On Tue, Jul 18, 2023 at 06:08:50PM +0800, Hao Jia wrote: > https://github.com/jiaozhouxiaojia/cgv2-stat-percpu_test/tree/main Isn't that just adding the same numbers twice and verifying that? Maybe I'm misunderstanding you. Here's a simpler case: # cd /sys/fs/cgroup # mkdir -p asdf/test0 # grep usage_usec asdf/test0/cpu.stat usage_usec 0 # echo $$ > asdf/test0/cgroup.procs # stress -c 1 & sleep 1; kill %% [1] 122329 stress: info: [122329] dispatching hogs: 1 cpu, 0 io, 0 vm, 0 hdd # grep usage_usec asdf/test0/cpu.stat usage_usec 1000956 [1]+ Terminated stress -c 1 # grep usage_usec asdf/cpu.stat usage_usec 1002548 # echo $$ > /sys/fs/cgroup/cgroup.procs # rmdir asdf/test0 # grep usage_usec asdf/cpu.stat usage_usec 1006338 So, we run `stress -c 1` for 1 second in the asdf/test0 cgroup and asdf/cpu.stat correctly reports the cumulative usage. After removing asdf/test0 cgroup, asdf's usage_usec is still there. What's missing here? What are you adding? Thanks. -- tejun
