Skip to contentSkip to conversation

perf test: Skip 6.2 kernel for bpf-filter test

PATCH In review Namhyung Kim 2 participants 4 messages Aug 25, 2023

Subsystems (3)
lore
Share & tools
lore

Namhyung Kim

author maintainer
#0 lore raw

PATCH 1/3 perf test: Skip 6.2 kernel for bpf-filter test

The BPF sample filtering requires two kernel changes below:
* bpf_cast_to_kernel_ctx() kfunc (added in v6.2)
* setting perf_sample_data->sample_flags (finished in v6.3)

The perf tools can check bpf_cast_to_kernel_ctx() easily so it can refuse BPF filters on those old kernels (v6.1 and earlier). But checking sample_flags appears to be difficult so current code won't work on v6.2 kernel. That's unfortunate but I don't know what's the correct way to handle it.

For now, let's skip v6.2 kernels explicitly (if failed) in the test.

Fixes 9575ecdd198a ("perf test: Add perf record sample filtering test")
---
 tools/perf/tests/shell/record_bpf_filter.sh | 6 ++++++
 1 file changed, 6 insertions(+)
tools/perf/tests/shell/record_bpf_filter.sh +6 -0
index e76ea861b92c..31c593966e8c 100755
@@ -49,6 +49,12 @@ test_bpf_filter_basic() {
   fi
   if perf script -i "${perfdata}" -F ip | grep 'ffffffff[0-9a-f]*'
   then
+    if uname -r | grep -q ^6.2
+    then
+      echo "Basic bpf-filter test [Skipped unsupported kernel]"
+      err=2
+      return
+    fi
     echo "Basic bpf-filter test [Failed invalid output]"
     err=1
     return
Signature · 1 line
2.42.0.rc1.204.g551eb34607-goog
author maintainer
#1 lore raw

PATCH 2/3 perf test: Fix perf stat bpf counters test on Intel

As of now, bpf counters (bperf) don't support event groups. But the default perf stat includes topdown metrics if supported (on recent Intel machines) which require groups. That makes perf stat exiting.

  $ sudo perf stat --bpf-counter true
  bpf managed perf events do not yet support groups.

Actually the test explicitly uses cycles event only, but it missed to pass the option when it checks the availability of the command.

Fixes 2c0cb9f56020d ("perf test: Add a shell test for 'perf stat --bpf-counters' new option")
Cc Song Liu <[email protected]>
---
 tools/perf/tests/shell/stat_bpf_counters.sh | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
tools/perf/tests/shell/stat_bpf_counters.sh +2 -2
index 513cd1e58e0e..a87bb2814b4c 100755
@@ -22,10 +22,10 @@ compare_number()
 }
 
 # skip if --bpf-counters is not supported
-if ! perf stat --bpf-counters true > /dev/null 2>&1; then
+if ! perf stat -e cycles --bpf-counters true > /dev/null 2>&1; then
 	if [ "$1" = "-v" ]; then
 		echo "Skipping: --bpf-counters not supported"
-		perf --no-pager stat --bpf-counters true || true
+		perf --no-pager stat -e cycles --bpf-counters true || true
 	fi
 	exit 2
 fi
Signature · 1 line
2.42.0.rc1.204.g551eb34607-goog
reviewer Reviewed-by on v1 2/3
#2 lore raw

Re: [PATCH 2/3] perf test: Fix perf stat bpf counters test on Intel

Quoted 8 lines · Namhyung Kim

As of now, bpf counters (bperf) don't support event groups. But the default perf stat includes topdown metrics if supported (on recent Intel machines) which require groups. That makes perf stat exiting.

  $ sudo perf stat --bpf-counter true
  bpf managed perf events do not yet support groups.
Quoted 7 lines

Actually the test explicitly uses cycles event only, but it missed to pass the option when it checks the availability of the command.

Fixes 2c0cb9f56020d ("perf test: Add a shell test for 'perf stat --bpf-counters' new option")
Cc Song Liu <[email protected]>

Thanks!
Song

Quoted 24 lines
---
 tools/perf/tests/shell/stat_bpf_counters.sh | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
tools/perf/tests/shell/stat_bpf_counters.sh +2 -3
index 513cd1e58e0e..a87bb2814b4c 100755
@@ -22,10 +22,10 @@ compare_number()
 }
 
 # skip if --bpf-counters is not supported
-if ! perf stat --bpf-counters true > /dev/null 2>&1; then
+if ! perf stat -e cycles --bpf-counters true > /dev/null 2>&1; then
        if [ "$1" = "-v" ]; then
                echo "Skipping: --bpf-counters not supported"
-               perf --no-pager stat --bpf-counters true || true
+               perf --no-pager stat -e cycles --bpf-counters true || true
        fi
        exit 2
 fi
--

2.42.0.rc1.204.g551eb34607-goog

author maintainer
#3 lore raw

PATCH 3/3 perf test: Enhance perf stat cgroup BPF counter test

In reply to Namhyung Kim

It has system-wide test and cpu-list test but the cpu-list test fails sometimes. It runs sleep command on CPU1 and measure both user.slice and system.slice cgroups by default (on systemd-based systems).

But if the system was idle enough, sometime the system.slice gets no count and it makes the test failing. Maybe that's because it only looks at the CPU1, let's add CPU0 to increase the chance it finds some tasks.

Fixes 7901086014bba ("perf test: Add a new test for perf stat cgroup BPF counter")
---
 tools/perf/tests/shell/stat_bpf_counters_cgrp.sh | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
tools/perf/tests/shell/stat_bpf_counters_cgrp.sh +1 -1
index a74440a00b6b..e75d0780dc78 100755
@@ -60,7 +60,7 @@ check_system_wide_counted()
 
 check_cpu_list_counted()
 {
-	check_cpu_list_counted_output=$(perf stat -C 1 --bpf-counters --for-each-cgroup ${test_cgroups} -e cpu-clock -x, taskset -c 1 sleep 1  2>&1)
+	check_cpu_list_counted_output=$(perf stat -C 0,1 --bpf-counters --for-each-cgroup ${test_cgroups} -e cpu-clock -x, taskset -c 1 sleep 1  2>&1)
 	if echo ${check_cpu_list_counted_output} | grep -q -F "<not "; then
 		echo "Some CPU events are not counted"
 		if [ "${verbose}" = "1" ]; then
Signature · 1 line
2.42.0.rc1.204.g551eb34607-goog