[v2,3/5] mm/page_alloc: move pages to tail in move_to_free_list()

Commit Message

David Hildenbrand Oct. 5, 2020, 12:15 p.m. UTC

Whenever we move pages between freelists via move_to_free_list()/
move_freepages_block(), we don't actually touch the pages:
1. Page isolation doesn't actually touch the pages, it simply isolates
   pageblocks and moves all free pages to the MIGRATE_ISOLATE freelist.
   When undoing isolation, we move the pages back to the target list.
2. Page stealing (steal_suitable_fallback()) moves free pages directly
   between lists without touching them.
3. reserve_highatomic_pageblock()/unreserve_highatomic_pageblock() moves
   free pages directly between freelists without touching them.

We already place pages to the tail of the freelists when undoing isolation
via __putback_isolated_page(), let's do it in any case (e.g., if order <=
pageblock_order) and document the behavior. To simplify, let's move the
pages to the tail for all move_to_free_list()/move_freepages_block() users.

In 2., the target list is empty, so there should be no change. In 3.,
we might observe a change, however, highatomic is more concerned about
allocations succeeding than cache hotness - if we ever realize this
change degrades a workload, we can special-case this instance and add a
proper comment.

This change results in all pages getting onlined via online_pages() to
be placed to the tail of the freelist.

Reviewed-by: Oscar Salvador <[e-mail]>
Acked-by: Pankaj Gupta <[e-mail]>
Reviewed-by: Wei Yang <[e-mail]>
Cc: Andrew Morton <[e-mail]>
Cc: Alexander Duyck <[e-mail]>
Cc: Mel Gorman <[e-mail]>
Cc: Michal Hocko <[e-mail]>
Cc: Dave Hansen <[e-mail]>
Cc: Vlastimil Babka <[e-mail]>
Cc: Wei Yang <[e-mail]>
Cc: Oscar Salvador <[e-mail]>
Cc: Mike Rapoport <[e-mail]>
Cc: Scott Cheloha <[e-mail]>
Cc: Michael Ellerman <[e-mail]>
Signed-off-by: David Hildenbrand <[e-mail]>
---
 mm/page_alloc.c     | 10 +++++++---
 mm/page_isolation.c |  5 +++++
 2 files changed, 12 insertions(+), 3 deletions(-)

Comments

Michal Hocko Oct. 6, 2020, 12:12 p.m. UTC | #1

On Mon 05-10-20 14:15:32, David Hildenbrand wrote:
> Whenever we move pages between freelists via move_to_free_list()/
> move_freepages_block(), we don't actually touch the pages:
> 1. Page isolation doesn't actually touch the pages, it simply isolates
>    pageblocks and moves all free pages to the MIGRATE_ISOLATE freelist.
>    When undoing isolation, we move the pages back to the target list.
> 2. Page stealing (steal_suitable_fallback()) moves free pages directly
>    between lists without touching them.
> 3. reserve_highatomic_pageblock()/unreserve_highatomic_pageblock() moves
>    free pages directly between freelists without touching them.
> 
> We already place pages to the tail of the freelists when undoing isolation
> via __putback_isolated_page(), let's do it in any case (e.g., if order <=
> pageblock_order) and document the behavior. To simplify, let's move the
> pages to the tail for all move_to_free_list()/move_freepages_block() users.
> 
> In 2., the target list is empty, so there should be no change. In 3.,
> we might observe a change, however, highatomic is more concerned about
> allocations succeeding than cache hotness - if we ever realize this
> change degrades a workload, we can special-case this instance and add a
> proper comment.
> 
> This change results in all pages getting onlined via online_pages() to
> be placed to the tail of the freelist.
> 
> Reviewed-by: Oscar Salvador <[e-mail]>
> Acked-by: Pankaj Gupta <[e-mail]>
> Reviewed-by: Wei Yang <[e-mail]>
> Cc: Andrew Morton <[e-mail]>
> Cc: Alexander Duyck <[e-mail]>
> Cc: Mel Gorman <[e-mail]>
> Cc: Michal Hocko <[e-mail]>
> Cc: Dave Hansen <[e-mail]>
> Cc: Vlastimil Babka <[e-mail]>
> Cc: Wei Yang <[e-mail]>
> Cc: Oscar Salvador <[e-mail]>
> Cc: Mike Rapoport <[e-mail]>
> Cc: Scott Cheloha <[e-mail]>
> Cc: Michael Ellerman <[e-mail]>
> Signed-off-by: David Hildenbrand <[e-mail]>

Much simpler!
Acked-by: Michal Hocko <[e-mail]>

Thanks!

> ---
>  mm/page_alloc.c     | 10 +++++++---
>  mm/page_isolation.c |  5 +++++
>  2 files changed, 12 insertions(+), 3 deletions(-)
> 
> diff --git a/mm/page_alloc.c b/mm/page_alloc.c
> index df5ff0cd6df1..b187e46cf640 100644
> --- a/mm/page_alloc.c
> +++ b/mm/page_alloc.c
> @@ -901,13 +901,17 @@ static inline void add_to_free_list_tail(struct page *page, struct zone *zone,
>  	area->nr_free++;
>  }
>  
> -/* Used for pages which are on another list */
> +/*
> + * Used for pages which are on another list. Move the pages to the tail
> + * of the list - so the moved pages won't immediately be considered for
> + * allocation again (e.g., optimization for memory onlining).
> + */
>  static inline void move_to_free_list(struct page *page, struct zone *zone,
>  				     unsigned int order, int migratetype)
>  {
>  	struct free_area *area = &zone->free_area[order];
>  
> -	list_move(&page->lru, &area->free_list[migratetype]);
> +	list_move_tail(&page->lru, &area->free_list[migratetype]);
>  }
>  
>  static inline void del_page_from_free_list(struct page *page, struct zone *zone,
> @@ -2340,7 +2344,7 @@ static inline struct page *__rmqueue_cma_fallback(struct zone *zone,
>  #endif
>  
>  /*
> - * Move the free pages in a range to the free lists of the requested type.
> + * Move the free pages in a range to the freelist tail of the requested type.
>   * Note that start_page and end_pages are not aligned on a pageblock
>   * boundary. If alignment is required, use move_freepages_block()
>   */
> diff --git a/mm/page_isolation.c b/mm/page_isolation.c
> index abfe26ad59fd..83692b937784 100644
> --- a/mm/page_isolation.c
> +++ b/mm/page_isolation.c
> @@ -106,6 +106,11 @@ static void unset_migratetype_isolate(struct page *page, unsigned migratetype)
>  	 * If we isolate freepage with more than pageblock_order, there
>  	 * should be no freepage in the range, so we could avoid costly
>  	 * pageblock scanning for freepage moving.
> +	 *
> +	 * We didn't actually touch any of the isolated pages, so place them
> +	 * to the tail of the freelist. This is an optimization for memory
> +	 * onlining - just onlined memory won't immediately be considered for
> +	 * allocation.
>  	 */
>  	if (!isolated_page) {
>  		nr_pages = move_freepages_block(zone, page, migratetype, NULL);
> -- 
> 2.26.2

Vlastimil Babka Oct. 20, 2020, 5:20 p.m. UTC | #2

On 10/5/20 2:15 PM, David Hildenbrand wrote:
> Whenever we move pages between freelists via move_to_free_list()/
> move_freepages_block(), we don't actually touch the pages:
> 1. Page isolation doesn't actually touch the pages, it simply isolates
>     pageblocks and moves all free pages to the MIGRATE_ISOLATE freelist.
>     When undoing isolation, we move the pages back to the target list.
> 2. Page stealing (steal_suitable_fallback()) moves free pages directly
>     between lists without touching them.
> 3. reserve_highatomic_pageblock()/unreserve_highatomic_pageblock() moves
>     free pages directly between freelists without touching them.
> 
> We already place pages to the tail of the freelists when undoing isolation
> via __putback_isolated_page(), let's do it in any case (e.g., if order <=
> pageblock_order) and document the behavior. To simplify, let's move the
> pages to the tail for all move_to_free_list()/move_freepages_block() users.
> 
> In 2., the target list is empty, so there should be no change. In 3.,
> we might observe a change, however, highatomic is more concerned about
> allocations succeeding than cache hotness - if we ever realize this
> change degrades a workload, we can special-case this instance and add a
> proper comment.
> 
> This change results in all pages getting onlined via online_pages() to
> be placed to the tail of the freelist.
> 
> Reviewed-by: Oscar Salvador <[e-mail]>
> Acked-by: Pankaj Gupta <[e-mail]>
> Reviewed-by: Wei Yang <[e-mail]>

Reviewed-by: Vlastimil Babka <[e-mail]>

Patch

diff --git a/mm/page_alloc.c b/mm/page_alloc.c
index df5ff0cd6df1..b187e46cf640 100644
--- a/mm/page_alloc.c
+++ b/mm/page_alloc.c
@@ -901,13 +901,17 @@  static inline void add_to_free_list_tail(struct page *page, struct zone *zone,
 	area->nr_free++;
 }
 
-/* Used for pages which are on another list */
+/*
+ * Used for pages which are on another list. Move the pages to the tail
+ * of the list - so the moved pages won't immediately be considered for
+ * allocation again (e.g., optimization for memory onlining).
+ */
 static inline void move_to_free_list(struct page *page, struct zone *zone,
 				     unsigned int order, int migratetype)
 {
 	struct free_area *area = &zone->free_area[order];
 
-	list_move(&page->lru, &area->free_list[migratetype]);
+	list_move_tail(&page->lru, &area->free_list[migratetype]);
 }
 
 static inline void del_page_from_free_list(struct page *page, struct zone *zone,
@@ -2340,7 +2344,7 @@  static inline struct page *__rmqueue_cma_fallback(struct zone *zone,
 #endif
 
 /*
- * Move the free pages in a range to the free lists of the requested type.
+ * Move the free pages in a range to the freelist tail of the requested type.
  * Note that start_page and end_pages are not aligned on a pageblock
  * boundary. If alignment is required, use move_freepages_block()
  */
diff --git a/mm/page_isolation.c b/mm/page_isolation.c
index abfe26ad59fd..83692b937784 100644
--- a/mm/page_isolation.c
+++ b/mm/page_isolation.c
@@ -106,6 +106,11 @@  static void unset_migratetype_isolate(struct page *page, unsigned migratetype)
 	 * If we isolate freepage with more than pageblock_order, there
 	 * should be no freepage in the range, so we could avoid costly
 	 * pageblock scanning for freepage moving.
+	 *
+	 * We didn't actually touch any of the isolated pages, so place them
+	 * to the tail of the freelist. This is an optimization for memory
+	 * onlining - just onlined memory won't immediately be considered for
+	 * allocation.
 	 */
 	if (!isolated_page) {
 		nr_pages = move_freepages_block(zone, page, migratetype, NULL);