Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
Date:   Mon, 23 Mar 2020 13:17:44 -0400
From:   Johannes Weiner <hannes@cmpxchg.org>
To:     js1304@gmail.com
Cc:     Andrew Morton <akpm@linux-foundation.org>, linux-mm@kvack.org,
        linux-kernel@vger.kernel.org, Michal Hocko <mhocko@kernel.org>,
        Hugh Dickins <hughd@google.com>,
        Minchan Kim <minchan@kernel.org>,
        Vlastimil Babka <vbabka@suse.cz>,
        Mel Gorman <mgorman@techsingularity.net>, kernel-team@lge.com,
        Joonsoo Kim <iamjoonsoo.kim@lge.com>
Subject: Re: [PATCH v4 6/8] mm/swap: implement workingset detection for
 anonymous LRU
Message-ID: <20200323171744.GD204561@cmpxchg.org>
References: <1584942732-2184-1-git-send-email-iamjoonsoo.kim@lge.com>
 <1584942732-2184-7-git-send-email-iamjoonsoo.kim@lge.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <1584942732-2184-7-git-send-email-iamjoonsoo.kim@lge.com>
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Mon, Mar 23, 2020 at 02:52:10PM +0900, js1304@gmail.com wrote:
> From: Joonsoo Kim <iamjoonsoo.kim@lge.com>
> 
> This patch implements workingset detection for anonymous LRU.
> All the infrastructure is implemented by the previous patches so this patch
> just activates the workingset detection by installing/retrieving
> the shadow entry.
> 
> Signed-off-by: Joonsoo Kim <iamjoonsoo.kim@lge.com>
> ---
>  include/linux/swap.h |  6 ++++++
>  mm/memory.c          |  7 ++++++-
>  mm/swap_state.c      | 20 ++++++++++++++++++--
>  mm/vmscan.c          |  7 +++++--
>  4 files changed, 35 insertions(+), 5 deletions(-)
> 
> diff --git a/include/linux/swap.h b/include/linux/swap.h
> index 273de48..fb4772e 100644
> --- a/include/linux/swap.h
> +++ b/include/linux/swap.h
> @@ -408,6 +408,7 @@ extern struct address_space *swapper_spaces[];
>  extern unsigned long total_swapcache_pages(void);
>  extern void show_swap_cache_info(void);
>  extern int add_to_swap(struct page *page);
> +extern void *get_shadow_from_swap_cache(swp_entry_t entry);
>  extern int add_to_swap_cache(struct page *page, swp_entry_t entry,
>  			gfp_t gfp, void **shadowp);
>  extern int __add_to_swap_cache(struct page *page, swp_entry_t entry);
> @@ -566,6 +567,11 @@ static inline int add_to_swap(struct page *page)
>  	return 0;
>  }
>  
> +static inline void *get_shadow_from_swap_cache(swp_entry_t entry)
> +{
> +	return NULL;
> +}
> +
>  static inline int add_to_swap_cache(struct page *page, swp_entry_t entry,
>  					gfp_t gfp_mask, void **shadowp)
>  {
> diff --git a/mm/memory.c b/mm/memory.c
> index 5f7813a..91a2097 100644
> --- a/mm/memory.c
> +++ b/mm/memory.c
> @@ -2925,10 +2925,15 @@ vm_fault_t do_swap_page(struct vm_fault *vmf)
>  			page = alloc_page_vma(GFP_HIGHUSER_MOVABLE, vma,
>  							vmf->address);
>  			if (page) {
> +				void *shadow;
> +
>  				__SetPageLocked(page);
>  				__SetPageSwapBacked(page);
>  				set_page_private(page, entry.val);
> -				lru_cache_add_anon(page);
> +				shadow = get_shadow_from_swap_cache(entry);
> +				if (shadow)
> +					workingset_refault(page, shadow);

Hm, this is calling workingset_refault() on a page that isn't charged
to a cgroup yet. That means the refault stats and inactive age counter
will be bumped incorrectly in the root cgroup instead of the real one.

> +				lru_cache_add(page);
>  				swap_readpage(page, true);
>  			}
>  		} else {

You need to look up and remember the shadow entry at the top and call
workingset_refault() after mem_cgroup_commit_charge() has run.

It'd be nice if we could do the shadow lookup for everybody in
lookup_swap_cache(), but that's subject to race conditions if multiple
faults on the same swap page happen in multiple vmas concurrently. The
swapcache bypass scenario is only safe because it checks that there is
a single pte under the mmap sem to prevent forking. So it looks like
you have to bubble up the shadow entry through swapin_readahead().