Re: [patch 2/2] mm: memcg: count pte references from every member of the reclaimed hierarchy

[Konstantin Khlebnikov <khlebnikov@openvz.org>]

Johannes Weiner wrote:
> The rmap walker checking page table references has historically
> ignored references from VMAs that were not part of the memcg that was
> being reclaimed during memcg hard limit reclaim.
>
> When transitioning global reclaim to memcg hierarchy reclaim, I missed
> that bit and now references from outside a memcg are ignored even
> during global reclaim.
>
> Reverting back to traditional behaviour - count all references during
> global reclaim and only mind references of the memcg being reclaimed
> during limit reclaim would be one option.
>
> However, the more generic idea is to ignore references exactly then
> when they are outside the hierarchy that is currently under reclaim;
> because only then will their reclamation be of any use to help the
> pressure situation.  It makes no sense to ignore references from a
> sibling memcg and then evict a page that will be immediately refaulted
> by that sibling which contributes to the same usage of the common
> ancestor under reclaim.
>
> The solution: make the rmap walker ignore references from VMAs that
> are not part of the hierarchy that is being reclaimed.
>
> Flat limit reclaim will stay the same, hierarchical limit reclaim will
> mind the references only to pages that the hierarchy owns.  Global
> reclaim, since it reclaims from all memcgs, will be fixed to regard
> all references.
>
> Reported-by: Konstantin Khlebnikov<khlebnikov@openvz.org>
> Signed-off-by: Johannes Weiner<hannes@cmpxchg.org>

Thanks, it makes my patchset smaller.
One note below.

> ---
>   include/linux/memcontrol.h |    6 +++++-
>   mm/memcontrol.c            |   16 +++++++++++-----
>   mm/vmscan.c                |    6 ++++--
>   3 files changed, 20 insertions(+), 8 deletions(-)
>
> diff --git a/include/linux/memcontrol.h b/include/linux/memcontrol.h
> index 8537c5d..661b54a 100644
> --- a/include/linux/memcontrol.h
> +++ b/include/linux/memcontrol.h
> @@ -78,6 +78,7 @@ extern void mem_cgroup_uncharge_page(struct page *page);
>   extern void mem_cgroup_uncharge_cache_page(struct page *page);
>
>   extern void mem_cgroup_out_of_memory(struct mem_cgroup *memcg, gfp_t gfp_mask);
> +bool __mem_cgroup_same_or_subtree(const struct mem_cgroup *, struct mem_cgroup *);
>   int task_in_mem_cgroup(struct task_struct *task, const struct mem_cgroup *memcg);
>
>   extern struct mem_cgroup *try_get_mem_cgroup_from_page(struct page *page);
> @@ -88,10 +89,13 @@ static inline
>   int mm_match_cgroup(const struct mm_struct *mm, const struct mem_cgroup *cgroup)
>   {
>   	struct mem_cgroup *memcg;
> +	int match;
> +
>   	rcu_read_lock();
>   	memcg = mem_cgroup_from_task(rcu_dereference((mm)->owner));
> +	match = __mem_cgroup_same_or_subtree(cgroup, memcg);

I'm afraid mm->owner and memcg can be NULL here,
for example if sys_exit() is raced with sys_swapoff, so
match = memcg && __mem_cgroup_same_or_subtree(cgroup, memcg);
would be better.

>   	rcu_read_unlock();
> -	return cgroup == memcg;
> +	return match;
>   }
>
>   extern struct cgroup_subsys_state *mem_cgroup_css(struct mem_cgroup *memcg);
> diff --git a/mm/memcontrol.c b/mm/memcontrol.c
> index b4622fb..21004df 100644
> --- a/mm/memcontrol.c
> +++ b/mm/memcontrol.c
> @@ -1044,17 +1044,23 @@ struct lruvec *mem_cgroup_lru_move_lists(struct zone *zone,
>    * Checks whether given mem is same or in the root_mem_cgroup's
>    * hierarchy subtree
>    */
> -static bool mem_cgroup_same_or_subtree(const struct mem_cgroup *root_memcg,
> -		struct mem_cgroup *memcg)
> +bool __mem_cgroup_same_or_subtree(const struct mem_cgroup *root_memcg,
> +				  struct mem_cgroup *memcg)
>   {
> -	bool ret;
> -
>   	if (root_memcg == memcg)
>   		return true;
>   	if (!root_memcg->use_hierarchy)
>   		return false;
> +	return css_is_ancestor(&memcg->css,&root_memcg->css);
> +}
> +
> +static bool mem_cgroup_same_or_subtree(const struct mem_cgroup *root_memcg,
> +				       struct mem_cgroup *memcg)
> +{
> +	bool ret;
> +
>   	rcu_read_lock();
> -	ret = css_is_ancestor(&memcg->css,&root_memcg->css);
> +	ret = __mem_cgroup_same_or_subtree(root_memcg, memcg);
>   	rcu_read_unlock();
>   	return ret;
>   }
> diff --git a/mm/vmscan.c b/mm/vmscan.c
> index c631234..120646e 100644
> --- a/mm/vmscan.c
> +++ b/mm/vmscan.c
> @@ -708,7 +708,8 @@ static enum page_references page_check_references(struct page *page,
>   	int referenced_ptes, referenced_page;
>   	unsigned long vm_flags;
>
> -	referenced_ptes = page_referenced(page, 1, mz->mem_cgroup,&vm_flags);
> +	referenced_ptes = page_referenced(page, 1, sc->target_mem_cgroup,
> +					&vm_flags);
>   	referenced_page = TestClearPageReferenced(page);
>
>   	/* Lumpy reclaim - ignore references */
> @@ -1710,7 +1711,8 @@ static void shrink_active_list(unsigned long nr_pages,
>   			continue;
>   		}
>
> -		if (page_referenced(page, 0, mz->mem_cgroup,&vm_flags)) {
> +		if (page_referenced(page, 0, sc->target_mem_cgroup,
> +				&vm_flags)) {
>   			nr_rotated += hpage_nr_pages(page);
>   			/*
>   			 * Identify referenced, file-backed active pages and