Re: [BUGFIX][PATCH] memcg: fix wrong decision of noswap withsoftlimit.

[Daisuke Nishimura]

On Thu, 9 Jun 2011 09:54:45 +0900
KAMEZAWA Hiroyuki <kamezawa.hiroyu@xxxxxxxxxxxxxx> wrote:

> I wonder this should go stable...
hmm, IMHO, it's not necessary just because there have been no bug reports
about this bug.

> ==
> From e2565de1c764057b75b4d9a1674d163b6c873cdd Mon Sep 17 00:00:00 2001
> From: KAMEZAWA Hiroyuki <kamezawa.hiroyu@xxxxxxxxxxxxxx>
> Date: Thu, 9 Jun 2011 09:54:32 +0900
> Subject: [PATCH 2/2] Fix softlimit wrong check of noswap
> 
> Now, hierarchical reclaim doesn't make swap if memory's limit is
> equal to mem+swap limit. Because if reclaim does swap-out,
> it still hits mem+swap limit and there will be no progress.
> WHEN HITTING HARD LIMIT.
> 
> When it comes to softlimit, it works for kswapd. noswap is nonsense.
> 
> Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@xxxxxxxxxxxxxx>

If we implement "softlimit for memsw" in future, we might change the check again,
but it's another story.

Acked-by: Daisuke Nishimura <nishimura@xxxxxxxxxxxxxxxxx>

Thanks,
Daisuke Nishimura.

> ---
>  mm/memcontrol.c |    2 +-
>  1 files changed, 1 insertions(+), 1 deletions(-)
> 
> diff --git a/mm/memcontrol.c b/mm/memcontrol.c
> index 3baddcb..06825be 100644
> --- a/mm/memcontrol.c
> +++ b/mm/memcontrol.c
> @@ -1663,7 +1663,7 @@ static int mem_cgroup_hierarchical_reclaim(struct mem_cgroup *root_mem,
>  	excess = res_counter_soft_limit_excess(&root_mem->res) >> PAGE_SHIFT;
>  
>  	/* If memsw_is_minimum==1, swap-out is of-no-use. */
> -	if (root_mem->memsw_is_minimum)
> +	if (!check_soft && root_mem->memsw_is_minimum)
>  		noswap = true;
>  
>  	while (1) {
> -- 
> 1.7.4.1
> 
> 
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/