perf test: Update all metrics test like metricgroups test
Namhyung Kim
maintainerPATCH 1/1 perf test: Update all metrics test like metricgroups test
Like in the metricgroup tests, it should check the permission first and then skip relevant failures accordingly.
Also it needs to try again with the system wide flag properly. On the second round, check if the result has the metric name because other failure cases are checked in the first round already.
--- tools/perf/tests/shell/stat_all_metrics.sh | 87 ++++++++++++++++------ 1 file changed, 66 insertions(+), 21 deletions(-)
tools/perf/tests/shell/stat_all_metrics.sh +66 -21
@@ -2,42 +2,87 @@ # perf all metrics test # SPDX-License-Identifier: GPL-2.0 +ParanoidAndNotRoot() +{ + [ "$(id -u)" != 0 ] && [ "$(cat /proc/sys/kernel/perf_event_paranoid)" -gt $1 ] +} + +system_wide_flag="-a" +if ParanoidAndNotRoot 0 +then + system_wide_flag="" +fi + err=0 for m in $(perf list --raw-dump metrics); do echo "Testing $m" - result=$(perf stat -M "$m" true 2>&1) - if [[ "$result" =~ ${m:0:50} ]] || [[ "$result" =~ "<not supported>" ]] + result=$(perf stat -M "$m" $system_wide_flag -- sleep 0.01 2>&1) + result_err=$? + if [[ $result_err -gt 0 ]] then - continue + if [[ "$result" =~ \ + "Access to performance monitoring and observability operations is limited" ]] + then + echo "Permission failure" + echo $result + if [[ $err -eq 0 ]] + then + err=2 # Skip + fi + continue + elif [[ "$result" =~ "in per-thread mode, enable system wide" ]] + then + echo "Permissions - need system wide mode" + echo $result + if [[ $err -eq 0 ]] + then + err=2 # Skip + fi + continue + elif [[ "$result" =~ "<not supported>" ]] + then + echo "Not supported events" + echo $result + if [[ $err -eq 0 ]] + then + err=2 # Skip + fi + continue + elif [[ "$result" =~ "FP_ARITH" || "$result" =~ "AMX" ]] + then + echo "FP issues" + echo $result + if [[ $err -eq 0 ]] + then + err=2 # Skip + fi + continue + elif [[ "$result" =~ "PMM" ]] + then + echo "Optane memory issues" + echo $result + if [[ $err -eq 0 ]] + then + err=2 # Skip + fi + continue + fi fi - # Failed so try system wide. - result=$(perf stat -M "$m" -a sleep 0.01 2>&1) + if [[ "$result" =~ ${m:0:50} ]] then continue fi - # Failed again, possibly the workload was too small so retry with something - # longer. - result=$(perf stat -M "$m" perf bench internals synthesize 2>&1) + + # Failed, possibly the workload was too small so retry with something longer. + result=$(perf stat -M "$m" $system_wide_flag -- perf bench internals synthesize 2>&1) if [[ "$result" =~ ${m:0:50} ]] then continue fi echo "Metric '$m' not printed in:" echo "$result" - if [[ "$err" != "1" ]] - then - err=2 - if [[ "$result" =~ "FP_ARITH" || "$result" =~ "AMX" ]] - then - echo "Skip, not fail, for FP issues" - elif [[ "$result" =~ "PMM" ]] - then - echo "Skip, not fail, for Optane memory issues" - else - err=1 - fi - fi + err=1 done exit "$err"
Signature · 1 line
2.47.0.rc1.288.g06298d1525-goog
Quoted 9 lines · Namhyung Kim
Like in the metricgroup tests, it should check the permission first and then skip relevant failures accordingly.
Also it needs to try again with the system wide flag properly. On the second round, check if the result has the metric name because other failure cases are checked in the first round already.
Signed-off-by Namhyung Kim <[email protected]>
Thanks,
Ian
Quoted 120 lines
--- tools/perf/tests/shell/stat_all_metrics.sh | 87 ++++++++++++++++------ 1 file changed, 66 insertions(+), 21 deletions(-)tools/perf/tests/shell/stat_all_metrics.sh +66
-22@@ -2,42 +2,87 @@ # perf all metrics test # SPDX-License-Identifier: GPL-2.0 +ParanoidAndNotRoot() +{ + [ "$(id -u)" != 0 ] && [ "$(cat /proc/sys/kernel/perf_event_paranoid)" -gt $1 ] +} + +system_wide_flag="-a" +if ParanoidAndNotRoot 0 +then + system_wide_flag="" +fi + err=0 for m in $(perf list --raw-dump metrics); do echo "Testing $m" - result=$(perf stat -M "$m" true 2>&1) - if [[ "$result" =~ ${m:0:50} ]] || [[ "$result" =~ "<not supported>" ]] + result=$(perf stat -M "$m" $system_wide_flag -- sleep 0.01 2>&1) + result_err=$? + if [[ $result_err -gt 0 ]] then - continue + if [[ "$result" =~ \ + "Access to performance monitoring and observability operations is limited" ]] + then + echo "Permission failure" + echo $result + if [[ $err -eq 0 ]] + then + err=2 # Skip + fi + continue + elif [[ "$result" =~ "in per-thread mode, enable system wide" ]] + then + echo "Permissions - need system wide mode" + echo $result + if [[ $err -eq 0 ]] + then + err=2 # Skip + fi + continue + elif [[ "$result" =~ "<not supported>" ]] + then + echo "Not supported events" + echo $result + if [[ $err -eq 0 ]] + then + err=2 # Skip + fi + continue + elif [[ "$result" =~ "FP_ARITH" || "$result" =~ "AMX" ]] + then + echo "FP issues" + echo $result + if [[ $err -eq 0 ]] + then + err=2 # Skip + fi + continue + elif [[ "$result" =~ "PMM" ]] + then + echo "Optane memory issues" + echo $result + if [[ $err -eq 0 ]] + then + err=2 # Skip + fi + continue + fi fi - # Failed so try system wide. - result=$(perf stat -M "$m" -a sleep 0.01 2>&1) + if [[ "$result" =~ ${m:0:50} ]] then continue fi - # Failed again, possibly the workload was too small so retry with something - # longer. - result=$(perf stat -M "$m" perf bench internals synthesize 2>&1) + + # Failed, possibly the workload was too small so retry with something longer. + result=$(perf stat -M "$m" $system_wide_flag -- perf bench internals synthesize 2>&1) if [[ "$result" =~ ${m:0:50} ]] then continue fi echo "Metric '$m' not printed in:" echo "$result" - if [[ "$err" != "1" ]] - then - err=2 - if [[ "$result" =~ "FP_ARITH" || "$result" =~ "AMX" ]] - then - echo "Skip, not fail, for FP issues" - elif [[ "$result" =~ "PMM" ]] - then - echo "Skip, not fail, for Optane memory issues" - else - err=1 - fi - fi + err=1 done exit "$err" --2.47.0.rc1.288.g06298d1525-goog
Quoted 8 lines · Namhyung Kim
Like in the metricgroup tests, it should check the permission first and then skip relevant failures accordingly.
Also it needs to try again with the system wide flag properly. On the second round, check if the result has the metric name because other failure cases are checked in the first round already.
[...]
Applied to perf-tools-next, thanks!
Best regards,
Namhyung