Re: [PATCH] mm/vmscan: get number of pages on the LRU list in memcgroup base on lru_zone_size

[Michal Hocko <mhocko@kernel.org>]

On Tue 03-09-19 16:54:16, Honglei Wang wrote:
> lruvec_lru_size() is involving lruvec_page_state_local() to get the
> lru_size in the current code. It's base on lruvec_stat_local.count[]
> of mem_cgroup_per_node. This counter is updated in batch. It won't
> do charge if the number of coming pages doesn't meet the needs of
> MEMCG_CHARGE_BATCH who's defined as 32 now.
> 
> This causes small section of memory can't be handled as expected in
> some scenario. For example, if we have only 32 pages madvise free
> memory in memcgroup, these pages won't be freed as expected when it
> meets memory pressure in this group.

Could you be more specific please?

> Getting lru_size base on lru_zone_size of mem_cgroup_per_node which
> is not updated in batch can make this a bit more accurate.

This is effectivelly reverting 1a61ab8038e72. There were no numbers
backing that commit, neither this one has. The only hot path I can see
is workingset_refault. All others seems to be in the reclaim path.
 
> Signed-off-by: Honglei Wang <honglei.wang@oracle.com>

That being said, I am not against this patch but the changelog should be
more specific about the particular problem and how serious it is.

> ---
>  mm/vmscan.c | 9 +++++----
>  1 file changed, 5 insertions(+), 4 deletions(-)
> 
> diff --git a/mm/vmscan.c b/mm/vmscan.c
> index c77d1e3761a7..c28672460868 100644
> --- a/mm/vmscan.c
> +++ b/mm/vmscan.c
> @@ -354,12 +354,13 @@ unsigned long zone_reclaimable_pages(struct zone *zone)
>   */
>  unsigned long lruvec_lru_size(struct lruvec *lruvec, enum lru_list lru, int zone_idx)
>  {
> -	unsigned long lru_size;
> +	unsigned long lru_size = 0;
>  	int zid;
>  
> -	if (!mem_cgroup_disabled())
> -		lru_size = lruvec_page_state_local(lruvec, NR_LRU_BASE + lru);
> -	else
> +	if (!mem_cgroup_disabled()) {
> +		for (zid = 0; zid < MAX_NR_ZONES; zid++)
> +			lru_size += mem_cgroup_get_zone_lru_size(lruvec, lru, zid);
> +	} else
>  		lru_size = node_page_state(lruvec_pgdat(lruvec), NR_LRU_BASE + lru);
>  
>  	for (zid = zone_idx + 1; zid < MAX_NR_ZONES; zid++) {
> -- 
> 2.17.0

-- 
Michal Hocko
SUSE Labs