Re: [PATCH] perf test: Add a perf event fallback test

["Chen, Zide" <zide.chen@intel.com>]

On 11/11/2025 5:06 PM, Namhyung Kim wrote:
> On Tue, Nov 11, 2025 at 02:42:46PM -0800, Zide Chen wrote:
>> This adds test cases to verify the precise ip fallback logic:
>>
>> - If the system supports precise ip, for an event given with the maximum
>>   precision level, it should be able to decrease precise_ip to find a
>>   supported level.
>> - The same fallback behavior should also work in more complex scenarios,
>>   such as event groups or when PEBS is involved
>>
>> Additional fallback tests, such as those covering missing feature cases,
>> can be added in the future.
>>
>> Suggested-by: Ian Rogers <irogers@google.com>
>> Reviewed-by: Dapeng Mi <dapeng1.mi@linux.intel.com>
>> Signed-off-by: Zide Chen <zide.chen@intel.com>
>> ---
>>  .../tests/shell/test_event_open_fallback.sh   | 86 +++++++++++++++++++
>>  1 file changed, 86 insertions(+)
>>  create mode 100755 tools/perf/tests/shell/test_event_open_fallback.sh
>>
>> diff --git a/tools/perf/tests/shell/test_event_open_fallback.sh b/tools/perf/tests/shell/test_event_open_fallback.sh
>> new file mode 100755
>> index 000000000000..72c1ac32c785
>> --- /dev/null
>> +++ b/tools/perf/tests/shell/test_event_open_fallback.sh
>> @@ -0,0 +1,86 @@
>> +#!/bin/bash
>> +# Perf event open fallback test
>> +# SPDX-License-Identifier: GPL-2.0
>> +
>> +skip_cnt=0
>> +ok_cnt=0
>> +err_cnt=0
>> +
>> +cleanup()
>> +{
>> +	rm -f perf.data
>> +	rm -f perf.data.old
>> +	trap - EXIT TERM INT
>> +}
>> +
>> +trap_cleanup()
>> +{
>> +	cleanup
>> +	exit 1
>> +}
>> +
>> +trap trap_cleanup EXIT TERM INT
>> +
>> +perf_record()
>> +{
>> +	perf record "$@" -- true 1>/dev/null 2>&1
>> +}
>> +
>> +test_decrease_precise_ip()
>> +{
>> +	echo "Decrease precise ip test"
>> +
>> +	perf list pmu | grep -q 'cycles' || return 2
>> +
>> +	if ! perf_record -e cycles; then
>> +		return 2
>> +	fi
>> +
>> +	# It should reduce precision level down to 0 if needed.
>> +	if ! perf_record -e cycles:ppp; then
> 
> I think you need 'P' instead of 'ppp' for automatic precision.

Yes, it's better with "-e cycles:P" to explicitly set evsel.precise_max
to 1.

> Thanks,
> Namhyung
> 
> 
>> +		return 1
>> +	fi
>> +	return 0
>> +}
>> +
>> +test_decrease_precise_ip_complicated()
>> +{
>> +	echo "Decrease precise ip test (complicated case)"
>> +
>> +	perf list pmu | grep -q 'mem-loads-aux' || return 2
>> +
>> +	if ! perf_record -e '{cpu/mem-loads-aux/S,cpu/mem-loads/PS}'; then
>> +		return 1
>> +	fi
>> +	return 0
>> +}
>> +
>> +count_result()
>> +{
>> +	if [ "$1" -eq 2 ] ; then
>> +		skip_cnt=$((skip_cnt + 1))
>> +		return
>> +	fi
>> +	if [ "$1" -eq 0 ] ; then
>> +		ok_cnt=$((ok_cnt + 1))
>> +		return
>> +	fi
>> +	err_cnt=$((err_cnt + 1))
>> +}
>> +
>> +ret=0
>> +test_decrease_precise_ip		|| ret=$? ; count_result $ret ; ret=0
>> +test_decrease_precise_ip_complicated	|| ret=$? ; count_result $ret ; ret=0
>> +
>> +cleanup
>> +
>> +if [ ${err_cnt} -gt 0 ] ; then
>> +	exit 1
>> +fi
>> +
>> +if [ ${ok_cnt} -gt 0 ] ; then
>> +	exit 0
>> +fi
>> +
>> +# Skip
>> +exit 2
>> -- 
>> 2.51.1
>>