Message ID | 20221208130805.66660-2-zhaogongyi@huawei.com (mailing list archive) |
---|---|
State | New |
Headers | show |
Series | Add setup/cleanup for run.sh | expand |
diff --git a/tools/testing/selftests/intel_pstate/run.sh b/tools/testing/selftests/intel_pstate/run.sh index e7008f614ad7..ee78c61cccb9 100755 --- a/tools/testing/selftests/intel_pstate/run.sh +++ b/tools/testing/selftests/intel_pstate/run.sh @@ -46,6 +46,12 @@ fi max_cpus=$(($(nproc)-1)) +# Save the scaling_governor +for i in `seq 0 $max_cpus` +do + cat /sys/devices/system/cpu/cpu$i/cpufreq/scaling_governor >> /tmp/governor +done + function run_test () { file_ext=$1 @@ -125,4 +131,14 @@ done # print the table pr -aTt -5 < /tmp/result.tab +# restore the scaling_governor +i=0 +while read line +do + echo $line > /sys/devices/system/cpu/cpu$i/cpufreq/scaling_governor + let i=$i+1 +done < /tmp/governor + +rm -f /tmp/governor + exit 0
Add saving and restore of scaling_governor for cpus, otherwise the cpus will run in the policy of powersave, and it would reduce the performance as unexpected. Signed-off-by: Zhao Gongyi <zhaogongyi@huawei.com> --- tools/testing/selftests/intel_pstate/run.sh | 16 ++++++++++++++++ 1 file changed, 16 insertions(+) -- 2.17.1