memcg

Reparenting mostly solves a problem with zombie memcg, where the memcg lingers because some resource is still charged to it.

It may make sense for other things in the OOMed memcg to wait for the OOM killer.

The author has identified the issue to be memcg reparenting causing a spike in CPU usage.

The zombie memcg problem is not unique to cgroup v2, but reparenting is fairly recent.

There is, memcg is a thing, and it has both minimum free and maximum bounds.

Instead of having allocating tasks go straight to global reclaim, they could try to reclaim the pages in the memcg they are part of first, as long as the group is not thrashing.

In Linux kernel, Different memory cgroups have distinct lruvecs, for memcg reclaim.

In the past, Ali and Google used to prioritize their tasks by setting the user's priority on the memcg priority setting by users to achieve this function, but BPF may be more flexible and applicable to more scenarios.

Subsystems affected by this patch series: slub, memcg, gup, kasan, pagealloc, hugetlb, vmscan, tools, mempolicy, memblock, hugetlbfs, thp, mmap, kconfig * More MM work.

With mm subsystems: slab-generic, slub, debug, pagecache, gup, swap, memcg, pagemap, memory-failure, vmalloc, kasan" * More mm/ work, plenty more to come.