Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
Date:   Thu, 5 Nov 2020 10:50:28 -0500
From:   Johannes Weiner <hannes@cmpxchg.org>
To:     Muchun Song <songmuchun@bytedance.com>
Cc:     akpm@linux-foundation.org, shakeelb@google.com, guro@fb.com,
        mhocko@suse.com, laoar.shao@gmail.com, chris@chrisdown.name,
        tj@kernel.org, linux-kernel@vger.kernel.org
Subject: Re: [PATCH] mm: memcontrol: fix missing wakeup polling thread
Message-ID: <20201105155028.GD744831@cmpxchg.org>
References: <20201104125558.49472-1-songmuchun@bytedance.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20201104125558.49472-1-songmuchun@bytedance.com>
Precedence: bulk

On Wed, Nov 04, 2020 at 08:55:58PM +0800, Muchun Song wrote:
> When wen poll the memory.swap.events, we can miss being waked up when the
> swap event occurs. Because we didn't notify.
> 
> Fixes: f3a53a3a1e5b ("mm, memcontrol: implement memory.swap.events")
> Signed-off-by: Muchun Song <songmuchun@bytedance.com>

Good catch!

> ---
>  include/linux/memcontrol.h | 18 ++++++++++++++++--
>  1 file changed, 16 insertions(+), 2 deletions(-)
> 
> diff --git a/include/linux/memcontrol.h b/include/linux/memcontrol.h
> index 0f4dd7829fb2..2456cb737329 100644
> --- a/include/linux/memcontrol.h
> +++ b/include/linux/memcontrol.h
> @@ -1071,15 +1071,29 @@ static inline void count_memcg_event_mm(struct mm_struct *mm,
>  	rcu_read_unlock();
>  }
>  
> +static inline bool is_swap_memory_event(enum memcg_memory_event event)
> +{
> +	return event == MEMCG_SWAP_HIGH || event == MEMCG_SWAP_MAX ||
> +	       event == MEMCG_SWAP_FAIL;
> +}

Please inline this, it's not really worth the indirection.

>  static inline void memcg_memory_event(struct mem_cgroup *memcg,
>  				      enum memcg_memory_event event)
>  {
> +	bool swap_event = is_swap_memory_event(event);
> +	struct cgroup_file *cfile;
> +
>  	atomic_long_inc(&memcg->memory_events_local[event]);
> -	cgroup_file_notify(&memcg->events_local_file);
> +	if (swap_event) {
> +		cfile = &memcg->swap_events_file;
> +	} else {
> +		cfile = &memcg->events_file;
> +		cgroup_file_notify(&memcg->events_local_file);
> +	}
>  
>  	do {
>  		atomic_long_inc(&memcg->memory_events[event]);
> -		cgroup_file_notify(&memcg->events_file);
> +		cgroup_file_notify(cfile);

This loop is a walk up the hierarchy and memcg keeps changing, so you
cannot cache cfile up front.

		if (swap_event)
			cgroup_file_notify(&memcg->swap_events_file);
		else
			cgroup_file_notify(&memcg->events_file);