perf stat: Include PMU name and split uncore events per PMU in metric-only JSON output
PATCH 1/1 perf stat: Include PMU name and split uncore events per PMU in metric-only JSON output
When running perf stat with -A (--no-aggr) and --metric-only in JSON output mode (-j), perf stat evaluates metric expressions across all matching PMUs (including uncore PMUs like uncore_iio_0, uncore_iio_1, etc.).
However, perf stat previously formatted JSON output by printing only "cpu" : "<id>" and grouping all metric values on a single line per CPU without identifying which PMU instance evaluated each metric. As a result, when an uncore event spans multiple PMU boxes, perf stat printed repeated, ambiguous metric keys without PMU names.
Fix this by:
1. Including "pmu" : "<pmu_name>" in print_aggr_id_json when evsel->pmu is a non-core or hybrid PMU in AGGR_NONE mode (-A).
2. Starting a new JSON metric line in AGGR_NONE mode (-A) whenever the underlying PMU instance changes across PMU events.
3. Updating perf_json_output_lint.py to recognize the new "pmu" key in the JSON test suite.
Example output:
$ perf stat -M iio_bandwidth_read -a -A --metric-only -j -I 1000
{"interval" : 1.000314908, "cpu" : "0", "pmu" : "uncore_iio_0", "MB/s iio_bandwidth_read" : "0.0"}
{"interval" : 1.000314908, "cpu" : "0", "pmu" : "uncore_iio_1", "MB/s iio_bandwidth_read" : "0.1"}
{"interval" : 1.000314908, "cpu" : "0", "pmu" : "uncore_iio_11", "MB/s iio_bandwidth_read" : "0.0"}
...
{"interval" : 1.000314908, "cpu" : "56", "pmu" : "uncore_iio_0", "MB/s iio_bandwidth_read" : "0.0"}
{"interval" : 1.000314908, "cpu" : "56", "pmu" : "uncore_iio_1", "MB/s iio_bandwidth_read" : "0.0"}
{"interval" : 1.000314908, "cpu" : "56", "pmu" : "uncore_iio_11", "MB/s iio_bandwidth_read" : "0.0"}--- .../tests/shell/lib/perf_json_output_lint.py | 1 + tools/perf/util/stat-display.c | 41 +++++++++++++------ 2 files changed, 30 insertions(+), 12 deletions(-)
tools/perf/tests/shell/lib/perf_json_output_lint.py +1 -0
@@ -64,6 +64,7 @@ def check_json_output(expected_items): 'metric-threshold': lambda x: x in ['unknown', 'good', 'less good', 'nearly bad', 'bad'], 'metricgroup': lambda x: True, 'node': lambda x: True, + 'pmu': lambda x: True, 'pcnt-running': lambda x: isfloat(x), 'socket': lambda x: True, 'thread': lambda x: True,
tools/perf/util/stat-display.c +29 -12
105 lines not on this page · raw message
Signature · 1 line
2.55.0.141.g00534a21ce-goog
[email protected]
AI reviewThank you for your contribution! Sashiko AI review found 2 potential issue(s) to consider:
Thank you for your contribution! Sashiko AI review found 2 potential issue(s) to consider:
New issues:
- [Medium] Bypassing print_metric_begin for tool events produces invalid, corrupted JSON and CSV output in --metric-only mode.
Pre-existing issues:
- [High] Missing bounds check on aggr_idx after searching config->aggr_map results in an out-of-bounds array read. --
--- Patch [1]: [PATCH] perf stat: Include PMU name and split uncore events per PMU in metric-only JSON output --- commit 418eb65331465e855519851f0438c3a1fe047563
Author: Chun-Tse Shao <[email protected]>
perf stat: Include PMU name and split uncore events per PMU in metric-only JSON output
This commit modifies perf stat to include the PMU name in the metric-only JSON output when running in no-aggr mode. It ensures that when an uncore event spans multiple PMU instances, each instance is printed on a separate line with the correct PMU identifier. The JSON linting script is also updated to recognize the new PMU key.
Quoted 4 lines
tools/perf/util/stat-display.c +0
-0
[ ... ]
tools/perf/util/stat-display.c static void print_no_aggr_metric(struct perf_stat_config *config, · line 1156
Quoted 12 lines
tools/perf/util/stat-display.c +5
-3@@ -1146,10 +1149,24 @@ static void print_no_aggr_metric(struct perf_stat_config *config, os->evsel = counter; os->id =aggr_cpu_id__cpu (cpu,/*data=*/ NULL ); -if (first) { -print_metric_begin (config, evlist, os, aggr_idx); - first = false; + +if (config->metric_only) { +struct perf_pmu *pmu = counter->pmu; + +if (!evsel__is_tool (counter)) {
[Severity: Medium]
Does bypassing the print_metric_begin() initialization for tool events cause invalid JSON output?
If a metric uses a tool event like duration_time as the metric leader, skipping print_metric_begin() means the opening bracket, timestamp, and CPU ID are not printed. This can leave the metric key-value pair as a bare string, followed by a mid-stream bracket when a subsequent hardware PMU event is processed.
Quoted 4 lines
tools/perf/util/stat-display.c +4
-0+if (config->json_output && line_open && pmu != last_pmu) { +print_metric_end (config, os); + line_open = false; + }
tools/perf/util/stat-display.c static void print_no_aggr_metric(struct perf_stat_config *config, · line 1170
Quoted 10 lines
tools/perf/util/stat-display.c +8
-0+if (!line_open) { +print_metric_begin (config, evlist, os, + aggr_idx, counter); + line_open = true; + last_pmu = pmu; + } + } } + val = ps->aggr[aggr_idx].counts.val;
[Severity: High]
This is a pre-existing issue, but does this code read past the end of the ps->aggr array if the CPU is not found in the aggregation map?
In print_no_aggr_metric(), there is a loop right before this block that searches for the CPU:
cpu_aggr_map__for_each_idx(aggr_idx, config->aggr_map) {
if (config->aggr_map->map[aggr_idx].cpu.cpu == cpu.cpu)
break;
}If a system lacks complete sysfs CPU topology, aggr_cpu_id__cpu() might return an empty ID, and the CPU may not be found in config->aggr_map.
Could the loop terminate with aggr_idx == config->aggr_map->nr, causing this line to unconditionally read out of bounds?
Quoted 2 lines
tools/perf/util/stat-display.c +0
-0ena = ps->aggr[aggr_idx].counts.ena; run = ps->aggr[aggr_idx].counts.run;
Signature · 1 line
Sashiko AI review · https://sashiko.dev/#/patchset/[email protected]?part=1