On Wed 20-10-21 11:07:02, Vasily Averin wrote:
[...]
I haven't read through the changelog and only focused on the patch this
time.
[...]
> @@ -1810,11 +1810,21 @@ static enum oom_status mem_cgroup_oom(struct mem_cgroup *memcg, gfp_t mask, int
> mem_cgroup_oom_notify(memcg);
>
> mem_cgroup_unmark_under_oom(memcg);
> - if (mem_cgroup_out_of_memory(memcg, mask, order))
> + if (mem_cgroup_out_of_memory(memcg, mask, order)) {
> ret = OOM_SUCCESS;
> - else
> + } else {
> ret = OOM_FAILED;
> -
> + /*
> + * In some rare cases mem_cgroup_out_of_memory() can return false.
> + * If it was called from #PF it forces handle_mm_fault()
> + * return VM_FAULT_OOM and executes pagefault_out_of_memory().
> + * memcg_in_oom is set here to notify pagefault_out_of_memory()
> + * that it was a memcg-related failure and not allow to run
> + * global OOM.
> + */
> + if (current->in_user_fault)
> + current->memcg_in_oom = (struct mem_cgroup *)ret;
> + }
> if (locked)
> mem_cgroup_oom_unlock(memcg);
>
> @@ -1848,6 +1858,15 @@ bool mem_cgroup_oom_synchronize(bool handle)
> if (!memcg)
> return false;
>
> + /* OOM is memcg, however out_of_memory() found no victim */
> + if (memcg == (struct mem_cgroup *)OOM_FAILED) {
> + /*
> + * Should be called from pagefault_out_of_memory() only,
> + * where it is used to prevent false global OOM.
> + */
> + current->memcg_in_oom = NULL;
> + return true;
> + }
> if (!handle)
> goto cleanup;
I have to say I am not a great fan of this but this belongs to a
separate patch on top of all the previous ones.
[...]
> diff --git a/mm/oom_kill.c b/mm/oom_kill.c
> index 831340e7ad8b..1deef8c7a71b 100644
> --- a/mm/oom_kill.c
> +++ b/mm/oom_kill.c
> @@ -1137,6 +1137,9 @@ void pagefault_out_of_memory(void)
> if (mem_cgroup_oom_synchronize(true))
> return;
>
> + if (fatal_signal_pending(current))
> + return;
> +
This belongs to its own patch as well.
All that being said I would go with pagefault_out_of_memory as the first
patch because it is necessary to handle charge failures. Then go with a
patch to remove charge forcing when OOM killer succeeds but the retry
still fails and finally go with one that tries to handle oom failures.
--
Michal Hocko
SUSE Labs
