| Message ID | 20210720143414.14599-1-krzysztof.kozlowski@canonical.com |
|---|---|
| State | Accepted |
| Headers | show |
| Series | controllers/memcg_stress: fix order of process cleanup | expand |
Hi! Good catch, applied, thanks.
diff --git a/testcases/kernel/controllers/memcg/stress/memcg_stress_test.sh b/testcases/kernel/controllers/memcg/stress/memcg_stress_test.sh index cd75c704a9a4..c2501e164018 100755 --- a/testcases/kernel/controllers/memcg/stress/memcg_stress_test.sh +++ b/testcases/kernel/controllers/memcg/stress/memcg_stress_test.sh @@ -78,7 +78,7 @@ run_stress() mkdir /dev/memcg/$i 2> /dev/null memcg_process_stress $mem_size $interval & echo $! > /dev/memcg/$i/tasks - pids="$! $pids" + pids="$pids $!" done for pid in $pids; do
The memcg_stress_test was forking test processes and adding them to new cgroup. However cleanup - killing forked process and cgroup removal - was iterating in mixed up order: killing the newest process but trying to remove the oldest cgroup. This could work if all processes died already via earlier "kill -USR1" but at least in case of Ubuntu with v4.15 kernel this failed and left cgroups not cleaned up: + kill -KILL 3147 /home/azure/ltp-install/testcases/bin/memcg_stress_test.sh: 96: kill: No such process + wait 3147 + rmdir /dev/memcg/0 /home/azure/ltp-install/testcases/bin/memcg_stress_test.sh: 98: /home/azure/ltp-install/testcases/bin/memcg_stress_test.sh: Cannot fork $ ll /dev/memcg 0 107 116 125 134 143 18 27 36 45 54 63 72 81 90 1 108 117 126 135 144 19 28 37 46 55 64 73 82 91 Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com> --- testcases/kernel/controllers/memcg/stress/memcg_stress_test.sh | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)