linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* Re: [PATCH 06/27] mm, vmscan: Make kswapd reclaim in terms of nodes
       [not found] <02fe01d1c48b$c44e9e80$4cebdb80$@alibaba-inc.com>
@ 2016-06-12  9:33 ` Hillf Danton
  2016-06-14 14:52   ` Mel Gorman
  0 siblings, 1 reply; 9+ messages in thread
From: Hillf Danton @ 2016-06-12  9:33 UTC (permalink / raw)
  To: Mel Gorman; +Cc: linux-kernel, linux-mm

> 
>  /*
> - * kswapd shrinks the zone by the number of pages required to reach
> - * the high watermark.
> + * kswapd shrinks a node of pages that are at or below the highest usable
> + * zone that is currently unbalanced.
>   *
>   * Returns true if kswapd scanned at least the requested number of pages to
>   * reclaim or if the lack of progress was due to pages under writeback.
>   * This is used to determine if the scanning priority needs to be raised.
>   */
> -static bool kswapd_shrink_zone(struct zone *zone,
> +static bool kswapd_shrink_node(pg_data_t *pgdat,
>  			       int classzone_idx,
>  			       struct scan_control *sc)
>  {
> -	unsigned long balance_gap;
> -	bool lowmem_pressure;
> -	struct pglist_data *pgdat = zone->zone_pgdat;
> +	struct zone *zone;
> +	unsigned long nr_to_reclaim = 0;
> +	int z;
> 
> -	/* Reclaim above the high watermark. */
> -	sc->nr_to_reclaim = max(SWAP_CLUSTER_MAX, high_wmark_pages(zone));
> +	/* Reclaim a number of pages proportional to the number of zones */
> +	for (z = 0; z <= classzone_idx; z++) {
> +		zone = pgdat->node_zones + z;
> +		if (!populated_zone(zone))
> +			continue;
> 
> -	/*
> -	 * We put equal pressure on every zone, unless one zone has way too
> -	 * many pages free already. The "too many pages" is defined as the
> -	 * high wmark plus a "gap" where the gap is either the low
> -	 * watermark or 1% of the zone, whichever is smaller.
> -	 */
> -	balance_gap = min(low_wmark_pages(zone), DIV_ROUND_UP(
> -			zone->managed_pages, KSWAPD_ZONE_BALANCE_GAP_RATIO));
> +		nr_to_reclaim += max(high_wmark_pages(zone), SWAP_CLUSTER_MAX);
> +	}

Missing sc->nr_to_reclaim = nr_to_reclaim; ?

> 
>  	/*
> -	 * If there is no low memory pressure or the zone is balanced then no
> -	 * reclaim is necessary
> +	 * Historically care was taken to put equal pressure on all zones but
> +	 * now pressure is applied based on node LRU order.
>  	 */
> -	lowmem_pressure = (buffer_heads_over_limit && is_highmem(zone));
> -	if (!lowmem_pressure && zone_balanced(zone, sc->order, false,
> -						balance_gap, classzone_idx))
> -		return true;
> -
> -	shrink_node(zone->zone_pgdat, sc, classzone_idx);
> -
> -	/* TODO: ANOMALY */
> -	clear_bit(PGDAT_WRITEBACK, &pgdat->flags);
> +	shrink_node(pgdat, sc, classzone_idx);
> 
>  	/*
> -	 * If a zone reaches its high watermark, consider it to be no longer
> -	 * congested. It's possible there are dirty pages backed by congested
> -	 * BDIs but as pressure is relieved, speculatively avoid congestion
> -	 * waits.
> +	 * Fragmentation may mean that the system cannot be rebalanced for
> +	 * high-order allocations. If twice the allocation size has been
> +	 * reclaimed then recheck watermarks only at order-0 to prevent
> +	 * excessive reclaim. Assume that a process requested a high-order
> +	 * can direct reclaim/compact.
>  	 */
> -	if (pgdat_reclaimable(zone->zone_pgdat) &&
> -	    zone_balanced(zone, sc->order, false, 0, classzone_idx)) {
> -		clear_bit(PGDAT_CONGESTED, &pgdat->flags);
> -		clear_bit(PGDAT_DIRTY, &pgdat->flags);
> -	}
> +	if (sc->order && sc->nr_reclaimed >= 2UL << sc->order)
> +		sc->order = 0;
> 
>  	return sc->nr_scanned >= sc->nr_to_reclaim;
>  }

^ permalink raw reply	[flat|nested] 9+ messages in thread

* Re: [PATCH 06/27] mm, vmscan: Make kswapd reclaim in terms of nodes
  2016-06-12  9:33 ` [PATCH 06/27] mm, vmscan: Make kswapd reclaim in terms of nodes Hillf Danton
@ 2016-06-14 14:52   ` Mel Gorman
  0 siblings, 0 replies; 9+ messages in thread
From: Mel Gorman @ 2016-06-14 14:52 UTC (permalink / raw)
  To: Hillf Danton; +Cc: linux-kernel, linux-mm

On Sun, Jun 12, 2016 at 05:33:24PM +0800, Hillf Danton wrote:
> > -	/*
> > -	 * We put equal pressure on every zone, unless one zone has way too
> > -	 * many pages free already. The "too many pages" is defined as the
> > -	 * high wmark plus a "gap" where the gap is either the low
> > -	 * watermark or 1% of the zone, whichever is smaller.
> > -	 */
> > -	balance_gap = min(low_wmark_pages(zone), DIV_ROUND_UP(
> > -			zone->managed_pages, KSWAPD_ZONE_BALANCE_GAP_RATIO));
> > +		nr_to_reclaim += max(high_wmark_pages(zone), SWAP_CLUSTER_MAX);
> > +	}
> 
> Missing sc->nr_to_reclaim = nr_to_reclaim; ?
> 

Yes. It may explain why I saw lower than expected kswapd in more
detailed tests recently. Thanks.

-- 
Mel Gorman
SUSE Labs

^ permalink raw reply	[flat|nested] 9+ messages in thread

* Re: [PATCH 06/27] mm, vmscan: Make kswapd reclaim in terms of nodes
  2016-06-22  8:42 ` Hillf Danton
@ 2016-06-23 11:31   ` Mel Gorman
  0 siblings, 0 replies; 9+ messages in thread
From: Mel Gorman @ 2016-06-23 11:31 UTC (permalink / raw)
  To: Hillf Danton; +Cc: Johannes Weiner, Vlastimil Babka, linux-kernel, linux-mm

On Wed, Jun 22, 2016 at 04:42:06PM +0800, Hillf Danton wrote:
> >  	/*
> > -	 * If a zone reaches its high watermark, consider it to be no longer
> > -	 * congested. It's possible there are dirty pages backed by congested
> > -	 * BDIs but as pressure is relieved, speculatively avoid congestion
> > -	 * waits.
> > +	 * Fragmentation may mean that the system cannot be rebalanced for
> > +	 * high-order allocations. If twice the allocation size has been
> > +	 * reclaimed then recheck watermarks only at order-0 to prevent
> > +	 * excessive reclaim. Assume that a process requested a high-order
> > +	 * can direct reclaim/compact.
> >  	 */
> > -	if (pgdat_reclaimable(zone->zone_pgdat) &&
> > -	    zone_balanced(zone, sc->order, false, 0, classzone_idx)) {
> > -		clear_bit(PGDAT_CONGESTED, &pgdat->flags);
> > -		clear_bit(PGDAT_DIRTY, &pgdat->flags);
> > -	}
> > +	if (sc->order && sc->nr_reclaimed >= 2UL << sc->order)
> > +		sc->order = 0;
> > 
> 
> Reclaim order is changed here.
> Btw, I find no such change in current code.
> 

It is reintroducing a check removed by commit accf62422b3a ("mm, kswapd: replace
kswapd compaction with waking up kcompactd"). That patch had kswapd
always check at order-0 once kswapd is awake in pgdat_balanced but would
still take at least one pass through reclaiming so kcompactd potentially
makes progress.

This patch removes pgdat_balanced entirely and zone_balanced() checks the
order it is asked like it used to. Hence, it is necessary to reset sc->order
once progress is made or kswapd potentially stays awake reclaiming pages
until a high-order page is freed.

-- 
Mel Gorman
SUSE Labs

^ permalink raw reply	[flat|nested] 9+ messages in thread

* Re: [PATCH 06/27] mm, vmscan: Make kswapd reclaim in terms of nodes
       [not found] <071801d1cc5c$245087d0$6cf19770$@alibaba-inc.com>
@ 2016-06-22  8:42 ` Hillf Danton
  2016-06-23 11:31   ` Mel Gorman
  0 siblings, 1 reply; 9+ messages in thread
From: Hillf Danton @ 2016-06-22  8:42 UTC (permalink / raw)
  To: Mel Gorman; +Cc: Johannes Weiner, Vlastimil Babka, linux-kernel, linux-mm

>  /*
> - * kswapd shrinks the zone by the number of pages required to reach
> - * the high watermark.
> + * kswapd shrinks a node of pages that are at or below the highest usable
> + * zone that is currently unbalanced.
>   *
>   * Returns true if kswapd scanned at least the requested number of pages to
>   * reclaim or if the lack of progress was due to pages under writeback.
>   * This is used to determine if the scanning priority needs to be raised.
>   */
> -static bool kswapd_shrink_zone(struct zone *zone,
> +static bool kswapd_shrink_node(pg_data_t *pgdat,
>  			       int classzone_idx,
>  			       struct scan_control *sc)
>  {
> -	unsigned long balance_gap;
> -	bool lowmem_pressure;
> -	struct pglist_data *pgdat = zone->zone_pgdat;
> +	struct zone *zone;
> +	int z;
> 
> -	/* Reclaim above the high watermark. */
> -	sc->nr_to_reclaim = max(SWAP_CLUSTER_MAX, high_wmark_pages(zone));
> +	/* Reclaim a number of pages proportional to the number of zones */
> +	sc->nr_to_reclaim = 0;
> +	for (z = 0; z <= classzone_idx; z++) {
> +		zone = pgdat->node_zones + z;
> +		if (!populated_zone(zone))
> +			continue;
> 
> -	/*
> -	 * We put equal pressure on every zone, unless one zone has way too
> -	 * many pages free already. The "too many pages" is defined as the
> -	 * high wmark plus a "gap" where the gap is either the low
> -	 * watermark or 1% of the zone, whichever is smaller.
> -	 */
> -	balance_gap = min(low_wmark_pages(zone), DIV_ROUND_UP(
> -			zone->managed_pages, KSWAPD_ZONE_BALANCE_GAP_RATIO));
> +		sc->nr_to_reclaim += max(high_wmark_pages(zone), SWAP_CLUSTER_MAX);
> +	}
> 
>  	/*
> -	 * If there is no low memory pressure or the zone is balanced then no
> -	 * reclaim is necessary
> +	 * Historically care was taken to put equal pressure on all zones but
> +	 * now pressure is applied based on node LRU order.
>  	 */
> -	lowmem_pressure = (buffer_heads_over_limit && is_highmem(zone));
> -	if (!lowmem_pressure && zone_balanced(zone, sc->order, false,
> -						balance_gap, classzone_idx))
> -		return true;
> -
> -	shrink_node(zone->zone_pgdat, sc, classzone_idx);
> -
> -	/* TODO: ANOMALY */
> -	clear_bit(PGDAT_WRITEBACK, &pgdat->flags);
> +	shrink_node(pgdat, sc, classzone_idx);
> 
>  	/*
> -	 * If a zone reaches its high watermark, consider it to be no longer
> -	 * congested. It's possible there are dirty pages backed by congested
> -	 * BDIs but as pressure is relieved, speculatively avoid congestion
> -	 * waits.
> +	 * Fragmentation may mean that the system cannot be rebalanced for
> +	 * high-order allocations. If twice the allocation size has been
> +	 * reclaimed then recheck watermarks only at order-0 to prevent
> +	 * excessive reclaim. Assume that a process requested a high-order
> +	 * can direct reclaim/compact.
>  	 */
> -	if (pgdat_reclaimable(zone->zone_pgdat) &&
> -	    zone_balanced(zone, sc->order, false, 0, classzone_idx)) {
> -		clear_bit(PGDAT_CONGESTED, &pgdat->flags);
> -		clear_bit(PGDAT_DIRTY, &pgdat->flags);
> -	}
> +	if (sc->order && sc->nr_reclaimed >= 2UL << sc->order)
> +		sc->order = 0;
> 

Reclaim order is changed here.
Btw, I find no such change in current code.

>  	return sc->nr_scanned >= sc->nr_to_reclaim;
>  }
> 
>  /*
> - * For kswapd, balance_pgdat() will work across all this node's zones until
> - * they are all at high_wmark_pages(zone).
> - *
> - * Returns the highest zone idx kswapd was reclaiming at
> + * For kswapd, balance_pgdat() will reclaim pages across a node from zones
> + * that are eligible for use by the caller until at least one zone is
> + * balanced.
>   *
> - * There is special handling here for zones which are full of pinned pages.
> - * This can happen if the pages are all mlocked, or if they are all used by
> - * device drivers (say, ZONE_DMA).  Or if they are all in use by hugetlb.
> - * What we do is to detect the case where all pages in the zone have been
> - * scanned twice and there has been zero successful reclaim.  Mark the zone as
> - * dead and from now on, only perform a short scan.  Basically we're polling
> - * the zone for when the problem goes away.
> + * Returns the order kswapd finished reclaiming at.
>   *
>   * kswapd scans the zones in the highmem->normal->dma direction.  It skips
>   * zones which have free_pages > high_wmark_pages(zone), but once a zone is
> - * found to have free_pages <= high_wmark_pages(zone), we scan that zone and the
> - * lower zones regardless of the number of free pages in the lower zones. This
> - * interoperates with the page allocator fallback scheme to ensure that aging
> - * of pages is balanced across the zones.
> + * found to have free_pages <= high_wmark_pages(zone), any page is that zone
> + * or lower is eligible for reclaim until at least one usable zone is
> + * balanced.
>   */
>  static int balance_pgdat(pg_data_t *pgdat, int order, int classzone_idx)
>  {
>  	int i;
> -	int end_zone = 0;	/* Inclusive.  0 = ZONE_DMA */
>  	unsigned long nr_soft_reclaimed;
>  	unsigned long nr_soft_scanned;
> +	struct zone *zone;
>  	struct scan_control sc = {
>  		.gfp_mask = GFP_KERNEL,
> -		.reclaim_idx = MAX_NR_ZONES - 1,
>  		.order = order,
>  		.priority = DEF_PRIORITY,
>  		.may_writepage = !laptop_mode,
>  		.may_unmap = 1,
>  		.may_swap = 1,
> +		.reclaim_idx = classzone_idx,
>  	};
>  	count_vm_event(PAGEOUTRUN);
> 
> @@ -3203,21 +3125,10 @@ static int balance_pgdat(pg_data_t *pgdat, int order, int classzone_idx)
> 
>  		/* Scan from the highest requested zone to dma */
>  		for (i = classzone_idx; i >= 0; i--) {
> -			struct zone *zone = pgdat->node_zones + i;
> -
> +			zone = pgdat->node_zones + i;
>  			if (!populated_zone(zone))
>  				continue;
> 
> -			if (sc.priority != DEF_PRIORITY &&
> -			    !pgdat_reclaimable(zone->zone_pgdat))
> -				continue;
> -
> -			/*
> -			 * Do some background aging of the anon list, to give
> -			 * pages a chance to be referenced before reclaiming.
> -			 */
> -			age_active_anon(zone, &sc);
> -
>  			/*
>  			 * If the number of buffer_heads in the machine
>  			 * exceeds the maximum allowed level and this node
> @@ -3225,19 +3136,17 @@ static int balance_pgdat(pg_data_t *pgdat, int order, int classzone_idx)
>  			 * it to relieve lowmem pressure.
>  			 */
>  			if (buffer_heads_over_limit && is_highmem_idx(i)) {
> -				end_zone = i;
> +				classzone_idx = i;
>  				break;
>  			}
> 
> -			if (!zone_balanced(zone, order, false, 0, 0)) {
> -				end_zone = i;
> +			if (!zone_balanced(zone, order, 0, 0)) {

We need to sync order with the above change?

> +				classzone_idx = i;
>  				break;
>  			} else {
>  				/*
> -				 * If balanced, clear the dirty and congested
> -				 * flags
> -				 *
> -				 * TODO: ANOMALY
> +				 * If any eligible zone is balanced then the
> +				 * node is not considered congested or dirty.
>  				 */
>  				clear_bit(PGDAT_CONGESTED, &zone->zone_pgdat->flags);
>  				clear_bit(PGDAT_DIRTY, &zone->zone_pgdat->flags);
> @@ -3248,51 +3157,34 @@ static int balance_pgdat(pg_data_t *pgdat, int order, int classzone_idx)
>  			goto out;
> 
>  		/*
> +		 * Do some background aging of the anon list, to give
> +		 * pages a chance to be referenced before reclaiming. All
> +		 * pages are rotated regardless of classzone as this is
> +		 * about consistent aging.
> +		 */
> +		age_active_anon(pgdat, &pgdat->node_zones[MAX_NR_ZONES - 1], &sc);
> +
> +		/*
>  		 * If we're getting trouble reclaiming, start doing writepage
>  		 * even in laptop mode.
>  		 */
> -		if (sc.priority < DEF_PRIORITY - 2)
> +		if (sc.priority < DEF_PRIORITY - 2 || !pgdat_reclaimable(pgdat))
>  			sc.may_writepage = 1;
> 
> +		/* Call soft limit reclaim before calling shrink_node. */
> +		sc.nr_scanned = 0;
> +		nr_soft_scanned = 0;
> +		nr_soft_reclaimed = mem_cgroup_soft_limit_reclaim(zone, sc.order,
> +						sc.gfp_mask, &nr_soft_scanned);
> +		sc.nr_reclaimed += nr_soft_reclaimed;
> +
>  		/*
> -		 * Continue scanning in the highmem->dma direction stopping at
> -		 * the last zone which needs scanning. This may reclaim lowmem
> -		 * pages that are not necessary for zone balancing but it
> -		 * preserves LRU ordering. It is assumed that the bulk of
> -		 * allocation requests can use arbitrary zones with the
> -		 * possible exception of big highmem:lowmem configurations.
> +		 * There should be no need to raise the scanning priority if
> +		 * enough pages are already being scanned that that high
> +		 * watermark would be met at 100% efficiency.
>  		 */
> -		for (i = end_zone; i >= 0; i--) {
> -			struct zone *zone = pgdat->node_zones + i;
> -
> -			if (!populated_zone(zone))
> -				continue;
> -
> -			if (sc.priority != DEF_PRIORITY &&
> -			    !pgdat_reclaimable(zone->zone_pgdat))
> -				continue;
> -
> -			sc.nr_scanned = 0;
> -			sc.reclaim_idx = i;
> -
> -			nr_soft_scanned = 0;
> -			/*
> -			 * Call soft limit reclaim before calling shrink_zone.
> -			 */
> -			nr_soft_reclaimed = mem_cgroup_soft_limit_reclaim(zone,
> -							order, sc.gfp_mask,
> -							&nr_soft_scanned);
> -			sc.nr_reclaimed += nr_soft_reclaimed;
> -
> -			/*
> -			 * There should be no need to raise the scanning
> -			 * priority if enough pages are already being scanned
> -			 * that that high watermark would be met at 100%
> -			 * efficiency.
> -			 */
> -			if (kswapd_shrink_zone(zone, end_zone, &sc))
> -				raise_priority = false;
> -		}
> +		if (kswapd_shrink_node(pgdat, classzone_idx, &sc))
> +			raise_priority = false;
> 
>  		/*
>  		 * If the low watermark is met there is no need for processes
> @@ -3308,20 +3200,37 @@ static int balance_pgdat(pg_data_t *pgdat, int order, int classzone_idx)
>  			break;
> 
>  		/*
> +		 * Stop reclaiming if any eligible zone is balanced and clear
> +		 * node writeback or congested.
> +		 */
> +		for (i = 0; i <= classzone_idx; i++) {
> +			zone = pgdat->node_zones + i;
> +			if (!populated_zone(zone))
> +				continue;
> +
> +			if (zone_balanced(zone, sc.order, 0, classzone_idx)) {
> +				clear_bit(PGDAT_CONGESTED, &pgdat->flags);
> +				clear_bit(PGDAT_DIRTY, &pgdat->flags);
> +				goto out;
> +			}
> +		}
> +
> +		/*
>  		 * Raise priority if scanning rate is too low or there was no
>  		 * progress in reclaiming pages
>  		 */
>  		if (raise_priority || !sc.nr_reclaimed)
>  			sc.priority--;
> -	} while (sc.priority >= 1 &&
> -			!pgdat_balanced(pgdat, order, classzone_idx));
> +	} while (sc.priority >= 1);
> 
>  out:
>  	/*
> -	 * Return the highest zone idx we were reclaiming at so
> -	 * prepare_kswapd_sleep() makes the same decisions as here.
> +	 * Return the order kswapd stopped reclaiming at as
> +	 * prepare_kswapd_sleep() takes it into account. If another caller
> +	 * entered the allocator slow path while kswapd was awake, order will
> +	 * remain at the higher level.
>  	 */
> -	return end_zone;
> +	return sc.order;
>  }
> 

^ permalink raw reply	[flat|nested] 9+ messages in thread

* [PATCH 06/27] mm, vmscan: Make kswapd reclaim in terms of nodes
  2016-06-21 14:15 [PATCH 00/27] Move LRU page reclaim from zones to nodes v7 Mel Gorman
@ 2016-06-21 14:15 ` Mel Gorman
  0 siblings, 0 replies; 9+ messages in thread
From: Mel Gorman @ 2016-06-21 14:15 UTC (permalink / raw)
  To: Andrew Morton, Linux-MM
  Cc: Rik van Riel, Vlastimil Babka, Johannes Weiner, LKML, Mel Gorman

Patch "mm: vmscan: Begin reclaiming pages on a per-node basis" started
thinking of reclaim in terms of nodes but kswapd is still zone-centric. This
patch gets rid of many of the node-based versus zone-based decisions.

o A node is considered balanced when any eligible lower zone is balanced.
  This eliminates one class of age-inversion problem because we avoid
  reclaiming a newer page just because it's in the wrong zone
o pgdat_balanced disappears because we now only care about one zone being
  balanced.
o Some anomalies related to writeback and congestion tracking being based on
  zones disappear.
o kswapd no longer has to take care to reclaim zones in the reverse order
  that the page allocator uses.
o Most importantly of all, reclaim from node 0 with multiple zones will
  have similar aging and reclaiming characteristics as every
  other node.

Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
Acked-by: Johannes Weiner <hannes@cmpxchg.org>
Acked-by: Vlastimil Babka <vbabka@suse.cz>
---
 mm/vmscan.c | 292 +++++++++++++++++++++---------------------------------------
 1 file changed, 101 insertions(+), 191 deletions(-)

diff --git a/mm/vmscan.c b/mm/vmscan.c
index b5b355db97cb..5873f5003078 100644
--- a/mm/vmscan.c
+++ b/mm/vmscan.c
@@ -2972,7 +2972,8 @@ unsigned long try_to_free_mem_cgroup_pages(struct mem_cgroup *memcg,
 }
 #endif
 
-static void age_active_anon(struct zone *zone, struct scan_control *sc)
+static void age_active_anon(struct pglist_data *pgdat,
+				struct zone *zone, struct scan_control *sc)
 {
 	struct mem_cgroup *memcg;
 
@@ -2991,85 +2992,15 @@ static void age_active_anon(struct zone *zone, struct scan_control *sc)
 	} while (memcg);
 }
 
-static bool zone_balanced(struct zone *zone, int order, bool highorder,
+static bool zone_balanced(struct zone *zone, int order,
 			unsigned long balance_gap, int classzone_idx)
 {
 	unsigned long mark = high_wmark_pages(zone) + balance_gap;
 
-	/*
-	 * When checking from pgdat_balanced(), kswapd should stop and sleep
-	 * when it reaches the high order-0 watermark and let kcompactd take
-	 * over. Other callers such as wakeup_kswapd() want to determine the
-	 * true high-order watermark.
-	 */
-	if (IS_ENABLED(CONFIG_COMPACTION) && !highorder) {
-		mark += (1UL << order);
-		order = 0;
-	}
-
 	return zone_watermark_ok_safe(zone, order, mark, classzone_idx);
 }
 
 /*
- * pgdat_balanced() is used when checking if a node is balanced.
- *
- * For order-0, all zones must be balanced!
- *
- * For high-order allocations only zones that meet watermarks and are in a
- * zone allowed by the callers classzone_idx are added to balanced_pages. The
- * total of balanced pages must be at least 25% of the zones allowed by
- * classzone_idx for the node to be considered balanced. Forcing all zones to
- * be balanced for high orders can cause excessive reclaim when there are
- * imbalanced zones.
- * The choice of 25% is due to
- *   o a 16M DMA zone that is balanced will not balance a zone on any
- *     reasonable sized machine
- *   o On all other machines, the top zone must be at least a reasonable
- *     percentage of the middle zones. For example, on 32-bit x86, highmem
- *     would need to be at least 256M for it to be balance a whole node.
- *     Similarly, on x86-64 the Normal zone would need to be at least 1G
- *     to balance a node on its own. These seemed like reasonable ratios.
- */
-static bool pgdat_balanced(pg_data_t *pgdat, int order, int classzone_idx)
-{
-	unsigned long managed_pages = 0;
-	unsigned long balanced_pages = 0;
-	int i;
-
-	/* Check the watermark levels */
-	for (i = 0; i <= classzone_idx; i++) {
-		struct zone *zone = pgdat->node_zones + i;
-
-		if (!populated_zone(zone))
-			continue;
-
-		managed_pages += zone->managed_pages;
-
-		/*
-		 * A special case here:
-		 *
-		 * balance_pgdat() skips over all_unreclaimable after
-		 * DEF_PRIORITY. Effectively, it considers them balanced so
-		 * they must be considered balanced here as well!
-		 */
-		if (!pgdat_reclaimable(zone->zone_pgdat)) {
-			balanced_pages += zone->managed_pages;
-			continue;
-		}
-
-		if (zone_balanced(zone, order, false, 0, i))
-			balanced_pages += zone->managed_pages;
-		else if (!order)
-			return false;
-	}
-
-	if (order)
-		return balanced_pages >= (managed_pages >> 2);
-	else
-		return true;
-}
-
-/*
  * Prepare kswapd for sleeping. This verifies that there are no processes
  * waiting in throttle_direct_reclaim() and that watermarks have been met.
  *
@@ -3078,6 +3009,8 @@ static bool pgdat_balanced(pg_data_t *pgdat, int order, int classzone_idx)
 static bool prepare_kswapd_sleep(pg_data_t *pgdat, int order, long remaining,
 					int classzone_idx)
 {
+	int i;
+
 	/* If a direct reclaimer woke kswapd within HZ/10, it's premature */
 	if (remaining)
 		return false;
@@ -3098,101 +3031,90 @@ static bool prepare_kswapd_sleep(pg_data_t *pgdat, int order, long remaining,
 	if (waitqueue_active(&pgdat->pfmemalloc_wait))
 		wake_up_all(&pgdat->pfmemalloc_wait);
 
-	return pgdat_balanced(pgdat, order, classzone_idx);
+	for (i = 0; i <= classzone_idx; i++) {
+		struct zone *zone = pgdat->node_zones + i;
+
+		if (!populated_zone(zone))
+			continue;
+
+		if (zone_balanced(zone, order, 0, classzone_idx))
+			return true;
+	}
+
+	return false;
 }
 
 /*
- * kswapd shrinks the zone by the number of pages required to reach
- * the high watermark.
+ * kswapd shrinks a node of pages that are at or below the highest usable
+ * zone that is currently unbalanced.
  *
  * Returns true if kswapd scanned at least the requested number of pages to
  * reclaim or if the lack of progress was due to pages under writeback.
  * This is used to determine if the scanning priority needs to be raised.
  */
-static bool kswapd_shrink_zone(struct zone *zone,
+static bool kswapd_shrink_node(pg_data_t *pgdat,
 			       int classzone_idx,
 			       struct scan_control *sc)
 {
-	unsigned long balance_gap;
-	bool lowmem_pressure;
-	struct pglist_data *pgdat = zone->zone_pgdat;
+	struct zone *zone;
+	int z;
 
-	/* Reclaim above the high watermark. */
-	sc->nr_to_reclaim = max(SWAP_CLUSTER_MAX, high_wmark_pages(zone));
+	/* Reclaim a number of pages proportional to the number of zones */
+	sc->nr_to_reclaim = 0;
+	for (z = 0; z <= classzone_idx; z++) {
+		zone = pgdat->node_zones + z;
+		if (!populated_zone(zone))
+			continue;
 
-	/*
-	 * We put equal pressure on every zone, unless one zone has way too
-	 * many pages free already. The "too many pages" is defined as the
-	 * high wmark plus a "gap" where the gap is either the low
-	 * watermark or 1% of the zone, whichever is smaller.
-	 */
-	balance_gap = min(low_wmark_pages(zone), DIV_ROUND_UP(
-			zone->managed_pages, KSWAPD_ZONE_BALANCE_GAP_RATIO));
+		sc->nr_to_reclaim += max(high_wmark_pages(zone), SWAP_CLUSTER_MAX);
+	}
 
 	/*
-	 * If there is no low memory pressure or the zone is balanced then no
-	 * reclaim is necessary
+	 * Historically care was taken to put equal pressure on all zones but
+	 * now pressure is applied based on node LRU order.
 	 */
-	lowmem_pressure = (buffer_heads_over_limit && is_highmem(zone));
-	if (!lowmem_pressure && zone_balanced(zone, sc->order, false,
-						balance_gap, classzone_idx))
-		return true;
-
-	shrink_node(zone->zone_pgdat, sc, classzone_idx);
-
-	/* TODO: ANOMALY */
-	clear_bit(PGDAT_WRITEBACK, &pgdat->flags);
+	shrink_node(pgdat, sc, classzone_idx);
 
 	/*
-	 * If a zone reaches its high watermark, consider it to be no longer
-	 * congested. It's possible there are dirty pages backed by congested
-	 * BDIs but as pressure is relieved, speculatively avoid congestion
-	 * waits.
+	 * Fragmentation may mean that the system cannot be rebalanced for
+	 * high-order allocations. If twice the allocation size has been
+	 * reclaimed then recheck watermarks only at order-0 to prevent
+	 * excessive reclaim. Assume that a process requested a high-order
+	 * can direct reclaim/compact.
 	 */
-	if (pgdat_reclaimable(zone->zone_pgdat) &&
-	    zone_balanced(zone, sc->order, false, 0, classzone_idx)) {
-		clear_bit(PGDAT_CONGESTED, &pgdat->flags);
-		clear_bit(PGDAT_DIRTY, &pgdat->flags);
-	}
+	if (sc->order && sc->nr_reclaimed >= 2UL << sc->order)
+		sc->order = 0;
 
 	return sc->nr_scanned >= sc->nr_to_reclaim;
 }
 
 /*
- * For kswapd, balance_pgdat() will work across all this node's zones until
- * they are all at high_wmark_pages(zone).
- *
- * Returns the highest zone idx kswapd was reclaiming at
+ * For kswapd, balance_pgdat() will reclaim pages across a node from zones
+ * that are eligible for use by the caller until at least one zone is
+ * balanced.
  *
- * There is special handling here for zones which are full of pinned pages.
- * This can happen if the pages are all mlocked, or if they are all used by
- * device drivers (say, ZONE_DMA).  Or if they are all in use by hugetlb.
- * What we do is to detect the case where all pages in the zone have been
- * scanned twice and there has been zero successful reclaim.  Mark the zone as
- * dead and from now on, only perform a short scan.  Basically we're polling
- * the zone for when the problem goes away.
+ * Returns the order kswapd finished reclaiming at.
  *
  * kswapd scans the zones in the highmem->normal->dma direction.  It skips
  * zones which have free_pages > high_wmark_pages(zone), but once a zone is
- * found to have free_pages <= high_wmark_pages(zone), we scan that zone and the
- * lower zones regardless of the number of free pages in the lower zones. This
- * interoperates with the page allocator fallback scheme to ensure that aging
- * of pages is balanced across the zones.
+ * found to have free_pages <= high_wmark_pages(zone), any page is that zone
+ * or lower is eligible for reclaim until at least one usable zone is
+ * balanced.
  */
 static int balance_pgdat(pg_data_t *pgdat, int order, int classzone_idx)
 {
 	int i;
-	int end_zone = 0;	/* Inclusive.  0 = ZONE_DMA */
 	unsigned long nr_soft_reclaimed;
 	unsigned long nr_soft_scanned;
+	struct zone *zone;
 	struct scan_control sc = {
 		.gfp_mask = GFP_KERNEL,
-		.reclaim_idx = MAX_NR_ZONES - 1,
 		.order = order,
 		.priority = DEF_PRIORITY,
 		.may_writepage = !laptop_mode,
 		.may_unmap = 1,
 		.may_swap = 1,
+		.reclaim_idx = classzone_idx,
 	};
 	count_vm_event(PAGEOUTRUN);
 
@@ -3203,21 +3125,10 @@ static int balance_pgdat(pg_data_t *pgdat, int order, int classzone_idx)
 
 		/* Scan from the highest requested zone to dma */
 		for (i = classzone_idx; i >= 0; i--) {
-			struct zone *zone = pgdat->node_zones + i;
-
+			zone = pgdat->node_zones + i;
 			if (!populated_zone(zone))
 				continue;
 
-			if (sc.priority != DEF_PRIORITY &&
-			    !pgdat_reclaimable(zone->zone_pgdat))
-				continue;
-
-			/*
-			 * Do some background aging of the anon list, to give
-			 * pages a chance to be referenced before reclaiming.
-			 */
-			age_active_anon(zone, &sc);
-
 			/*
 			 * If the number of buffer_heads in the machine
 			 * exceeds the maximum allowed level and this node
@@ -3225,19 +3136,17 @@ static int balance_pgdat(pg_data_t *pgdat, int order, int classzone_idx)
 			 * it to relieve lowmem pressure.
 			 */
 			if (buffer_heads_over_limit && is_highmem_idx(i)) {
-				end_zone = i;
+				classzone_idx = i;
 				break;
 			}
 
-			if (!zone_balanced(zone, order, false, 0, 0)) {
-				end_zone = i;
+			if (!zone_balanced(zone, order, 0, 0)) {
+				classzone_idx = i;
 				break;
 			} else {
 				/*
-				 * If balanced, clear the dirty and congested
-				 * flags
-				 *
-				 * TODO: ANOMALY
+				 * If any eligible zone is balanced then the
+				 * node is not considered congested or dirty.
 				 */
 				clear_bit(PGDAT_CONGESTED, &zone->zone_pgdat->flags);
 				clear_bit(PGDAT_DIRTY, &zone->zone_pgdat->flags);
@@ -3248,51 +3157,34 @@ static int balance_pgdat(pg_data_t *pgdat, int order, int classzone_idx)
 			goto out;
 
 		/*
+		 * Do some background aging of the anon list, to give
+		 * pages a chance to be referenced before reclaiming. All
+		 * pages are rotated regardless of classzone as this is
+		 * about consistent aging.
+		 */
+		age_active_anon(pgdat, &pgdat->node_zones[MAX_NR_ZONES - 1], &sc);
+
+		/*
 		 * If we're getting trouble reclaiming, start doing writepage
 		 * even in laptop mode.
 		 */
-		if (sc.priority < DEF_PRIORITY - 2)
+		if (sc.priority < DEF_PRIORITY - 2 || !pgdat_reclaimable(pgdat))
 			sc.may_writepage = 1;
 
+		/* Call soft limit reclaim before calling shrink_node. */
+		sc.nr_scanned = 0;
+		nr_soft_scanned = 0;
+		nr_soft_reclaimed = mem_cgroup_soft_limit_reclaim(zone, sc.order,
+						sc.gfp_mask, &nr_soft_scanned);
+		sc.nr_reclaimed += nr_soft_reclaimed;
+
 		/*
-		 * Continue scanning in the highmem->dma direction stopping at
-		 * the last zone which needs scanning. This may reclaim lowmem
-		 * pages that are not necessary for zone balancing but it
-		 * preserves LRU ordering. It is assumed that the bulk of
-		 * allocation requests can use arbitrary zones with the
-		 * possible exception of big highmem:lowmem configurations.
+		 * There should be no need to raise the scanning priority if
+		 * enough pages are already being scanned that that high
+		 * watermark would be met at 100% efficiency.
 		 */
-		for (i = end_zone; i >= 0; i--) {
-			struct zone *zone = pgdat->node_zones + i;
-
-			if (!populated_zone(zone))
-				continue;
-
-			if (sc.priority != DEF_PRIORITY &&
-			    !pgdat_reclaimable(zone->zone_pgdat))
-				continue;
-
-			sc.nr_scanned = 0;
-			sc.reclaim_idx = i;
-
-			nr_soft_scanned = 0;
-			/*
-			 * Call soft limit reclaim before calling shrink_zone.
-			 */
-			nr_soft_reclaimed = mem_cgroup_soft_limit_reclaim(zone,
-							order, sc.gfp_mask,
-							&nr_soft_scanned);
-			sc.nr_reclaimed += nr_soft_reclaimed;
-
-			/*
-			 * There should be no need to raise the scanning
-			 * priority if enough pages are already being scanned
-			 * that that high watermark would be met at 100%
-			 * efficiency.
-			 */
-			if (kswapd_shrink_zone(zone, end_zone, &sc))
-				raise_priority = false;
-		}
+		if (kswapd_shrink_node(pgdat, classzone_idx, &sc))
+			raise_priority = false;
 
 		/*
 		 * If the low watermark is met there is no need for processes
@@ -3308,20 +3200,37 @@ static int balance_pgdat(pg_data_t *pgdat, int order, int classzone_idx)
 			break;
 
 		/*
+		 * Stop reclaiming if any eligible zone is balanced and clear
+		 * node writeback or congested.
+		 */
+		for (i = 0; i <= classzone_idx; i++) {
+			zone = pgdat->node_zones + i;
+			if (!populated_zone(zone))
+				continue;
+
+			if (zone_balanced(zone, sc.order, 0, classzone_idx)) {
+				clear_bit(PGDAT_CONGESTED, &pgdat->flags);
+				clear_bit(PGDAT_DIRTY, &pgdat->flags);
+				goto out;
+			}
+		}
+
+		/*
 		 * Raise priority if scanning rate is too low or there was no
 		 * progress in reclaiming pages
 		 */
 		if (raise_priority || !sc.nr_reclaimed)
 			sc.priority--;
-	} while (sc.priority >= 1 &&
-			!pgdat_balanced(pgdat, order, classzone_idx));
+	} while (sc.priority >= 1);
 
 out:
 	/*
-	 * Return the highest zone idx we were reclaiming at so
-	 * prepare_kswapd_sleep() makes the same decisions as here.
+	 * Return the order kswapd stopped reclaiming at as
+	 * prepare_kswapd_sleep() takes it into account. If another caller
+	 * entered the allocator slow path while kswapd was awake, order will
+	 * remain at the higher level.
 	 */
-	return end_zone;
+	return sc.order;
 }
 
 static void kswapd_try_to_sleep(pg_data_t *pgdat, int order,
@@ -3478,8 +3387,9 @@ static int kswapd(void *p)
 		 */
 		if (!ret) {
 			trace_mm_vmscan_kswapd_wake(pgdat->node_id, order);
-			balanced_classzone_idx = balance_pgdat(pgdat, order,
-								classzone_idx);
+
+			/* return value ignored until next patch */
+			balance_pgdat(pgdat, order, classzone_idx);
 		}
 	}
 
@@ -3509,7 +3419,7 @@ void wakeup_kswapd(struct zone *zone, int order, enum zone_type classzone_idx)
 	}
 	if (!waitqueue_active(&pgdat->kswapd_wait))
 		return;
-	if (zone_balanced(zone, order, true, 0, 0))
+	if (zone_balanced(zone, order, 0, 0))
 		return;
 
 	trace_mm_vmscan_wakeup_kswapd(pgdat->node_id, zone_idx(zone), order);
-- 
2.6.4

^ permalink raw reply	[flat|nested] 9+ messages in thread

* Re: [PATCH 06/27] mm, vmscan: Make kswapd reclaim in terms of nodes
  2016-06-09 18:04 ` [PATCH 06/27] mm, vmscan: Make kswapd reclaim in terms of nodes Mel Gorman
@ 2016-06-15 14:23   ` Vlastimil Babka
  0 siblings, 0 replies; 9+ messages in thread
From: Vlastimil Babka @ 2016-06-15 14:23 UTC (permalink / raw)
  To: Mel Gorman, Andrew Morton, Linux-MM; +Cc: Rik van Riel, Johannes Weiner, LKML

On 06/09/2016 08:04 PM, Mel Gorman wrote:
> Patch "mm: vmscan: Begin reclaiming pages on a per-node basis" started
> thinking of reclaim in terms of nodes but kswapd is still zone-centric. This
> patch gets rid of many of the node-based versus zone-based decisions.
>
> o A node is considered balanced when any eligible lower zone is balanced.
>   This eliminates one class of age-inversion problem because we avoid
>   reclaiming a newer page just because it's in the wrong zone
> o pgdat_balanced disappears because we now only care about one zone being
>   balanced.
> o Some anomalies related to writeback and congestion tracking being based on
>   zones disappear.
> o kswapd no longer has to take care to reclaim zones in the reverse order
>   that the page allocator uses.
> o Most importantly of all, reclaim from node 0 with multiple zones will
>   have similar aging and reclaiming characteristics as every
>   other node.
>
> Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
> Acked-by: Johannes Weiner <hannes@cmpxchg.org>

Acked-by: Vlastimil Babka <vbabka@suse.cz>

^ permalink raw reply	[flat|nested] 9+ messages in thread

* [PATCH 06/27] mm, vmscan: Make kswapd reclaim in terms of nodes
  2016-06-09 18:04 [PATCH 00/27] Move LRU page reclaim from zones to nodes v6 Mel Gorman
@ 2016-06-09 18:04 ` Mel Gorman
  2016-06-15 14:23   ` Vlastimil Babka
  0 siblings, 1 reply; 9+ messages in thread
From: Mel Gorman @ 2016-06-09 18:04 UTC (permalink / raw)
  To: Andrew Morton, Linux-MM
  Cc: Rik van Riel, Vlastimil Babka, Johannes Weiner, LKML, Mel Gorman

Patch "mm: vmscan: Begin reclaiming pages on a per-node basis" started
thinking of reclaim in terms of nodes but kswapd is still zone-centric. This
patch gets rid of many of the node-based versus zone-based decisions.

o A node is considered balanced when any eligible lower zone is balanced.
  This eliminates one class of age-inversion problem because we avoid
  reclaiming a newer page just because it's in the wrong zone
o pgdat_balanced disappears because we now only care about one zone being
  balanced.
o Some anomalies related to writeback and congestion tracking being based on
  zones disappear.
o kswapd no longer has to take care to reclaim zones in the reverse order
  that the page allocator uses.
o Most importantly of all, reclaim from node 0 with multiple zones will
  have similar aging and reclaiming characteristics as every
  other node.

Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
Acked-by: Johannes Weiner <hannes@cmpxchg.org>
---
 mm/vmscan.c | 292 +++++++++++++++++++++---------------------------------------
 1 file changed, 101 insertions(+), 191 deletions(-)

diff --git a/mm/vmscan.c b/mm/vmscan.c
index 0a619241c576..9368af4cfb06 100644
--- a/mm/vmscan.c
+++ b/mm/vmscan.c
@@ -2942,7 +2942,8 @@ unsigned long try_to_free_mem_cgroup_pages(struct mem_cgroup *memcg,
 }
 #endif
 
-static void age_active_anon(struct zone *zone, struct scan_control *sc)
+static void age_active_anon(struct pglist_data *pgdat,
+				struct zone *zone, struct scan_control *sc)
 {
 	struct mem_cgroup *memcg;
 
@@ -2961,85 +2962,15 @@ static void age_active_anon(struct zone *zone, struct scan_control *sc)
 	} while (memcg);
 }
 
-static bool zone_balanced(struct zone *zone, int order, bool highorder,
+static bool zone_balanced(struct zone *zone, int order,
 			unsigned long balance_gap, int classzone_idx)
 {
 	unsigned long mark = high_wmark_pages(zone) + balance_gap;
 
-	/*
-	 * When checking from pgdat_balanced(), kswapd should stop and sleep
-	 * when it reaches the high order-0 watermark and let kcompactd take
-	 * over. Other callers such as wakeup_kswapd() want to determine the
-	 * true high-order watermark.
-	 */
-	if (IS_ENABLED(CONFIG_COMPACTION) && !highorder) {
-		mark += (1UL << order);
-		order = 0;
-	}
-
 	return zone_watermark_ok_safe(zone, order, mark, classzone_idx);
 }
 
 /*
- * pgdat_balanced() is used when checking if a node is balanced.
- *
- * For order-0, all zones must be balanced!
- *
- * For high-order allocations only zones that meet watermarks and are in a
- * zone allowed by the callers classzone_idx are added to balanced_pages. The
- * total of balanced pages must be at least 25% of the zones allowed by
- * classzone_idx for the node to be considered balanced. Forcing all zones to
- * be balanced for high orders can cause excessive reclaim when there are
- * imbalanced zones.
- * The choice of 25% is due to
- *   o a 16M DMA zone that is balanced will not balance a zone on any
- *     reasonable sized machine
- *   o On all other machines, the top zone must be at least a reasonable
- *     percentage of the middle zones. For example, on 32-bit x86, highmem
- *     would need to be at least 256M for it to be balance a whole node.
- *     Similarly, on x86-64 the Normal zone would need to be at least 1G
- *     to balance a node on its own. These seemed like reasonable ratios.
- */
-static bool pgdat_balanced(pg_data_t *pgdat, int order, int classzone_idx)
-{
-	unsigned long managed_pages = 0;
-	unsigned long balanced_pages = 0;
-	int i;
-
-	/* Check the watermark levels */
-	for (i = 0; i <= classzone_idx; i++) {
-		struct zone *zone = pgdat->node_zones + i;
-
-		if (!populated_zone(zone))
-			continue;
-
-		managed_pages += zone->managed_pages;
-
-		/*
-		 * A special case here:
-		 *
-		 * balance_pgdat() skips over all_unreclaimable after
-		 * DEF_PRIORITY. Effectively, it considers them balanced so
-		 * they must be considered balanced here as well!
-		 */
-		if (!pgdat_reclaimable(zone->zone_pgdat)) {
-			balanced_pages += zone->managed_pages;
-			continue;
-		}
-
-		if (zone_balanced(zone, order, false, 0, i))
-			balanced_pages += zone->managed_pages;
-		else if (!order)
-			return false;
-	}
-
-	if (order)
-		return balanced_pages >= (managed_pages >> 2);
-	else
-		return true;
-}
-
-/*
  * Prepare kswapd for sleeping. This verifies that there are no processes
  * waiting in throttle_direct_reclaim() and that watermarks have been met.
  *
@@ -3048,6 +2979,8 @@ static bool pgdat_balanced(pg_data_t *pgdat, int order, int classzone_idx)
 static bool prepare_kswapd_sleep(pg_data_t *pgdat, int order, long remaining,
 					int classzone_idx)
 {
+	int i;
+
 	/* If a direct reclaimer woke kswapd within HZ/10, it's premature */
 	if (remaining)
 		return false;
@@ -3068,101 +3001,90 @@ static bool prepare_kswapd_sleep(pg_data_t *pgdat, int order, long remaining,
 	if (waitqueue_active(&pgdat->pfmemalloc_wait))
 		wake_up_all(&pgdat->pfmemalloc_wait);
 
-	return pgdat_balanced(pgdat, order, classzone_idx);
+	for (i = 0; i <= classzone_idx; i++) {
+		struct zone *zone = pgdat->node_zones + i;
+
+		if (!populated_zone(zone))
+			continue;
+
+		if (zone_balanced(zone, order, 0, classzone_idx))
+			return true;
+	}
+
+	return false;
 }
 
 /*
- * kswapd shrinks the zone by the number of pages required to reach
- * the high watermark.
+ * kswapd shrinks a node of pages that are at or below the highest usable
+ * zone that is currently unbalanced.
  *
  * Returns true if kswapd scanned at least the requested number of pages to
  * reclaim or if the lack of progress was due to pages under writeback.
  * This is used to determine if the scanning priority needs to be raised.
  */
-static bool kswapd_shrink_zone(struct zone *zone,
+static bool kswapd_shrink_node(pg_data_t *pgdat,
 			       int classzone_idx,
 			       struct scan_control *sc)
 {
-	unsigned long balance_gap;
-	bool lowmem_pressure;
-	struct pglist_data *pgdat = zone->zone_pgdat;
+	struct zone *zone;
+	unsigned long nr_to_reclaim = 0;
+	int z;
 
-	/* Reclaim above the high watermark. */
-	sc->nr_to_reclaim = max(SWAP_CLUSTER_MAX, high_wmark_pages(zone));
+	/* Reclaim a number of pages proportional to the number of zones */
+	for (z = 0; z <= classzone_idx; z++) {
+		zone = pgdat->node_zones + z;
+		if (!populated_zone(zone))
+			continue;
 
-	/*
-	 * We put equal pressure on every zone, unless one zone has way too
-	 * many pages free already. The "too many pages" is defined as the
-	 * high wmark plus a "gap" where the gap is either the low
-	 * watermark or 1% of the zone, whichever is smaller.
-	 */
-	balance_gap = min(low_wmark_pages(zone), DIV_ROUND_UP(
-			zone->managed_pages, KSWAPD_ZONE_BALANCE_GAP_RATIO));
+		nr_to_reclaim += max(high_wmark_pages(zone), SWAP_CLUSTER_MAX);
+	}
 
 	/*
-	 * If there is no low memory pressure or the zone is balanced then no
-	 * reclaim is necessary
+	 * Historically care was taken to put equal pressure on all zones but
+	 * now pressure is applied based on node LRU order.
 	 */
-	lowmem_pressure = (buffer_heads_over_limit && is_highmem(zone));
-	if (!lowmem_pressure && zone_balanced(zone, sc->order, false,
-						balance_gap, classzone_idx))
-		return true;
-
-	shrink_node(zone->zone_pgdat, sc, classzone_idx);
-
-	/* TODO: ANOMALY */
-	clear_bit(PGDAT_WRITEBACK, &pgdat->flags);
+	shrink_node(pgdat, sc, classzone_idx);
 
 	/*
-	 * If a zone reaches its high watermark, consider it to be no longer
-	 * congested. It's possible there are dirty pages backed by congested
-	 * BDIs but as pressure is relieved, speculatively avoid congestion
-	 * waits.
+	 * Fragmentation may mean that the system cannot be rebalanced for
+	 * high-order allocations. If twice the allocation size has been
+	 * reclaimed then recheck watermarks only at order-0 to prevent
+	 * excessive reclaim. Assume that a process requested a high-order
+	 * can direct reclaim/compact.
 	 */
-	if (pgdat_reclaimable(zone->zone_pgdat) &&
-	    zone_balanced(zone, sc->order, false, 0, classzone_idx)) {
-		clear_bit(PGDAT_CONGESTED, &pgdat->flags);
-		clear_bit(PGDAT_DIRTY, &pgdat->flags);
-	}
+	if (sc->order && sc->nr_reclaimed >= 2UL << sc->order)
+		sc->order = 0;
 
 	return sc->nr_scanned >= sc->nr_to_reclaim;
 }
 
 /*
- * For kswapd, balance_pgdat() will work across all this node's zones until
- * they are all at high_wmark_pages(zone).
- *
- * Returns the highest zone idx kswapd was reclaiming at
+ * For kswapd, balance_pgdat() will reclaim pages across a node from zones
+ * that are eligible for use by the caller until at least one zone is
+ * balanced.
  *
- * There is special handling here for zones which are full of pinned pages.
- * This can happen if the pages are all mlocked, or if they are all used by
- * device drivers (say, ZONE_DMA).  Or if they are all in use by hugetlb.
- * What we do is to detect the case where all pages in the zone have been
- * scanned twice and there has been zero successful reclaim.  Mark the zone as
- * dead and from now on, only perform a short scan.  Basically we're polling
- * the zone for when the problem goes away.
+ * Returns the order kswapd finished reclaiming at.
  *
  * kswapd scans the zones in the highmem->normal->dma direction.  It skips
  * zones which have free_pages > high_wmark_pages(zone), but once a zone is
- * found to have free_pages <= high_wmark_pages(zone), we scan that zone and the
- * lower zones regardless of the number of free pages in the lower zones. This
- * interoperates with the page allocator fallback scheme to ensure that aging
- * of pages is balanced across the zones.
+ * found to have free_pages <= high_wmark_pages(zone), any page is that zone
+ * or lower is eligible for reclaim until at least one usable zone is
+ * balanced.
  */
 static int balance_pgdat(pg_data_t *pgdat, int order, int classzone_idx)
 {
 	int i;
-	int end_zone = 0;	/* Inclusive.  0 = ZONE_DMA */
 	unsigned long nr_soft_reclaimed;
 	unsigned long nr_soft_scanned;
+	struct zone *zone;
 	struct scan_control sc = {
 		.gfp_mask = GFP_KERNEL,
-		.reclaim_idx = MAX_NR_ZONES - 1,
 		.order = order,
 		.priority = DEF_PRIORITY,
 		.may_writepage = !laptop_mode,
 		.may_unmap = 1,
 		.may_swap = 1,
+		.reclaim_idx = classzone_idx,
 	};
 	count_vm_event(PAGEOUTRUN);
 
@@ -3173,21 +3095,10 @@ static int balance_pgdat(pg_data_t *pgdat, int order, int classzone_idx)
 
 		/* Scan from the highest requested zone to dma */
 		for (i = classzone_idx; i >= 0; i--) {
-			struct zone *zone = pgdat->node_zones + i;
-
+			zone = pgdat->node_zones + i;
 			if (!populated_zone(zone))
 				continue;
 
-			if (sc.priority != DEF_PRIORITY &&
-			    !pgdat_reclaimable(zone->zone_pgdat))
-				continue;
-
-			/*
-			 * Do some background aging of the anon list, to give
-			 * pages a chance to be referenced before reclaiming.
-			 */
-			age_active_anon(zone, &sc);
-
 			/*
 			 * If the number of buffer_heads in the machine
 			 * exceeds the maximum allowed level and this node
@@ -3195,19 +3106,17 @@ static int balance_pgdat(pg_data_t *pgdat, int order, int classzone_idx)
 			 * it to relieve lowmem pressure.
 			 */
 			if (buffer_heads_over_limit && is_highmem_idx(i)) {
-				end_zone = i;
+				classzone_idx = i;
 				break;
 			}
 
-			if (!zone_balanced(zone, order, false, 0, 0)) {
-				end_zone = i;
+			if (!zone_balanced(zone, order, 0, 0)) {
+				classzone_idx = i;
 				break;
 			} else {
 				/*
-				 * If balanced, clear the dirty and congested
-				 * flags
-				 *
-				 * TODO: ANOMALY
+				 * If any eligible zone is balanced then the
+				 * node is not considered congested or dirty.
 				 */
 				clear_bit(PGDAT_CONGESTED, &zone->zone_pgdat->flags);
 				clear_bit(PGDAT_DIRTY, &zone->zone_pgdat->flags);
@@ -3218,51 +3127,34 @@ static int balance_pgdat(pg_data_t *pgdat, int order, int classzone_idx)
 			goto out;
 
 		/*
+		 * Do some background aging of the anon list, to give
+		 * pages a chance to be referenced before reclaiming. All
+		 * pages are rotated regardless of classzone as this is
+		 * about consistent aging.
+		 */
+		age_active_anon(pgdat, &pgdat->node_zones[MAX_NR_ZONES - 1], &sc);
+
+		/*
 		 * If we're getting trouble reclaiming, start doing writepage
 		 * even in laptop mode.
 		 */
-		if (sc.priority < DEF_PRIORITY - 2)
+		if (sc.priority < DEF_PRIORITY - 2 || !pgdat_reclaimable(pgdat))
 			sc.may_writepage = 1;
 
+		/* Call soft limit reclaim before calling shrink_node. */
+		sc.nr_scanned = 0;
+		nr_soft_scanned = 0;
+		nr_soft_reclaimed = mem_cgroup_soft_limit_reclaim(zone, sc.order,
+						sc.gfp_mask, &nr_soft_scanned);
+		sc.nr_reclaimed += nr_soft_reclaimed;
+
 		/*
-		 * Continue scanning in the highmem->dma direction stopping at
-		 * the last zone which needs scanning. This may reclaim lowmem
-		 * pages that are not necessary for zone balancing but it
-		 * preserves LRU ordering. It is assumed that the bulk of
-		 * allocation requests can use arbitrary zones with the
-		 * possible exception of big highmem:lowmem configurations.
+		 * There should be no need to raise the scanning priority if
+		 * enough pages are already being scanned that that high
+		 * watermark would be met at 100% efficiency.
 		 */
-		for (i = end_zone; i >= end_zone; i--) {
-			struct zone *zone = pgdat->node_zones + i;
-
-			if (!populated_zone(zone))
-				continue;
-
-			if (sc.priority != DEF_PRIORITY &&
-			    !pgdat_reclaimable(zone->zone_pgdat))
-				continue;
-
-			sc.nr_scanned = 0;
-			sc.reclaim_idx = i;
-
-			nr_soft_scanned = 0;
-			/*
-			 * Call soft limit reclaim before calling shrink_zone.
-			 */
-			nr_soft_reclaimed = mem_cgroup_soft_limit_reclaim(zone,
-							order, sc.gfp_mask,
-							&nr_soft_scanned);
-			sc.nr_reclaimed += nr_soft_reclaimed;
-
-			/*
-			 * There should be no need to raise the scanning
-			 * priority if enough pages are already being scanned
-			 * that that high watermark would be met at 100%
-			 * efficiency.
-			 */
-			if (kswapd_shrink_zone(zone, end_zone, &sc))
-				raise_priority = false;
-		}
+		if (kswapd_shrink_node(pgdat, classzone_idx, &sc))
+			raise_priority = false;
 
 		/*
 		 * If the low watermark is met there is no need for processes
@@ -3278,20 +3170,37 @@ static int balance_pgdat(pg_data_t *pgdat, int order, int classzone_idx)
 			break;
 
 		/*
+		 * Stop reclaiming if any eligible zone is balanced and clear
+		 * node writeback or congested.
+		 */
+		for (i = 0; i <= classzone_idx; i++) {
+			zone = pgdat->node_zones + i;
+			if (!populated_zone(zone))
+				continue;
+
+			if (zone_balanced(zone, sc.order, 0, classzone_idx)) {
+				clear_bit(PGDAT_CONGESTED, &pgdat->flags);
+				clear_bit(PGDAT_DIRTY, &pgdat->flags);
+				goto out;
+			}
+		}
+
+		/*
 		 * Raise priority if scanning rate is too low or there was no
 		 * progress in reclaiming pages
 		 */
 		if (raise_priority || !sc.nr_reclaimed)
 			sc.priority--;
-	} while (sc.priority >= 1 &&
-			!pgdat_balanced(pgdat, order, classzone_idx));
+	} while (sc.priority >= 1);
 
 out:
 	/*
-	 * Return the highest zone idx we were reclaiming at so
-	 * prepare_kswapd_sleep() makes the same decisions as here.
+	 * Return the order kswapd stopped reclaiming at as
+	 * prepare_kswapd_sleep() takes it into account. If another caller
+	 * entered the allocator slow path while kswapd was awake, order will
+	 * remain at the higher level.
 	 */
-	return end_zone;
+	return sc.order;
 }
 
 static void kswapd_try_to_sleep(pg_data_t *pgdat, int order,
@@ -3448,8 +3357,9 @@ static int kswapd(void *p)
 		 */
 		if (!ret) {
 			trace_mm_vmscan_kswapd_wake(pgdat->node_id, order);
-			balanced_classzone_idx = balance_pgdat(pgdat, order,
-								classzone_idx);
+
+			/* return value ignored until next patch */
+			balance_pgdat(pgdat, order, classzone_idx);
 		}
 	}
 
@@ -3479,7 +3389,7 @@ void wakeup_kswapd(struct zone *zone, int order, enum zone_type classzone_idx)
 	}
 	if (!waitqueue_active(&pgdat->kswapd_wait))
 		return;
-	if (zone_balanced(zone, order, true, 0, 0))
+	if (zone_balanced(zone, order, 0, 0))
 		return;
 
 	trace_mm_vmscan_wakeup_kswapd(pgdat->node_id, zone_idx(zone), order);
-- 
2.6.4

^ permalink raw reply	[flat|nested] 9+ messages in thread

* Re: [PATCH 06/27] mm, vmscan: Make kswapd reclaim in terms of nodes
  2016-04-15  9:13 ` [PATCH 06/27] mm, vmscan: Make kswapd reclaim in terms of nodes Mel Gorman
@ 2016-04-28  8:36   ` Vlastimil Babka
  0 siblings, 0 replies; 9+ messages in thread
From: Vlastimil Babka @ 2016-04-28  8:36 UTC (permalink / raw)
  To: Mel Gorman, Andrew Morton, Linux-MM
  Cc: Rik van Riel, Johannes Weiner, Jesper Dangaard Brouer, LKML

On 04/15/2016 11:13 AM, Mel Gorman wrote:
>   	/*
> -	 * If a zone reaches its high watermark, consider it to be no longer
> -	 * congested. It's possible there are dirty pages backed by congested
> -	 * BDIs but as pressure is relieved, speculatively avoid congestion
> -	 * waits.
> +	 * Fragmentation may mean that the system cannot be rebalanced for
> +	 * high-order allocations. If twice the allocation size has been
> +	 * reclaimed then recheck watermarks only at order-0 to prevent
> +	 * excessive reclaim. Assume that a process requested a high-order
> +	 * can direct reclaim/compact.

Also kcompactd is woken up in this case...

>   	 */
> -	if (pgdat_reclaimable(zone->zone_pgdat) &&
> -	    zone_balanced(zone, sc->order, false, 0, classzone_idx)) {
> -		clear_bit(PGDAT_CONGESTED, &pgdat->flags);
> -		clear_bit(PGDAT_DIRTY, &pgdat->flags);
> -	}
> +	if (sc->order && sc->nr_reclaimed >= 2UL << sc->order)
> +		sc->order = 0;
>
>   	return sc->nr_scanned >= sc->nr_to_reclaim;

This looks indeed simpler than my earlier zone_balanced() modification 
you removed. However I think there's still potential of overreclaim due 
to a stream of kswapd_wakeups where each will have to reclaim 2UL << 
sc->order pages, regardless of watermarks. Could be some high-order 
wakeups from GFP_ATOMIC context that have order-0 fallbacks but will 
cause kswapd to keep reclaiming when kcompactd can't keep up due to 
fragmentation...

^ permalink raw reply	[flat|nested] 9+ messages in thread

* [PATCH 06/27] mm, vmscan: Make kswapd reclaim in terms of nodes
  2016-04-15  9:13 [PATCH 00/27] Move LRU page reclaim from zones to nodes v5 Mel Gorman
@ 2016-04-15  9:13 ` Mel Gorman
  2016-04-28  8:36   ` Vlastimil Babka
  0 siblings, 1 reply; 9+ messages in thread
From: Mel Gorman @ 2016-04-15  9:13 UTC (permalink / raw)
  To: Andrew Morton, Linux-MM
  Cc: Rik van Riel, Vlastimil Babka, Johannes Weiner,
	Jesper Dangaard Brouer, LKML, Mel Gorman

Patch "mm: vmscan: Begin reclaiming pages on a per-node basis" started
thinking of reclaim in terms of nodes but kswapd is still zone-centric. This
patch gets rid of many of the node-based versus zone-based decisions.

o A node is considered balanced when any eligible lower zone is balanced.
  This eliminates one class of age-inversion problem because we avoid
  reclaiming a newer page just because it's in the wrong zone
o pgdat_balanced disappears because we now only care about one zone being
  balanced.
o Some anomalies related to writeback and congestion tracking being based on
  zones disappear.
o kswapd no longer has to take care to reclaim zones in the reverse order
  that the page allocator uses.
o Most importantly of all, reclaim from node 0 with multiple zones will
  have similar aging and reclaiming characteristics as every
  other node.

Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
Acked-by: Johannes Weiner <hannes@cmpxchg.org>
---
 mm/vmscan.c | 292 +++++++++++++++++++++---------------------------------------
 1 file changed, 101 insertions(+), 191 deletions(-)

diff --git a/mm/vmscan.c b/mm/vmscan.c
index f2534e8f8527..c23d8f9722ad 100644
--- a/mm/vmscan.c
+++ b/mm/vmscan.c
@@ -2979,7 +2979,8 @@ unsigned long try_to_free_mem_cgroup_pages(struct mem_cgroup *memcg,
 }
 #endif
 
-static void age_active_anon(struct zone *zone, struct scan_control *sc)
+static void age_active_anon(struct pglist_data *pgdat,
+				struct zone *zone, struct scan_control *sc)
 {
 	struct mem_cgroup *memcg;
 
@@ -2998,85 +2999,15 @@ static void age_active_anon(struct zone *zone, struct scan_control *sc)
 	} while (memcg);
 }
 
-static bool zone_balanced(struct zone *zone, int order, bool highorder,
+static bool zone_balanced(struct zone *zone, int order,
 			unsigned long balance_gap, int classzone_idx)
 {
 	unsigned long mark = high_wmark_pages(zone) + balance_gap;
 
-	/*
-	 * When checking from pgdat_balanced(), kswapd should stop and sleep
-	 * when it reaches the high order-0 watermark and let kcompactd take
-	 * over. Other callers such as wakeup_kswapd() want to determine the
-	 * true high-order watermark.
-	 */
-	if (IS_ENABLED(CONFIG_COMPACTION) && !highorder) {
-		mark += (1UL << order);
-		order = 0;
-	}
-
 	return zone_watermark_ok_safe(zone, order, mark, classzone_idx);
 }
 
 /*
- * pgdat_balanced() is used when checking if a node is balanced.
- *
- * For order-0, all zones must be balanced!
- *
- * For high-order allocations only zones that meet watermarks and are in a
- * zone allowed by the callers classzone_idx are added to balanced_pages. The
- * total of balanced pages must be at least 25% of the zones allowed by
- * classzone_idx for the node to be considered balanced. Forcing all zones to
- * be balanced for high orders can cause excessive reclaim when there are
- * imbalanced zones.
- * The choice of 25% is due to
- *   o a 16M DMA zone that is balanced will not balance a zone on any
- *     reasonable sized machine
- *   o On all other machines, the top zone must be at least a reasonable
- *     percentage of the middle zones. For example, on 32-bit x86, highmem
- *     would need to be at least 256M for it to be balance a whole node.
- *     Similarly, on x86-64 the Normal zone would need to be at least 1G
- *     to balance a node on its own. These seemed like reasonable ratios.
- */
-static bool pgdat_balanced(pg_data_t *pgdat, int order, int classzone_idx)
-{
-	unsigned long managed_pages = 0;
-	unsigned long balanced_pages = 0;
-	int i;
-
-	/* Check the watermark levels */
-	for (i = 0; i <= classzone_idx; i++) {
-		struct zone *zone = pgdat->node_zones + i;
-
-		if (!populated_zone(zone))
-			continue;
-
-		managed_pages += zone->managed_pages;
-
-		/*
-		 * A special case here:
-		 *
-		 * balance_pgdat() skips over all_unreclaimable after
-		 * DEF_PRIORITY. Effectively, it considers them balanced so
-		 * they must be considered balanced here as well!
-		 */
-		if (!pgdat_reclaimable(zone->zone_pgdat)) {
-			balanced_pages += zone->managed_pages;
-			continue;
-		}
-
-		if (zone_balanced(zone, order, false, 0, i))
-			balanced_pages += zone->managed_pages;
-		else if (!order)
-			return false;
-	}
-
-	if (order)
-		return balanced_pages >= (managed_pages >> 2);
-	else
-		return true;
-}
-
-/*
  * Prepare kswapd for sleeping. This verifies that there are no processes
  * waiting in throttle_direct_reclaim() and that watermarks have been met.
  *
@@ -3085,6 +3016,8 @@ static bool pgdat_balanced(pg_data_t *pgdat, int order, int classzone_idx)
 static bool prepare_kswapd_sleep(pg_data_t *pgdat, int order, long remaining,
 					int classzone_idx)
 {
+	int i;
+
 	/* If a direct reclaimer woke kswapd within HZ/10, it's premature */
 	if (remaining)
 		return false;
@@ -3105,101 +3038,90 @@ static bool prepare_kswapd_sleep(pg_data_t *pgdat, int order, long remaining,
 	if (waitqueue_active(&pgdat->pfmemalloc_wait))
 		wake_up_all(&pgdat->pfmemalloc_wait);
 
-	return pgdat_balanced(pgdat, order, classzone_idx);
+	for (i = 0; i <= classzone_idx; i++) {
+		struct zone *zone = pgdat->node_zones + i;
+
+		if (!populated_zone(zone))
+			continue;
+
+		if (zone_balanced(zone, order, 0, classzone_idx))
+			return true;
+	}
+
+	return false;
 }
 
 /*
- * kswapd shrinks the zone by the number of pages required to reach
- * the high watermark.
+ * kswapd shrinks a node of pages that are at or below the highest usable
+ * zone that is currently unbalanced.
  *
  * Returns true if kswapd scanned at least the requested number of pages to
  * reclaim or if the lack of progress was due to pages under writeback.
  * This is used to determine if the scanning priority needs to be raised.
  */
-static bool kswapd_shrink_zone(struct zone *zone,
+static bool kswapd_shrink_node(pg_data_t *pgdat,
 			       int classzone_idx,
 			       struct scan_control *sc)
 {
-	unsigned long balance_gap;
-	bool lowmem_pressure;
-	struct pglist_data *pgdat = zone->zone_pgdat;
+	struct zone *zone;
+	unsigned long nr_to_reclaim = 0;
+	int z;
 
-	/* Reclaim above the high watermark. */
-	sc->nr_to_reclaim = max(SWAP_CLUSTER_MAX, high_wmark_pages(zone));
+	/* Reclaim a number of pages proportional to the number of zones */
+	for (z = 0; z <= classzone_idx; z++) {
+		zone = pgdat->node_zones + z;
+		if (!populated_zone(zone))
+			continue;
 
-	/*
-	 * We put equal pressure on every zone, unless one zone has way too
-	 * many pages free already. The "too many pages" is defined as the
-	 * high wmark plus a "gap" where the gap is either the low
-	 * watermark or 1% of the zone, whichever is smaller.
-	 */
-	balance_gap = min(low_wmark_pages(zone), DIV_ROUND_UP(
-			zone->managed_pages, KSWAPD_ZONE_BALANCE_GAP_RATIO));
+		nr_to_reclaim += max(high_wmark_pages(zone), SWAP_CLUSTER_MAX);
+	}
 
 	/*
-	 * If there is no low memory pressure or the zone is balanced then no
-	 * reclaim is necessary
+	 * Historically care was taken to put equal pressure on all zones but
+	 * now pressure is applied based on node LRU order.
 	 */
-	lowmem_pressure = (buffer_heads_over_limit && is_highmem(zone));
-	if (!lowmem_pressure && zone_balanced(zone, sc->order, false,
-						balance_gap, classzone_idx))
-		return true;
-
-	shrink_node(zone->zone_pgdat, sc, classzone_idx);
-
-	/* TODO: ANOMALY */
-	clear_bit(PGDAT_WRITEBACK, &pgdat->flags);
+	shrink_node(pgdat, sc, classzone_idx);
 
 	/*
-	 * If a zone reaches its high watermark, consider it to be no longer
-	 * congested. It's possible there are dirty pages backed by congested
-	 * BDIs but as pressure is relieved, speculatively avoid congestion
-	 * waits.
+	 * Fragmentation may mean that the system cannot be rebalanced for
+	 * high-order allocations. If twice the allocation size has been
+	 * reclaimed then recheck watermarks only at order-0 to prevent
+	 * excessive reclaim. Assume that a process requested a high-order
+	 * can direct reclaim/compact.
 	 */
-	if (pgdat_reclaimable(zone->zone_pgdat) &&
-	    zone_balanced(zone, sc->order, false, 0, classzone_idx)) {
-		clear_bit(PGDAT_CONGESTED, &pgdat->flags);
-		clear_bit(PGDAT_DIRTY, &pgdat->flags);
-	}
+	if (sc->order && sc->nr_reclaimed >= 2UL << sc->order)
+		sc->order = 0;
 
 	return sc->nr_scanned >= sc->nr_to_reclaim;
 }
 
 /*
- * For kswapd, balance_pgdat() will work across all this node's zones until
- * they are all at high_wmark_pages(zone).
- *
- * Returns the highest zone idx kswapd was reclaiming at
+ * For kswapd, balance_pgdat() will reclaim pages across a node from zones
+ * that are eligible for use by the caller until at least one zone is
+ * balanced.
  *
- * There is special handling here for zones which are full of pinned pages.
- * This can happen if the pages are all mlocked, or if they are all used by
- * device drivers (say, ZONE_DMA).  Or if they are all in use by hugetlb.
- * What we do is to detect the case where all pages in the zone have been
- * scanned twice and there has been zero successful reclaim.  Mark the zone as
- * dead and from now on, only perform a short scan.  Basically we're polling
- * the zone for when the problem goes away.
+ * Returns the order kswapd finished reclaiming at.
  *
  * kswapd scans the zones in the highmem->normal->dma direction.  It skips
  * zones which have free_pages > high_wmark_pages(zone), but once a zone is
- * found to have free_pages <= high_wmark_pages(zone), we scan that zone and the
- * lower zones regardless of the number of free pages in the lower zones. This
- * interoperates with the page allocator fallback scheme to ensure that aging
- * of pages is balanced across the zones.
+ * found to have free_pages <= high_wmark_pages(zone), any page is that zone
+ * or lower is eligible for reclaim until at least one usable zone is
+ * balanced.
  */
 static int balance_pgdat(pg_data_t *pgdat, int order, int classzone_idx)
 {
 	int i;
-	int end_zone = 0;	/* Inclusive.  0 = ZONE_DMA */
 	unsigned long nr_soft_reclaimed;
 	unsigned long nr_soft_scanned;
+	struct zone *zone;
 	struct scan_control sc = {
 		.gfp_mask = GFP_KERNEL,
-		.reclaim_idx = MAX_NR_ZONES - 1,
 		.order = order,
 		.priority = DEF_PRIORITY,
 		.may_writepage = !laptop_mode,
 		.may_unmap = 1,
 		.may_swap = 1,
+		.reclaim_idx = classzone_idx,
 	};
 	count_vm_event(PAGEOUTRUN);
 
@@ -3210,21 +3132,10 @@ static int balance_pgdat(pg_data_t *pgdat, int order, int classzone_idx)
 
 		/* Scan from the highest requested zone to dma */
 		for (i = classzone_idx; i >= 0; i--) {
-			struct zone *zone = pgdat->node_zones + i;
-
+			zone = pgdat->node_zones + i;
 			if (!populated_zone(zone))
 				continue;
 
-			if (sc.priority != DEF_PRIORITY &&
-			    !pgdat_reclaimable(zone->zone_pgdat))
-				continue;
-
-			/*
-			 * Do some background aging of the anon list, to give
-			 * pages a chance to be referenced before reclaiming.
-			 */
-			age_active_anon(zone, &sc);
-
 			/*
 			 * If the number of buffer_heads in the machine
 			 * exceeds the maximum allowed level and this node
@@ -3232,19 +3143,17 @@ static int balance_pgdat(pg_data_t *pgdat, int order, int classzone_idx)
 			 * it to relieve lowmem pressure.
 			 */
 			if (buffer_heads_over_limit && is_highmem_idx(i)) {
-				end_zone = i;
+				classzone_idx = i;
 				break;
 			}
 
-			if (!zone_balanced(zone, order, false, 0, 0)) {
-				end_zone = i;
+			if (!zone_balanced(zone, order, 0, 0)) {
+				classzone_idx = i;
 				break;
 			} else {
 				/*
-				 * If balanced, clear the dirty and congested
-				 * flags
-				 *
-				 * TODO: ANOMALY
+				 * If any eligible zone is balanced then the
+				 * node is not considered congested or dirty.
 				 */
 				clear_bit(PGDAT_CONGESTED, &zone->zone_pgdat->flags);
 				clear_bit(PGDAT_DIRTY, &zone->zone_pgdat->flags);
@@ -3255,51 +3164,34 @@ static int balance_pgdat(pg_data_t *pgdat, int order, int classzone_idx)
 			goto out;
 
 		/*
+		 * Do some background aging of the anon list, to give
+		 * pages a chance to be referenced before reclaiming. All
+		 * pages are rotated regardless of classzone as this is
+		 * about consistent aging.
+		 */
+		age_active_anon(pgdat, &pgdat->node_zones[MAX_NR_ZONES - 1], &sc);
+
+		/*
 		 * If we're getting trouble reclaiming, start doing writepage
 		 * even in laptop mode.
 		 */
-		if (sc.priority < DEF_PRIORITY - 2)
+		if (sc.priority < DEF_PRIORITY - 2 || !pgdat_reclaimable(pgdat))
 			sc.may_writepage = 1;
 
+		/* Call soft limit reclaim before calling shrink_node. */
+		sc.nr_scanned = 0;
+		nr_soft_scanned = 0;
+		nr_soft_reclaimed = mem_cgroup_soft_limit_reclaim(zone, sc.order,
+						sc.gfp_mask, &nr_soft_scanned);
+		sc.nr_reclaimed += nr_soft_reclaimed;
+
 		/*
-		 * Continue scanning in the highmem->dma direction stopping at
-		 * the last zone which needs scanning. This may reclaim lowmem
-		 * pages that are not necessary for zone balancing but it
-		 * preserves LRU ordering. It is assumed that the bulk of
-		 * allocation requests can use arbitrary zones with the
-		 * possible exception of big highmem:lowmem configurations.
+		 * There should be no need to raise the scanning priority if
+		 * enough pages are already being scanned that that high
+		 * watermark would be met at 100% efficiency.
 		 */
-		for (i = end_zone; i >= end_zone; i--) {
-			struct zone *zone = pgdat->node_zones + i;
-
-			if (!populated_zone(zone))
-				continue;
-
-			if (sc.priority != DEF_PRIORITY &&
-			    !pgdat_reclaimable(zone->zone_pgdat))
-				continue;
-
-			sc.nr_scanned = 0;
-			sc.reclaim_idx = i;
-
-			nr_soft_scanned = 0;
-			/*
-			 * Call soft limit reclaim before calling shrink_zone.
-			 */
-			nr_soft_reclaimed = mem_cgroup_soft_limit_reclaim(zone,
-							order, sc.gfp_mask,
-							&nr_soft_scanned);
-			sc.nr_reclaimed += nr_soft_reclaimed;
-
-			/*
-			 * There should be no need to raise the scanning
-			 * priority if enough pages are already being scanned
-			 * that that high watermark would be met at 100%
-			 * efficiency.
-			 */
-			if (kswapd_shrink_zone(zone, end_zone, &sc))
-				raise_priority = false;
-		}
+		if (kswapd_shrink_node(pgdat, classzone_idx, &sc))
+			raise_priority = false;
 
 		/*
 		 * If the low watermark is met there is no need for processes
@@ -3315,20 +3207,37 @@ static int balance_pgdat(pg_data_t *pgdat, int order, int classzone_idx)
 			break;
 
 		/*
+		 * Stop reclaiming if any eligible zone is balanced and clear
+		 * node writeback or congested.
+		 */
+		for (i = 0; i <= classzone_idx; i++) {
+			zone = pgdat->node_zones + i;
+			if (!populated_zone(zone))
+				continue;
+
+			if (zone_balanced(zone, sc.order, 0, classzone_idx)) {
+				clear_bit(PGDAT_CONGESTED, &pgdat->flags);
+				clear_bit(PGDAT_DIRTY, &pgdat->flags);
+				goto out;
+			}
+		}
+
+		/*
 		 * Raise priority if scanning rate is too low or there was no
 		 * progress in reclaiming pages
 		 */
 		if (raise_priority || !sc.nr_reclaimed)
 			sc.priority--;
-	} while (sc.priority >= 1 &&
-			!pgdat_balanced(pgdat, order, classzone_idx));
+	} while (sc.priority >= 1);
 
 out:
 	/*
-	 * Return the highest zone idx we were reclaiming at so
-	 * prepare_kswapd_sleep() makes the same decisions as here.
+	 * Return the order kswapd stopped reclaiming at as
+	 * prepare_kswapd_sleep() takes it into account. If another caller
+	 * entered the allocator slow path while kswapd was awake, order will
+	 * remain at the higher level.
 	 */
-	return end_zone;
+	return sc.order;
 }
 
 static void kswapd_try_to_sleep(pg_data_t *pgdat, int order,
@@ -3485,8 +3394,9 @@ static int kswapd(void *p)
 		 */
 		if (!ret) {
 			trace_mm_vmscan_kswapd_wake(pgdat->node_id, order);
-			balanced_classzone_idx = balance_pgdat(pgdat, order,
-								classzone_idx);
+
+			/* return value ignored until next patch */
+			balance_pgdat(pgdat, order, classzone_idx);
 		}
 	}
 
@@ -3516,7 +3426,7 @@ void wakeup_kswapd(struct zone *zone, int order, enum zone_type classzone_idx)
 	}
 	if (!waitqueue_active(&pgdat->kswapd_wait))
 		return;
-	if (zone_balanced(zone, order, true, 0, 0))
+	if (zone_balanced(zone, order, 0, 0))
 		return;
 
 	trace_mm_vmscan_wakeup_kswapd(pgdat->node_id, zone_idx(zone), order);
-- 
2.6.4

^ permalink raw reply	[flat|nested] 9+ messages in thread

end of thread, other threads:[~2016-06-23 11:31 UTC | newest]

Thread overview: 9+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
     [not found] <02fe01d1c48b$c44e9e80$4cebdb80$@alibaba-inc.com>
2016-06-12  9:33 ` [PATCH 06/27] mm, vmscan: Make kswapd reclaim in terms of nodes Hillf Danton
2016-06-14 14:52   ` Mel Gorman
     [not found] <071801d1cc5c$245087d0$6cf19770$@alibaba-inc.com>
2016-06-22  8:42 ` Hillf Danton
2016-06-23 11:31   ` Mel Gorman
2016-06-21 14:15 [PATCH 00/27] Move LRU page reclaim from zones to nodes v7 Mel Gorman
2016-06-21 14:15 ` [PATCH 06/27] mm, vmscan: Make kswapd reclaim in terms of nodes Mel Gorman
  -- strict thread matches above, loose matches on Subject: below --
2016-06-09 18:04 [PATCH 00/27] Move LRU page reclaim from zones to nodes v6 Mel Gorman
2016-06-09 18:04 ` [PATCH 06/27] mm, vmscan: Make kswapd reclaim in terms of nodes Mel Gorman
2016-06-15 14:23   ` Vlastimil Babka
2016-04-15  9:13 [PATCH 00/27] Move LRU page reclaim from zones to nodes v5 Mel Gorman
2016-04-15  9:13 ` [PATCH 06/27] mm, vmscan: Make kswapd reclaim in terms of nodes Mel Gorman
2016-04-28  8:36   ` Vlastimil Babka

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).