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 for Android: free password hash cracker in your pocket
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date:   Fri, 30 Sep 2016 11:37:37 +0900
From:   Minchan Kim <minchan@...nel.org>
To:     "ming.ling" <ming.ling@...eadtrum.com>
Cc:     akpm@...ux-foundation.org, mgorman@...hsingularity.net,
        vbabka@...e.cz, hannes@...xchg.org, mhocko@...e.com,
        baiyaowei@...s.chinamobile.com, iamjoonsoo.kim@....com,
        rientjes@...gle.com, hughd@...gle.com,
        kirill.shutemov@...ux.intel.com, riel@...hat.com, mgorman@...e.de,
        aquini@...hat.com, corbet@....net, linux-mm@...ck.org,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH] mm: exclude isolated non-lru pages from NR_ISOLATED_ANON
 or NR_ISOLATED_FILE.

Hello,

On Wed, Sep 28, 2016 at 05:31:03PM +0800, ming.ling wrote:
> Non-lru pages don't belong to any lru, so accounting them to
> NR_ISOLATED_ANON or NR_ISOLATED_FILE doesn't make any sense.
> It may misguide functions such as pgdat_reclaimable_pages and
> too_many_isolated.

I agree this part. It would be happier if you give any story you suffered
from. Although you don't have, it's okay because you are correcting
clearly wrong part. Thanks. :)

> 
> This patch adds NR_ISOLATED_NONLRU to vmstat and moves isolated non-lru
> pages from NR_ISOLATED_ANON or NR_ISOLATED_FILE to NR_ISOLATED_NONLRU.
> And with non-lru pages in vmstat, it helps to optimize algorithm of
> function too_many_isolated oneday.

Need more justfication to add new vmstat because once we add it, it's
really hard to change/remove it(i.e., maintainace trobule) so I want
to add it when it really would be helpful sometime, not now.

Could you resend the patch without part adding new vmstat?

Thanks.

> 
> Signed-off-by: ming.ling <ming.ling@...eadtrum.com>
> ---
>  include/linux/mmzone.h |  1 +
>  mm/compaction.c        | 12 +++++++++---
>  mm/migrate.c           | 14 ++++++++++----
>  3 files changed, 20 insertions(+), 7 deletions(-)
> 
> diff --git a/include/linux/mmzone.h b/include/linux/mmzone.h
> index 7f2ae99..dc0adba 100644
> --- a/include/linux/mmzone.h
> +++ b/include/linux/mmzone.h
> @@ -169,6 +169,7 @@ enum node_stat_item {
>  	NR_VMSCAN_IMMEDIATE,	/* Prioritise for reclaim when writeback ends */
>  	NR_DIRTIED,		/* page dirtyings since bootup */
>  	NR_WRITTEN,		/* page writings since bootup */
> +	NR_ISOLATED_NONLRU,	/* Temporary isolated pages from non-lru */
>  	NR_VM_NODE_STAT_ITEMS
>  };
>  
> diff --git a/mm/compaction.c b/mm/compaction.c
> index 9affb29..8da1dca 100644
> --- a/mm/compaction.c
> +++ b/mm/compaction.c
> @@ -638,16 +638,21 @@ isolate_freepages_range(struct compact_control *cc,
>  static void acct_isolated(struct zone *zone, struct compact_control *cc)
>  {
>  	struct page *page;
> -	unsigned int count[2] = { 0, };
> +	unsigned int count[3] = { 0, };
>  
>  	if (list_empty(&cc->migratepages))
>  		return;
>  
> -	list_for_each_entry(page, &cc->migratepages, lru)
> -		count[!!page_is_file_cache(page)]++;
> +	list_for_each_entry(page, &cc->migratepages, lru) {
> +		if (PageLRU(page))
> +			count[!!page_is_file_cache(page)]++;
> +		else
> +			count[2]++;
> +	}
>  
>  	mod_node_page_state(zone->zone_pgdat, NR_ISOLATED_ANON, count[0]);
>  	mod_node_page_state(zone->zone_pgdat, NR_ISOLATED_FILE, count[1]);
> +	mod_node_page_state(zone->zone_pgdat, NR_ISOLATED_NONLRU, count[2]);
>  }
>  
>  /* Similar to reclaim, but different enough that they don't share logic */
> @@ -659,6 +664,7 @@ static bool too_many_isolated(struct zone *zone)
>  			node_page_state(zone->zone_pgdat, NR_INACTIVE_ANON);
>  	active = node_page_state(zone->zone_pgdat, NR_ACTIVE_FILE) +
>  			node_page_state(zone->zone_pgdat, NR_ACTIVE_ANON);
> +	/* Is it necessary to add NR_ISOLATED_NONLRU?? */
>  	isolated = node_page_state(zone->zone_pgdat, NR_ISOLATED_FILE) +
>  			node_page_state(zone->zone_pgdat, NR_ISOLATED_ANON);
>  
> diff --git a/mm/migrate.c b/mm/migrate.c
> index f7ee04a..cd5abb2 100644
> --- a/mm/migrate.c
> +++ b/mm/migrate.c
> @@ -168,8 +168,11 @@ void putback_movable_pages(struct list_head *l)
>  			continue;
>  		}
>  		list_del(&page->lru);
> -		dec_node_page_state(page, NR_ISOLATED_ANON +
> -				page_is_file_cache(page));
> +		if (PageLRU(page))
> +			dec_node_page_state(page, NR_ISOLATED_ANON +
> +					page_is_file_cache(page));
> +		else
> +			dec_node_page_state(page, NR_ISOLATED_NONLRU);
>  		/*
>  		 * We isolated non-lru movable page so here we can use
>  		 * __PageMovable because LRU page's mapping cannot have
> @@ -1121,8 +1124,11 @@ out:
>  		 * restored.
>  		 */
>  		list_del(&page->lru);
> -		dec_node_page_state(page, NR_ISOLATED_ANON +
> -				page_is_file_cache(page));
> +		if (PageLRU(page))
> +			dec_node_page_state(page, NR_ISOLATED_ANON +
> +					page_is_file_cache(page));
> +		else
> +			dec_node_page_state(page, NR_ISOLATED_NONLRU);
>  	}
>  
>  	/*
> -- 
> 1.9.1
> 
> --
> To unsubscribe, send a message with 'unsubscribe linux-mm' in
> the body to majordomo@...ck.org.  For more info on Linux MM,
> see: http://www.linux-mm.org/ .
> Don't email: <a href=mailto:"dont@...ck.org"> email@...ck.org </a>

Powered by blists - more mailing lists