Description of problem: When a system OOM occur, and select a victim to kill, it's better to notify the victim's memcg's OOM event, thus, the container's cgroup OOM event watcher would be notified and mark the container's exit reason as "OOM". Version-Release number of selected component (if applicable): How reproducible: Steps to Reproduce: 1. create a process in a memory cgroup and let it allocate a larger memory 2. in the root memory cgroup, try to allocate large memory to tigger the system OOM 3. the process in the memory cgroup was selected as an victim Actual results: The victim process located cgroup didn't receive the OOM event Expected results: The victim process located cgroup should receive the OOM event Additional info:
The PR Link: https://gitee.com/anolis/cloud-kernel/pulls/2836