Re: [PATCH v2] mm/memcg: use exist interface to get css from memcg

2012-07-19 Thread Hugh Dickins
On Thu, 19 Jul 2012, Kamezawa Hiroyuki wrote:
> (2012/07/19 18:29), Michal Hocko wrote:
> > On Wed 18-07-12 19:08:54, Wanpeng Li wrote:
> > > use exist interface mem_cgroup_css instead of &mem->css.
> > 
> > This interface has been added to enable mem->css outside of
> > mm/memcontrol.c (where we define struct mem_cgroup). There is one user
> > left (hwpoison_filter_task) after recent clean ups.
> > 
> > I think we shouldn't spread the usage inside the mm/memcontrol.c. The
> > compiler inlines the function for all callers added by this patch but I
> > wouldn't rely on it. It is also unfortunate that we cannot convert all
> > dereferences (e.g. const mem_cgroup).
> > Moreover it doesn't add any additional type safety. So I would vote for
> > not taking the patch but if others like it I will not block it.
> > 
> 
> Agreed.

Very strongly agreed: I found it hard to be as polite as you have been.

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


Re: [PATCH v2] mm/memcg: use exist interface to get css from memcg

2012-07-19 Thread Wanpeng Li
On Thu, Jul 19, 2012 at 07:23:07PM +0900, Kamezawa Hiroyuki wrote:
>(2012/07/19 18:29), Michal Hocko wrote:
>>On Wed 18-07-12 19:08:54, Wanpeng Li wrote:
>>>use exist interface mem_cgroup_css instead of &mem->css.
>>
>>This interface has been added to enable mem->css outside of
>>mm/memcontrol.c (where we define struct mem_cgroup). There is one user
>>left (hwpoison_filter_task) after recent clean ups.
>>
>>I think we shouldn't spread the usage inside the mm/memcontrol.c. The
>>compiler inlines the function for all callers added by this patch but I
>>wouldn't rely on it. It is also unfortunate that we cannot convert all
>>dereferences (e.g. const mem_cgroup).
>>Moreover it doesn't add any additional type safety. So I would vote for
>>not taking the patch but if others like it I will not block it.
>>
>
>Agreed.

Hmm, I see, thank you. :-)

Best Regards,
Wanpeng Li
>
>-Kame

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


Re: [PATCH v2] mm/memcg: use exist interface to get css from memcg

2012-07-19 Thread Kamezawa Hiroyuki

(2012/07/19 18:29), Michal Hocko wrote:

On Wed 18-07-12 19:08:54, Wanpeng Li wrote:

use exist interface mem_cgroup_css instead of &mem->css.


This interface has been added to enable mem->css outside of
mm/memcontrol.c (where we define struct mem_cgroup). There is one user
left (hwpoison_filter_task) after recent clean ups.

I think we shouldn't spread the usage inside the mm/memcontrol.c. The
compiler inlines the function for all callers added by this patch but I
wouldn't rely on it. It is also unfortunate that we cannot convert all
dereferences (e.g. const mem_cgroup).
Moreover it doesn't add any additional type safety. So I would vote for
not taking the patch but if others like it I will not block it.



Agreed.

-Kame

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


Re: [PATCH v2] mm/memcg: use exist interface to get css from memcg

2012-07-19 Thread Michal Hocko
On Wed 18-07-12 19:08:54, Wanpeng Li wrote:
> use exist interface mem_cgroup_css instead of &mem->css.

This interface has been added to enable mem->css outside of
mm/memcontrol.c (where we define struct mem_cgroup). There is one user
left (hwpoison_filter_task) after recent clean ups.

I think we shouldn't spread the usage inside the mm/memcontrol.c. The
compiler inlines the function for all callers added by this patch but I
wouldn't rely on it. It is also unfortunate that we cannot convert all
dereferences (e.g. const mem_cgroup).
Moreover it doesn't add any additional type safety. So I would vote for
not taking the patch but if others like it I will not block it.

> V2->V1:
> * restore the mistake modify in mem_cgroup_css
> 
> Signed-off-by: Wanpeng Li 
> ---
>  mm/memcontrol.c |   89 --
>  1 files changed, 46 insertions(+), 43 deletions(-)
> 
> diff --git a/mm/memcontrol.c b/mm/memcontrol.c
> index f72b5e5..f27084c 100644
> --- a/mm/memcontrol.c
> +++ b/mm/memcontrol.c
> @@ -645,7 +645,7 @@ retry:
>*/
>   __mem_cgroup_remove_exceeded(mz->memcg, mz, mctz);
>   if (!res_counter_soft_limit_excess(&mz->memcg->res) ||
> - !css_tryget(&mz->memcg->css))
> + !css_tryget(mem_cgroup_css(mz->memcg)))
>   goto retry;
>  done:
>   return mz;
> @@ -899,7 +899,7 @@ struct mem_cgroup *try_get_mem_cgroup_from_mm(struct 
> mm_struct *mm)
>   memcg = mem_cgroup_from_task(rcu_dereference(mm->owner));
>   if (unlikely(!memcg))
>   break;
> - } while (!css_tryget(&memcg->css));
> + } while (!css_tryget(mem_cgroup_css(memcg)));
>   rcu_read_unlock();
>   return memcg;
>  }
> @@ -935,10 +935,10 @@ struct mem_cgroup *mem_cgroup_iter(struct mem_cgroup 
> *root,
>   root = root_mem_cgroup;
>  
>   if (prev && !reclaim)
> - id = css_id(&prev->css);
> + id = css_id(mem_cgroup_css(prev));
>  
>   if (prev && prev != root)
> - css_put(&prev->css);
> + css_put(mem_cgroup_css(prev));
>  
>   if (!root->use_hierarchy && root != root_mem_cgroup) {
>   if (prev)
> @@ -963,9 +963,10 @@ struct mem_cgroup *mem_cgroup_iter(struct mem_cgroup 
> *root,
>   }
>  
>   rcu_read_lock();
> - css = css_get_next(&mem_cgroup_subsys, id + 1, &root->css, &id);
> + css = css_get_next(&mem_cgroup_subsys, id + 1,
> + mem_cgroup_css(root), &id);
>   if (css) {
> - if (css == &root->css || css_tryget(css))
> + if (css == mem_cgroup_css(root) || css_tryget(css))
>   memcg = container_of(css,
>struct mem_cgroup, css);
>   } else
> @@ -997,7 +998,7 @@ void mem_cgroup_iter_break(struct mem_cgroup *root,
>   if (!root)
>   root = root_mem_cgroup;
>   if (prev && prev != root)
> - css_put(&prev->css);
> + css_put(mem_cgroup_css(prev));
>  }
>  
>  /*
> @@ -1150,7 +1151,8 @@ bool __mem_cgroup_same_or_subtree(const struct 
> mem_cgroup *root_memcg,
>   return true;
>   if (!root_memcg->use_hierarchy || !memcg)
>   return false;
> - return css_is_ancestor(&memcg->css, &root_memcg->css);
> + return css_is_ancestor(mem_cgroup_css(memcg),
> + &root_memcg->css);
>  }
>  
>  static bool mem_cgroup_same_or_subtree(const struct mem_cgroup *root_memcg,
> @@ -1183,7 +1185,7 @@ int task_in_mem_cgroup(struct task_struct *task, const 
> struct mem_cgroup *memcg)
>   task_lock(task);
>   curr = mem_cgroup_from_task(task);
>   if (curr)
> - css_get(&curr->css);
> + css_get(mem_cgroup_css(curr));
>   task_unlock(task);
>   }
>   if (!curr)
> @@ -1195,7 +1197,7 @@ int task_in_mem_cgroup(struct task_struct *task, const 
> struct mem_cgroup *memcg)
>* hierarchy(even if use_hierarchy is disabled in "memcg").
>*/
>   ret = mem_cgroup_same_or_subtree(memcg, curr);
> - css_put(&curr->css);
> + css_put(mem_cgroup_css(curr));
>   return ret;
>  }
>  
> @@ -1251,7 +1253,7 @@ static unsigned long mem_cgroup_margin(struct 
> mem_cgroup *memcg)
>  
>  int mem_cgroup_swappiness(struct mem_cgroup *memcg)
>  {
> - struct cgroup *cgrp = memcg->css.cgroup;
> + struct cgroup *cgrp = mem_cgroup_css(memcg)->cgroup;
>  
>   /* root ? */
>   if (cgrp->parent == NULL)
> @@ -1396,7 +1398,7 @@ void mem_cgroup_print_oom_info(struct mem_cgroup 
> *memcg, struct task_struct *p)
>  
>   rcu_read_lock();
>  
> - mem_cgrp = memcg->css.cgroup;
> + mem_cgrp = mem_cgroup_css(memcg)->cgroup;
>   task_cgrp = task_cgroup(p, mem_cgroup_subsys_id);
>  
>   ret = cgroup_path(task_cgrp