Re: [RFC][PATCH] memory cgroup enhancements updated [1/10] try_to_free_mem_cgroup_pages bugfix

[Balbir Singh <balbir-23VcF4HTsmIX0ybBhKVfKdBPR1lH4CV8@public.gmane.org>]

KAMEZAWA Hiroyuki wrote:
> Because NODE_DATA(node)->node_zonelists[] is guaranteed to contain
> all necessary zones, it is not necessary to use for_each_online_node.
> 
> And this for_each_online_node() makes reclaim routine start always
> from node 0. This is bad.
> 

But with this change, we'll start reclaim on the node that the task
hit the limit on - right?

> Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu-+CUm20s59erQFUHtdCDX3A@public.gmane.org>
> 
> 
> 
>  mm/vmscan.c |    8 +++-----
>  1 file changed, 3 insertions(+), 5 deletions(-)
> 
> Index: devel-2.6.23-mm1/mm/vmscan.c
> ===================================================================
> --- devel-2.6.23-mm1.orig/mm/vmscan.c
> +++ devel-2.6.23-mm1/mm/vmscan.c
> @@ -1375,15 +1375,13 @@ unsigned long try_to_free_mem_cgroup_pag
>  		.mem_cgroup = mem_cont,
>  		.isolate_pages = mem_cgroup_isolate_pages,
>  	};
> -	int node;
> +	int node = numa_node_id();
>  	struct zone **zones;
>  	int target_zone = gfp_zone(GFP_HIGHUSER_MOVABLE);
> 
> -	for_each_online_node(node) {
> -		zones = NODE_DATA(node)->node_zonelists[target_zone].zones;
> -		if (do_try_to_free_pages(zones, sc.gfp_mask, &sc))
> +	zones = NODE_DATA(node)->node_zonelists[target_zone].zones;
> +	if (do_try_to_free_pages(zones, sc.gfp_mask, &sc))
>  			return 1;
> -	}
>  	return 0;
>  }
>  #endif
> 


-- 
	Warm Regards,
	Balbir Singh
	Linux Technology Center
	IBM, ISTL