tjh.dev / kernel
Linux kernel mirror (for testing) git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git
kernel os linux
1
fork

Configure Feed

Select the types of activity you want to include in your feed.

perf test: Remove cpu-list BPF cgroup counter test

The cpu-list part of this testcase has proven itself to be unreliable.
Sometimes, we get "<not counted>" for system.slice when pinned to CPUs
0 and 1. In such case, the test fails.

Since we cannot simply guarantee that any system.slice load will run
on any arbitrary list of CPUs, except the whole set of all CPUs, let's
rather remove the cpu-list subtest.

Fixes: a84260e314029e6dc9904fd ("perf test stat_bpf_counters_cgrp: Enhance perf stat cgroup BPF counter test")
Signed-off-by: Michael Petlan <mpetlan@redhat.com>
Cc: vmolnaro@redhat.com
Cc: bpf@vger.kernel.org
Link: https://lore.kernel.org/r/20241101102812.576425-1-mpetlan@redhat.com
Signed-off-by: Namhyung Kim <namhyung@kernel.org>

authored by

Michael Petlan and committed by
Namhyung Kim c741c7b5 13e17c9f

-13
-13
tools/perf/tests/shell/stat_bpf_counters_cgrp.sh
··· 58 58 fi 59 59 } 60 60 61 - check_cpu_list_counted() 62 - { 63 - 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) 64 - if echo ${check_cpu_list_counted_output} | grep -q -F "<not "; then 65 - echo "Some CPU events are not counted" 66 - if [ "${verbose}" = "1" ]; then 67 - echo ${check_cpu_list_counted_output} 68 - fi 69 - exit 1 70 - fi 71 - } 72 - 73 61 check_bpf_counter 74 62 find_cgroups 75 63 76 64 check_system_wide_counted 77 - check_cpu_list_counted 78 65 79 66 exit 0