linux-kernel - Re: [PATCH 1/2] mm/memcontrol.c: remove meaningless while loop in mem_cgroup

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:	Tue, 22 Apr 2014 11:47:59 +0200
From:	Michal Hocko <mhocko@...e.cz>
To:	Jianyu Zhan <nasa4836@...il.com>
Cc:	hannes@...xchg.org, bsingharora@...il.com,
	kamezawa.hiroyu@...fujitsu.com, cgroups@...r.kernel.org,
	linux-mm@...ck.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH 1/2] mm/memcontrol.c: remove meaningless while loop in
 mem_cgroup_iter()

On Sat 19-04-14 06:58:55, Jianyu Zhan wrote:
> Currently, the iteration job in mem_cgroup_iter() all delegates
> to __mem_cgroup_iter_next(), which will skip dead node.
> 
> Thus, the outer while loop in mem_cgroup_iter() is meaningless.
> It could be proven by this:
> 
> 1. memcg != NULL
>     we are done, no loop needed.
> 2. memcg == NULL
>    2.1 prev != NULL, a round-trip is done, break out, no loop.
>    2.2 prev == NULL, this is impossible, since prev==NULL means
>        the initial interation, it will returns memcg==root.

What about
  3. last_visited == last_node in the tree

__mem_cgroup_iter_next returns NULL and the iterator would return
without visiting anything.

The patch is not correct, I am afraid.

> So, this patches remove this meaningless while loop.
> 
> Signed-off-by: Jianyu Zhan <nasa4836@...il.com>
> ---
>  mm/memcontrol.c | 49 ++++++++++++++++++++++---------------------------
>  1 file changed, 22 insertions(+), 27 deletions(-)
> 
> diff --git a/mm/memcontrol.c b/mm/memcontrol.c
> index 29501f0..e0ce15c 100644
> --- a/mm/memcontrol.c
> +++ b/mm/memcontrol.c
> @@ -1212,6 +1212,8 @@ struct mem_cgroup *mem_cgroup_iter(struct mem_cgroup *root,
>  {
>  	struct mem_cgroup *memcg = NULL;
>  	struct mem_cgroup *last_visited = NULL;
> +	struct mem_cgroup_reclaim_iter *uninitialized_var(iter);
> +	int uninitialized_var(seq);
>  
>  	if (mem_cgroup_disabled())
>  		return NULL;
> @@ -1229,40 +1231,33 @@ struct mem_cgroup *mem_cgroup_iter(struct mem_cgroup *root,
>  	}
>  
>  	rcu_read_lock();
> -	while (!memcg) {
> -		struct mem_cgroup_reclaim_iter *uninitialized_var(iter);
> -		int uninitialized_var(seq);
> -
> -		if (reclaim) {
> -			int nid = zone_to_nid(reclaim->zone);
> -			int zid = zone_idx(reclaim->zone);
> -			struct mem_cgroup_per_zone *mz;
> -
> -			mz = mem_cgroup_zoneinfo(root, nid, zid);
> -			iter = &mz->reclaim_iter[reclaim->priority];
> -			if (prev && reclaim->generation != iter->generation) {
> -				iter->last_visited = NULL;
> -				goto out_unlock;
> -			}
> +	if (reclaim) {
> +		int nid = zone_to_nid(reclaim->zone);
> +		int zid = zone_idx(reclaim->zone);
> +		struct mem_cgroup_per_zone *mz;
>  
> -			last_visited = mem_cgroup_iter_load(iter, root, &seq);
> +		mz = mem_cgroup_zoneinfo(root, nid, zid);
> +		iter = &mz->reclaim_iter[reclaim->priority];
> +		if (prev && reclaim->generation != iter->generation) {
> +			iter->last_visited = NULL;
> +			goto out_unlock;
>  		}
>  
> -		memcg = __mem_cgroup_iter_next(root, last_visited);
> +		last_visited = mem_cgroup_iter_load(iter, root, &seq);
> +	}
>  
> -		if (reclaim) {
> -			mem_cgroup_iter_update(iter, last_visited, memcg, root,
> -					seq);
> +	memcg = __mem_cgroup_iter_next(root, last_visited);
>  
> -			if (!memcg)
> -				iter->generation++;
> -			else if (!prev && memcg)
> -				reclaim->generation = iter->generation;
> -		}
> +	if (reclaim) {
> +		mem_cgroup_iter_update(iter, last_visited, memcg, root,
> +				seq);
>  
> -		if (prev && !memcg)
> -			goto out_unlock;
> +		if (!memcg)
> +			iter->generation++;
> +		else if (!prev && memcg)
> +			reclaim->generation = iter->generation;
>  	}
> +
>  out_unlock:
>  	rcu_read_unlock();
>  out_css_put:
> -- 
> 1.9.0.GIT
> 

-- 
Michal Hocko
SUSE Labs
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/