Re: [PATCH 4/4] mm/page_alloc : relieve zone->lock's pressure for memory free

[Minchan Kim <minchan.kim@gmail.com>]

On Mon, 11 Jan 2010 14:27:57 +0800
Huang Shijie <shijie8@gmail.com> wrote:

>   Move the __mod_zone_page_state out the guard region of
> the spinlock to relieve the pressure for memory free.
> 
>   Using the zone->lru_lock to replace the zone->lock for
> zone->pages_scanned and zone's flag ZONE_ALL_UNRECLAIMABLE.
> 
> Signed-off-by: Huang Shijie <shijie8@gmail.com>
> ---
>  mm/page_alloc.c |   33 +++++++++++++++++++++++----------
>  1 files changed, 23 insertions(+), 10 deletions(-)
> 
> diff --git a/mm/page_alloc.c b/mm/page_alloc.c
> index 290dfc3..dfd4be0 100644
> --- a/mm/page_alloc.c
> +++ b/mm/page_alloc.c
> @@ -530,12 +530,9 @@ static void free_pcppages_bulk(struct zone *zone, int count,
>  {
>  	int migratetype = 0;
>  	int batch_free = 0;
> +	int free_ok = 0;
>  
>  	spin_lock(&zone->lock);
> -	zone_clear_flag(zone, ZONE_ALL_UNRECLAIMABLE);
> -	zone->pages_scanned = 0;
> -
> -	__mod_zone_page_state(zone, NR_FREE_PAGES, count);
>  	while (count) {
>  		struct page *page;
>  		struct list_head *list;
> @@ -558,23 +555,39 @@ static void free_pcppages_bulk(struct zone *zone, int count,
>  			page = list_entry(list->prev, struct page, lru);
>  			/* must delete as __free_one_page list manipulates */
>  			list_del(&page->lru);
> -			__free_one_page(page, zone, 0, migratetype);
> +			free_ok += __free_one_page(page, zone, 0, migratetype);
>  			trace_mm_page_pcpu_drain(page, 0, migratetype);
>  		} while (--count && --batch_free && !list_empty(list));
>  	}
>  	spin_unlock(&zone->lock);
> +
> +	if (likely(free_ok)) {
> +		spin_lock(&zone->lru_lock);
> +		zone_clear_flag(zone, ZONE_ALL_UNRECLAIMABLE);
> +		zone->pages_scanned = 0;
> +		spin_unlock(&zone->lru_lock);
> +
> +		__mod_zone_page_state(zone, NR_FREE_PAGES, free_ok);
> +	}
>  }
>  
>  static void free_one_page(struct zone *zone, struct page *page, int order,
>  				int migratetype)
>  {
> -	spin_lock(&zone->lock);
> -	zone_clear_flag(zone, ZONE_ALL_UNRECLAIMABLE);
> -	zone->pages_scanned = 0;
> +	int free_ok;
>  
> -	__mod_zone_page_state(zone, NR_FREE_PAGES, 1 << order);
> -	__free_one_page(page, zone, order, migratetype);
> +	spin_lock(&zone->lock);
> +	free_ok = __free_one_page(page, zone, order, migratetype);
>  	spin_unlock(&zone->lock);
> +
> +	if (likely(free_ok)) {
> +		spin_lock(&zone->lru_lock);
> +		zone_clear_flag(zone, ZONE_ALL_UNRECLAIMABLE);
> +		zone->pages_scanned = 0;
> +		spin_unlock(&zone->lru_lock);
> +
> +		__mod_zone_page_state(zone, NR_FREE_PAGES, free_ok << order);
> +	}
>  }
>  
>  static void __free_pages_ok(struct page *page, unsigned int order)
> -- 
> 1.6.5.2
> 

Thanks, Huang. 

Frankly speaking, I am not sure this ir right way.
This patch is adding to fine-grained locking overhead

As you know, this functions are one of hot pathes.
In addition, we didn't see the any problem, until now.
It means out of synchronization in ZONE_ALL_UNRECLAIMABLE 
and pages_scanned are all right?

If it is, we can move them out of zone->lock, too.
If it isn't, we need one more lock, then. 

Let's listen other mm guys's opinion.

-- 
Kind regards,
Minchan Kim

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>