summaryrefslogtreecommitdiff
path: root/tools/perf/ui/hist.c
diff options
context:
space:
mode:
Diffstat (limited to 'tools/perf/ui/hist.c')
-rw-r--r--tools/perf/ui/hist.c303
1 files changed, 202 insertions, 101 deletions
diff --git a/tools/perf/ui/hist.c b/tools/perf/ui/hist.c
index 2bf959d08354..ae3b7fe1dadc 100644
--- a/tools/perf/ui/hist.c
+++ b/tools/perf/ui/hist.c
@@ -23,85 +23,90 @@
__ret; \
})
+static int __hpp__fmt_print(struct perf_hpp *hpp, struct hists *hists, u64 val,
+ int nr_samples, const char *fmt, int len,
+ hpp_snprint_fn print_fn, enum perf_hpp_fmt_type fmtype)
+{
+ if (fmtype == PERF_HPP_FMT_TYPE__PERCENT || fmtype == PERF_HPP_FMT_TYPE__LATENCY) {
+ double percent = 0.0;
+ u64 total = fmtype == PERF_HPP_FMT_TYPE__PERCENT ? hists__total_period(hists) :
+ hists__total_latency(hists);
+
+ if (total)
+ percent = 100.0 * val / total;
+
+ return hpp__call_print_fn(hpp, print_fn, fmt, len, percent);
+ }
+
+ if (fmtype == PERF_HPP_FMT_TYPE__AVERAGE) {
+ double avg = nr_samples ? (1.0 * val / nr_samples) : 0;
+
+ return hpp__call_print_fn(hpp, print_fn, fmt, len, avg);
+ }
+
+ return hpp__call_print_fn(hpp, print_fn, fmt, len, val);
+}
+
+struct hpp_fmt_value {
+ struct hists *hists;
+ u64 val;
+ int samples;
+};
+
static int __hpp__fmt(struct perf_hpp *hpp, struct hist_entry *he,
hpp_field_fn get_field, const char *fmt, int len,
- hpp_snprint_fn print_fn, bool fmt_percent)
+ hpp_snprint_fn print_fn, enum perf_hpp_fmt_type fmtype)
{
- int ret;
+ int ret = 0;
struct hists *hists = he->hists;
struct evsel *evsel = hists_to_evsel(hists);
+ struct evsel *pos;
char *buf = hpp->buf;
size_t size = hpp->size;
+ int i = 0, nr_members = 1;
+ struct hpp_fmt_value *values;
- if (fmt_percent) {
- double percent = 0.0;
- u64 total = hists__total_period(hists);
+ if (evsel__is_group_event(evsel))
+ nr_members = evsel->core.nr_members;
- if (total)
- percent = 100.0 * get_field(he) / total;
+ values = calloc(nr_members, sizeof(*values));
+ if (values == NULL)
+ return 0;
- ret = hpp__call_print_fn(hpp, print_fn, fmt, len, percent);
- } else
- ret = hpp__call_print_fn(hpp, print_fn, fmt, len, get_field(he));
+ values[0].hists = evsel__hists(evsel);
+ values[0].val = get_field(he);
+ values[0].samples = he->stat.nr_events;
if (evsel__is_group_event(evsel)) {
- int prev_idx, idx_delta;
struct hist_entry *pair;
- int nr_members = evsel->core.nr_members;
- prev_idx = evsel__group_idx(evsel);
+ for_each_group_member(pos, evsel)
+ values[++i].hists = evsel__hists(pos);
list_for_each_entry(pair, &he->pairs.head, pairs.node) {
- u64 period = get_field(pair);
- u64 total = hists__total_period(pair->hists);
+ for (i = 0; i < nr_members; i++) {
+ if (values[i].hists != pair->hists)
+ continue;
- if (!total)
- continue;
-
- evsel = hists_to_evsel(pair->hists);
- idx_delta = evsel__group_idx(evsel) - prev_idx - 1;
-
- while (idx_delta--) {
- /*
- * zero-fill group members in the middle which
- * have no sample
- */
- if (fmt_percent) {
- ret += hpp__call_print_fn(hpp, print_fn,
- fmt, len, 0.0);
- } else {
- ret += hpp__call_print_fn(hpp, print_fn,
- fmt, len, 0ULL);
- }
+ values[i].val = get_field(pair);
+ values[i].samples = pair->stat.nr_events;
+ break;
}
-
- if (fmt_percent) {
- ret += hpp__call_print_fn(hpp, print_fn, fmt, len,
- 100.0 * period / total);
- } else {
- ret += hpp__call_print_fn(hpp, print_fn, fmt,
- len, period);
- }
-
- prev_idx = evsel__group_idx(evsel);
}
+ }
- idx_delta = nr_members - prev_idx - 1;
-
- while (idx_delta--) {
- /*
- * zero-fill group members at last which have no sample
- */
- if (fmt_percent) {
- ret += hpp__call_print_fn(hpp, print_fn,
- fmt, len, 0.0);
- } else {
- ret += hpp__call_print_fn(hpp, print_fn,
- fmt, len, 0ULL);
- }
- }
+ for (i = 0; i < nr_members; i++) {
+ if (symbol_conf.skip_empty &&
+ values[i].hists->stats.nr_samples == 0)
+ continue;
+
+ ret += __hpp__fmt_print(hpp, values[i].hists, values[i].val,
+ values[i].samples, fmt, len,
+ print_fn, fmtype);
}
+ free(values);
+
/*
* Restore original buf and size as it's where caller expects
* the result will be saved.
@@ -114,33 +119,35 @@ static int __hpp__fmt(struct perf_hpp *hpp, struct hist_entry *he,
int hpp__fmt(struct perf_hpp_fmt *fmt, struct perf_hpp *hpp,
struct hist_entry *he, hpp_field_fn get_field,
- const char *fmtstr, hpp_snprint_fn print_fn, bool fmt_percent)
+ const char *fmtstr, hpp_snprint_fn print_fn,
+ enum perf_hpp_fmt_type fmtype)
{
- int len = fmt->user_len ?: fmt->len;
+ int len = max(fmt->user_len ?: fmt->len, (int)strlen(fmt->name));
if (symbol_conf.field_sep) {
return __hpp__fmt(hpp, he, get_field, fmtstr, 1,
- print_fn, fmt_percent);
+ print_fn, fmtype);
}
- if (fmt_percent)
+ if (fmtype == PERF_HPP_FMT_TYPE__PERCENT || fmtype == PERF_HPP_FMT_TYPE__LATENCY)
len -= 2; /* 2 for a space and a % sign */
else
len -= 1;
- return __hpp__fmt(hpp, he, get_field, fmtstr, len, print_fn, fmt_percent);
+ return __hpp__fmt(hpp, he, get_field, fmtstr, len, print_fn, fmtype);
}
int hpp__fmt_acc(struct perf_hpp_fmt *fmt, struct perf_hpp *hpp,
struct hist_entry *he, hpp_field_fn get_field,
- const char *fmtstr, hpp_snprint_fn print_fn, bool fmt_percent)
+ const char *fmtstr, hpp_snprint_fn print_fn,
+ enum perf_hpp_fmt_type fmtype)
{
if (!symbol_conf.cumulate_callchain) {
int len = fmt->user_len ?: fmt->len;
return snprintf(hpp->buf, hpp->size, " %*s", len - 1, "N/A");
}
- return hpp__fmt(fmt, hpp, he, get_field, fmtstr, print_fn, fmt_percent);
+ return hpp__fmt(fmt, hpp, he, get_field, fmtstr, print_fn, fmtype);
}
static int field_cmp(u64 field_a, u64 field_b)
@@ -294,8 +301,18 @@ static int hpp__width_fn(struct perf_hpp_fmt *fmt,
int len = fmt->user_len ?: fmt->len;
struct evsel *evsel = hists_to_evsel(hists);
- if (symbol_conf.event_group)
- len = max(len, evsel->core.nr_members * fmt->len);
+ if (symbol_conf.event_group) {
+ int nr = 0;
+ struct evsel *pos;
+
+ for_each_group_evsel(pos, evsel) {
+ if (!symbol_conf.skip_empty ||
+ evsel__hists(pos)->stats.nr_samples)
+ nr++;
+ }
+
+ len = max(len, nr * fmt->len);
+ }
if (len < (int)strlen(fmt->name))
len = strlen(fmt->name);
@@ -340,7 +357,7 @@ static int hpp_entry_scnprintf(struct perf_hpp *hpp, const char *fmt, ...)
return (ret >= ssize) ? (ssize - 1) : ret;
}
-#define __HPP_COLOR_PERCENT_FN(_type, _field) \
+#define __HPP_COLOR_PERCENT_FN(_type, _field, _fmttype) \
static u64 he_get_##_field(struct hist_entry *he) \
{ \
return he->stat._field; \
@@ -350,15 +367,15 @@ static int hpp__color_##_type(struct perf_hpp_fmt *fmt, \
struct perf_hpp *hpp, struct hist_entry *he) \
{ \
return hpp__fmt(fmt, hpp, he, he_get_##_field, " %*.2f%%", \
- hpp_color_scnprintf, true); \
+ hpp_color_scnprintf, _fmttype); \
}
-#define __HPP_ENTRY_PERCENT_FN(_type, _field) \
+#define __HPP_ENTRY_PERCENT_FN(_type, _field, _fmttype) \
static int hpp__entry_##_type(struct perf_hpp_fmt *fmt, \
struct perf_hpp *hpp, struct hist_entry *he) \
{ \
return hpp__fmt(fmt, hpp, he, he_get_##_field, " %*.2f%%", \
- hpp_entry_scnprintf, true); \
+ hpp_entry_scnprintf, _fmttype); \
}
#define __HPP_SORT_FN(_type, _field) \
@@ -368,7 +385,7 @@ static int64_t hpp__sort_##_type(struct perf_hpp_fmt *fmt __maybe_unused, \
return __hpp__sort(a, b, he_get_##_field); \
}
-#define __HPP_COLOR_ACC_PERCENT_FN(_type, _field) \
+#define __HPP_COLOR_ACC_PERCENT_FN(_type, _field, _fmttype) \
static u64 he_get_acc_##_field(struct hist_entry *he) \
{ \
return he->stat_acc->_field; \
@@ -378,15 +395,15 @@ static int hpp__color_##_type(struct perf_hpp_fmt *fmt, \
struct perf_hpp *hpp, struct hist_entry *he) \
{ \
return hpp__fmt_acc(fmt, hpp, he, he_get_acc_##_field, " %*.2f%%", \
- hpp_color_scnprintf, true); \
+ hpp_color_scnprintf, _fmttype); \
}
-#define __HPP_ENTRY_ACC_PERCENT_FN(_type, _field) \
+#define __HPP_ENTRY_ACC_PERCENT_FN(_type, _field, _fmttype) \
static int hpp__entry_##_type(struct perf_hpp_fmt *fmt, \
struct perf_hpp *hpp, struct hist_entry *he) \
{ \
return hpp__fmt_acc(fmt, hpp, he, he_get_acc_##_field, " %*.2f%%", \
- hpp_entry_scnprintf, true); \
+ hpp_entry_scnprintf, _fmttype); \
}
#define __HPP_SORT_ACC_FN(_type, _field) \
@@ -406,7 +423,7 @@ static int hpp__entry_##_type(struct perf_hpp_fmt *fmt, \
struct perf_hpp *hpp, struct hist_entry *he) \
{ \
return hpp__fmt(fmt, hpp, he, he_get_raw_##_field, " %*"PRIu64, \
- hpp_entry_scnprintf, false); \
+ hpp_entry_scnprintf, PERF_HPP_FMT_TYPE__RAW); \
}
#define __HPP_SORT_RAW_FN(_type, _field) \
@@ -416,31 +433,61 @@ static int64_t hpp__sort_##_type(struct perf_hpp_fmt *fmt __maybe_unused, \
return __hpp__sort(a, b, he_get_raw_##_field); \
}
+#define __HPP_ENTRY_AVERAGE_FN(_type, _field) \
+static u64 he_get_##_field(struct hist_entry *he) \
+{ \
+ return he->stat._field; \
+} \
+ \
+static int hpp__entry_##_type(struct perf_hpp_fmt *fmt, \
+ struct perf_hpp *hpp, struct hist_entry *he) \
+{ \
+ return hpp__fmt(fmt, hpp, he, he_get_##_field, " %*.1f", \
+ hpp_entry_scnprintf, PERF_HPP_FMT_TYPE__AVERAGE); \
+}
-#define HPP_PERCENT_FNS(_type, _field) \
-__HPP_COLOR_PERCENT_FN(_type, _field) \
-__HPP_ENTRY_PERCENT_FN(_type, _field) \
+#define __HPP_SORT_AVERAGE_FN(_type, _field) \
+static int64_t hpp__sort_##_type(struct perf_hpp_fmt *fmt __maybe_unused, \
+ struct hist_entry *a, struct hist_entry *b) \
+{ \
+ return __hpp__sort(a, b, he_get_##_field); \
+}
+
+
+#define HPP_PERCENT_FNS(_type, _field, _fmttype) \
+__HPP_COLOR_PERCENT_FN(_type, _field, _fmttype) \
+__HPP_ENTRY_PERCENT_FN(_type, _field, _fmttype) \
__HPP_SORT_FN(_type, _field)
-#define HPP_PERCENT_ACC_FNS(_type, _field) \
-__HPP_COLOR_ACC_PERCENT_FN(_type, _field) \
-__HPP_ENTRY_ACC_PERCENT_FN(_type, _field) \
+#define HPP_PERCENT_ACC_FNS(_type, _field, _fmttype) \
+__HPP_COLOR_ACC_PERCENT_FN(_type, _field, _fmttype) \
+__HPP_ENTRY_ACC_PERCENT_FN(_type, _field, _fmttype) \
__HPP_SORT_ACC_FN(_type, _field)
#define HPP_RAW_FNS(_type, _field) \
__HPP_ENTRY_RAW_FN(_type, _field) \
__HPP_SORT_RAW_FN(_type, _field)
-HPP_PERCENT_FNS(overhead, period)
-HPP_PERCENT_FNS(overhead_sys, period_sys)
-HPP_PERCENT_FNS(overhead_us, period_us)
-HPP_PERCENT_FNS(overhead_guest_sys, period_guest_sys)
-HPP_PERCENT_FNS(overhead_guest_us, period_guest_us)
-HPP_PERCENT_ACC_FNS(overhead_acc, period)
+#define HPP_AVERAGE_FNS(_type, _field) \
+__HPP_ENTRY_AVERAGE_FN(_type, _field) \
+__HPP_SORT_AVERAGE_FN(_type, _field)
+
+HPP_PERCENT_FNS(overhead, period, PERF_HPP_FMT_TYPE__PERCENT)
+HPP_PERCENT_FNS(latency, latency, PERF_HPP_FMT_TYPE__LATENCY)
+HPP_PERCENT_FNS(overhead_sys, period_sys, PERF_HPP_FMT_TYPE__PERCENT)
+HPP_PERCENT_FNS(overhead_us, period_us, PERF_HPP_FMT_TYPE__PERCENT)
+HPP_PERCENT_FNS(overhead_guest_sys, period_guest_sys, PERF_HPP_FMT_TYPE__PERCENT)
+HPP_PERCENT_FNS(overhead_guest_us, period_guest_us, PERF_HPP_FMT_TYPE__PERCENT)
+HPP_PERCENT_ACC_FNS(overhead_acc, period, PERF_HPP_FMT_TYPE__PERCENT)
+HPP_PERCENT_ACC_FNS(latency_acc, latency, PERF_HPP_FMT_TYPE__LATENCY)
HPP_RAW_FNS(samples, nr_events)
HPP_RAW_FNS(period, period)
+HPP_AVERAGE_FNS(weight1, weight1)
+HPP_AVERAGE_FNS(weight2, weight2)
+HPP_AVERAGE_FNS(weight3, weight3)
+
static int64_t hpp__nop_cmp(struct perf_hpp_fmt *fmt __maybe_unused,
struct hist_entry *a __maybe_unused,
struct hist_entry *b __maybe_unused)
@@ -504,13 +551,18 @@ static bool hpp__equal(struct perf_hpp_fmt *a, struct perf_hpp_fmt *b)
struct perf_hpp_fmt perf_hpp__format[] = {
HPP__COLOR_PRINT_FNS("Overhead", overhead, OVERHEAD),
+ HPP__COLOR_PRINT_FNS("Latency", latency, LATENCY),
HPP__COLOR_PRINT_FNS("sys", overhead_sys, OVERHEAD_SYS),
HPP__COLOR_PRINT_FNS("usr", overhead_us, OVERHEAD_US),
HPP__COLOR_PRINT_FNS("guest sys", overhead_guest_sys, OVERHEAD_GUEST_SYS),
HPP__COLOR_PRINT_FNS("guest usr", overhead_guest_us, OVERHEAD_GUEST_US),
HPP__COLOR_ACC_PRINT_FNS("Children", overhead_acc, OVERHEAD_ACC),
+ HPP__COLOR_ACC_PRINT_FNS("Latency", latency_acc, LATENCY_ACC),
HPP__PRINT_FNS("Samples", samples, SAMPLES),
- HPP__PRINT_FNS("Period", period, PERIOD)
+ HPP__PRINT_FNS("Period", period, PERIOD),
+ HPP__PRINT_FNS("Weight1", weight1, WEIGHT1),
+ HPP__PRINT_FNS("Weight2", weight2, WEIGHT2),
+ HPP__PRINT_FNS("Weight3", weight3, WEIGHT3),
};
struct perf_hpp_list perf_hpp_list = {
@@ -526,6 +578,7 @@ struct perf_hpp_list perf_hpp_list = {
#undef HPP_PERCENT_FNS
#undef HPP_PERCENT_ACC_FNS
#undef HPP_RAW_FNS
+#undef HPP_AVERAGE_FNS
#undef __HPP_HEADER_FN
#undef __HPP_WIDTH_FN
@@ -534,9 +587,11 @@ struct perf_hpp_list perf_hpp_list = {
#undef __HPP_COLOR_ACC_PERCENT_FN
#undef __HPP_ENTRY_ACC_PERCENT_FN
#undef __HPP_ENTRY_RAW_FN
+#undef __HPP_ENTRY_AVERAGE_FN
#undef __HPP_SORT_FN
#undef __HPP_SORT_ACC_FN
#undef __HPP_SORT_RAW_FN
+#undef __HPP_SORT_AVERAGE_FN
static void fmt_free(struct perf_hpp_fmt *fmt)
{
@@ -551,6 +606,11 @@ static void fmt_free(struct perf_hpp_fmt *fmt)
fmt->free(fmt);
}
+static bool fmt_equal(struct perf_hpp_fmt *a, struct perf_hpp_fmt *b)
+{
+ return a->equal && a->equal(a, b);
+}
+
void perf_hpp__init(void)
{
int i;
@@ -571,28 +631,48 @@ void perf_hpp__init(void)
if (is_strict_order(field_order))
return;
+ /*
+ * Overhead and latency columns are added in setup_overhead(),
+ * so they are added implicitly here only if they were added
+ * by setup_overhead() before (have was_taken flag set).
+ * This is required because setup_overhead() has more complex
+ * logic, in particular it does not add "overhead" if user
+ * specified "latency" in sort order, and vise versa.
+ */
if (symbol_conf.cumulate_callchain) {
- hpp_dimension__add_output(PERF_HPP__OVERHEAD_ACC);
+ /*
+ * Addition of fields is idempotent, so we add latency
+ * column twice to get desired order with simpler logic.
+ */
+ if (symbol_conf.prefer_latency)
+ hpp_dimension__add_output(PERF_HPP__LATENCY_ACC, true);
+ hpp_dimension__add_output(PERF_HPP__OVERHEAD_ACC, true);
+ if (symbol_conf.enable_latency)
+ hpp_dimension__add_output(PERF_HPP__LATENCY_ACC, true);
perf_hpp__format[PERF_HPP__OVERHEAD].name = "Self";
}
- hpp_dimension__add_output(PERF_HPP__OVERHEAD);
+ if (symbol_conf.prefer_latency)
+ hpp_dimension__add_output(PERF_HPP__LATENCY, true);
+ hpp_dimension__add_output(PERF_HPP__OVERHEAD, true);
+ if (symbol_conf.enable_latency)
+ hpp_dimension__add_output(PERF_HPP__LATENCY, true);
if (symbol_conf.show_cpu_utilization) {
- hpp_dimension__add_output(PERF_HPP__OVERHEAD_SYS);
- hpp_dimension__add_output(PERF_HPP__OVERHEAD_US);
+ hpp_dimension__add_output(PERF_HPP__OVERHEAD_SYS, false);
+ hpp_dimension__add_output(PERF_HPP__OVERHEAD_US, false);
if (perf_guest) {
- hpp_dimension__add_output(PERF_HPP__OVERHEAD_GUEST_SYS);
- hpp_dimension__add_output(PERF_HPP__OVERHEAD_GUEST_US);
+ hpp_dimension__add_output(PERF_HPP__OVERHEAD_GUEST_SYS, false);
+ hpp_dimension__add_output(PERF_HPP__OVERHEAD_GUEST_US, false);
}
}
if (symbol_conf.show_nr_samples)
- hpp_dimension__add_output(PERF_HPP__SAMPLES);
+ hpp_dimension__add_output(PERF_HPP__SAMPLES, false);
if (symbol_conf.show_total_period)
- hpp_dimension__add_output(PERF_HPP__PERIOD);
+ hpp_dimension__add_output(PERF_HPP__PERIOD, false);
}
void perf_hpp_list__column_register(struct perf_hpp_list *list,
@@ -621,28 +701,42 @@ static void perf_hpp__column_unregister(struct perf_hpp_fmt *format)
void perf_hpp__cancel_cumulate(void)
{
- struct perf_hpp_fmt *fmt, *acc, *ovh, *tmp;
+ struct perf_hpp_fmt *fmt, *acc, *ovh, *acc_lat, *tmp;
if (is_strict_order(field_order))
return;
ovh = &perf_hpp__format[PERF_HPP__OVERHEAD];
acc = &perf_hpp__format[PERF_HPP__OVERHEAD_ACC];
+ acc_lat = &perf_hpp__format[PERF_HPP__LATENCY_ACC];
perf_hpp_list__for_each_format_safe(&perf_hpp_list, fmt, tmp) {
- if (acc->equal(acc, fmt)) {
+ if (fmt_equal(acc, fmt) || fmt_equal(acc_lat, fmt)) {
perf_hpp__column_unregister(fmt);
continue;
}
- if (ovh->equal(ovh, fmt))
+ if (fmt_equal(ovh, fmt))
fmt->name = "Overhead";
}
}
-static bool fmt_equal(struct perf_hpp_fmt *a, struct perf_hpp_fmt *b)
+void perf_hpp__cancel_latency(void)
{
- return a->equal && a->equal(a, b);
+ struct perf_hpp_fmt *fmt, *lat, *acc, *tmp;
+
+ if (is_strict_order(field_order))
+ return;
+ if (sort_order && strstr(sort_order, "latency"))
+ return;
+
+ lat = &perf_hpp__format[PERF_HPP__LATENCY];
+ acc = &perf_hpp__format[PERF_HPP__LATENCY_ACC];
+
+ perf_hpp_list__for_each_format_safe(&perf_hpp_list, fmt, tmp) {
+ if (fmt_equal(lat, fmt) || fmt_equal(acc, fmt))
+ perf_hpp__column_unregister(fmt);
+ }
}
void perf_hpp__setup_output_field(struct perf_hpp_list *list)
@@ -769,6 +863,7 @@ void perf_hpp__reset_width(struct perf_hpp_fmt *fmt, struct hists *hists)
switch (fmt->idx) {
case PERF_HPP__OVERHEAD:
+ case PERF_HPP__LATENCY:
case PERF_HPP__OVERHEAD_SYS:
case PERF_HPP__OVERHEAD_US:
case PERF_HPP__OVERHEAD_ACC:
@@ -785,6 +880,12 @@ void perf_hpp__reset_width(struct perf_hpp_fmt *fmt, struct hists *hists)
fmt->len = 12;
break;
+ case PERF_HPP__WEIGHT1:
+ case PERF_HPP__WEIGHT2:
+ case PERF_HPP__WEIGHT3:
+ fmt->len = 8;
+ break;
+
default:
break;
}
generated by cgit (git 2.34.1) at 2025-07-23 10:02:39 +0000