> > I am not objecting to the patch I would just like to understand the > runaway case. ep_insert seems to limit the maximum number of watches to > max_user_watches which should be ~4% of lowmem if I am following the > code properly. pwq_cache should be bound by the number of watches as > well, or am I misunderstanding the code? > You are absolutely right that there is a per-user limit (~4% of total memory if no highmem) on these caches. I think it is too generous particularly in the scenario where jobs of multiple users are running on the system and the administrator is reducing cost by overcomitting the memory. This is unaccounted kernel memory and will not be considered by the oom-killer. I think by accounting it to kmemcg, for systems with kmem accounting enabled, we can provide better isolation between jobs of different users.
