Rogers <irogers@google.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Andi Kleen <ak@linux.intel.com>
Cc: Jin Yao <yao.jin@linux.intel.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: John Garry <john.garry@huawei.com>
Cc: Kajol Jain <kjain@linux.ibm.com>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Leo Yan <leo.yan@linaro.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Paul Clarke <pc@us.ibm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lore.kernel.org/lkml/20200513212933.41273-1-irogers@google.com
[ split from a larger patch ]
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2020-05-14 05:29:33 +08:00
|
|
|
expr_failure("Parse other failed", map, pe);
|
|
|
|
ret++;
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Add all ids with a made up value. The value may
|
|
|
|
* trigger divide by zero when subtracted and so try to
|
|
|
|
* make them unique.
|
|
|
|
*/
|
2020-05-16 06:17:32 +08:00
|
|
|
k = 1;
|
|
|
|
hashmap__for_each_entry((&ctx.ids), cur, bkt)
|
|
|
|
expr__add_id(&ctx, strdup(cur->key), k++);
|
perf test: Improve pmu event metric testing
Break pmu-events test into 2 and add a test to verify that all pmu
metric expressions simply parse. Try to parse all metric ids/events,
skip/warn if metrics for the current architecture fail to parse. To
support warning for a skip, and an ability for a subtest to describe why
it skips.
Tested on power9, skylakex, haswell, broadwell, westmere, sandybridge and
ivybridge.
May skip/warn on other architectures if metrics are invalid. In
particular s390 is untested, but its expressions are trivial. The
untested architectures with expressions are power8, cascadelakex,
tremontx, skylake, jaketown, ivytown and variants of haswell and
broadwell.
v3. addresses review comments from John Garry <john.garry@huawei.com>,
Jiri Olsa <jolsa@redhat.com> and Arnaldo Carvalho de Melo
<acme@kernel.org>.
v2. changes the commit message as event parsing errors no longer cause
the test to fail.
Committer notes:
Check the return value of strtod() to fix the build in systems where
that function is declared with attribute warn_unused_result.
Signed-off-by: Ian Rogers <irogers@google.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Andi Kleen <ak@linux.intel.com>
Cc: Jin Yao <yao.jin@linux.intel.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: John Garry <john.garry@huawei.com>
Cc: Kajol Jain <kjain@linux.ibm.com>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Leo Yan <leo.yan@linaro.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Paul Clarke <pc@us.ibm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lore.kernel.org/lkml/20200513212933.41273-1-irogers@google.com
[ split from a larger patch ]
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2020-05-14 05:29:33 +08:00
|
|
|
|
2020-05-16 06:17:32 +08:00
|
|
|
hashmap__for_each_entry((&ctx.ids), cur, bkt) {
|
|
|
|
if (check_parse_id(cur->key, map == cpus_map,
|
|
|
|
pe))
|
perf test: Improve pmu event metric testing
Break pmu-events test into 2 and add a test to verify that all pmu
metric expressions simply parse. Try to parse all metric ids/events,
skip/warn if metrics for the current architecture fail to parse. To
support warning for a skip, and an ability for a subtest to describe why
it skips.
Tested on power9, skylakex, haswell, broadwell, westmere, sandybridge and
ivybridge.
May skip/warn on other architectures if metrics are invalid. In
particular s390 is untested, but its expressions are trivial. The
untested architectures with expressions are power8, cascadelakex,
tremontx, skylake, jaketown, ivytown and variants of haswell and
broadwell.
v3. addresses review comments from John Garry <john.garry@huawei.com>,
Jiri Olsa <jolsa@redhat.com> and Arnaldo Carvalho de Melo
<acme@kernel.org>.
v2. changes the commit message as event parsing errors no longer cause
the test to fail.
Committer notes:
Check the return value of strtod() to fix the build in systems where
that function is declared with attribute warn_unused_result.
Signed-off-by: Ian Rogers <irogers@google.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Andi Kleen <ak@linux.intel.com>
Cc: Jin Yao <yao.jin@linux.intel.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: John Garry <john.garry@huawei.com>
Cc: Kajol Jain <kjain@linux.ibm.com>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Leo Yan <leo.yan@linaro.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Paul Clarke <pc@us.ibm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lore.kernel.org/lkml/20200513212933.41273-1-irogers@google.com
[ split from a larger patch ]
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2020-05-14 05:29:33 +08:00
|
|
|
ret++;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (expr__parse(&result, &ctx, pe->metric_expr, 0)) {
|
|
|
|
expr_failure("Parse failed", map, pe);
|
|
|
|
ret++;
|
|
|
|
}
|
2020-05-16 06:17:32 +08:00
|
|
|
expr__ctx_clear(&ctx);
|
perf test: Improve pmu event metric testing
Break pmu-events test into 2 and add a test to verify that all pmu
metric expressions simply parse. Try to parse all metric ids/events,
skip/warn if metrics for the current architecture fail to parse. To
support warning for a skip, and an ability for a subtest to describe why
it skips.
Tested on power9, skylakex, haswell, broadwell, westmere, sandybridge and
ivybridge.
May skip/warn on other architectures if metrics are invalid. In
particular s390 is untested, but its expressions are trivial. The
untested architectures with expressions are power8, cascadelakex,
tremontx, skylake, jaketown, ivytown and variants of haswell and
broadwell.
v3. addresses review comments from John Garry <john.garry@huawei.com>,
Jiri Olsa <jolsa@redhat.com> and Arnaldo Carvalho de Melo
<acme@kernel.org>.
v2. changes the commit message as event parsing errors no longer cause
the test to fail.
Committer notes:
Check the return value of strtod() to fix the build in systems where
that function is declared with attribute warn_unused_result.
Signed-off-by: Ian Rogers <irogers@google.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Andi Kleen <ak@linux.intel.com>
Cc: Jin Yao <yao.jin@linux.intel.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: John Garry <john.garry@huawei.com>
Cc: Kajol Jain <kjain@linux.ibm.com>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Leo Yan <leo.yan@linaro.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Paul Clarke <pc@us.ibm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lore.kernel.org/lkml/20200513212933.41273-1-irogers@google.com
[ split from a larger patch ]
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2020-05-14 05:29:33 +08:00
|
|
|
}
|
|
|
|
}
|
|
|
|
/* TODO: fail when not ok */
|
|
|
|
return ret == 0 ? TEST_OK : TEST_SKIP;
|
|
|
|
}
|
|
|
|
|
|
|
|
static const struct {
|
|
|
|
int (*func)(void);
|
|
|
|
const char *desc;
|
|
|
|
} pmu_events_testcase_table[] = {
|
|
|
|
{
|
|
|
|
.func = test_pmu_event_table,
|
|
|
|
.desc = "PMU event table sanity",
|
|
|
|
},
|
|
|
|
{
|
|
|
|
.func = test_aliases,
|
|
|
|
.desc = "PMU event map aliases",
|
|
|
|
},
|
|
|
|
{
|
|
|
|
.func = test_parsing,
|
|
|
|
.desc = "Parsing of PMU event table metrics",
|
|
|
|
},
|
|
|
|
};
|
|
|
|
|
|
|
|
const char *test__pmu_events_subtest_get_desc(int subtest)
|
|
|
|
{
|
|
|
|
if (subtest < 0 ||
|
|
|
|
subtest >= (int)ARRAY_SIZE(pmu_events_testcase_table))
|
|
|
|
return NULL;
|
|
|
|
return pmu_events_testcase_table[subtest].desc;
|
|
|
|
}
|
|
|
|
|
|
|
|
const char *test__pmu_events_subtest_skip_reason(int subtest)
|
|
|
|
{
|
|
|
|
if (subtest < 0 ||
|
|
|
|
subtest >= (int)ARRAY_SIZE(pmu_events_testcase_table))
|
|
|
|
return NULL;
|
|
|
|
if (pmu_events_testcase_table[subtest].func != test_parsing)
|
|
|
|
return NULL;
|
|
|
|
return "some metrics failed";
|
|
|
|
}
|
|
|
|
|
|
|
|
int test__pmu_events_subtest_get_nr(void)
|
|
|
|
{
|
|
|
|
return (int)ARRAY_SIZE(pmu_events_testcase_table);
|
|
|
|
}
|
|
|
|
|
|
|
|
int test__pmu_events(struct test *test __maybe_unused, int subtest)
|
|
|
|
{
|
|
|
|
if (subtest < 0 ||
|
|
|
|
subtest >= (int)ARRAY_SIZE(pmu_events_testcase_table))
|
|
|
|
return TEST_FAIL;
|
|
|
|
return pmu_events_testcase_table[subtest].func();
|
|
|
|
}
|