mm: __rmqueue_fallback() should respect pageblock type

When __rmqueue_fallback() doesn't find a free block with the required size it splits a larger page and puts the rest of the page onto the free list. But it has one serious mistake. When putting back, __rmqueue_fallback() always use start_migratetype if type is not CMA. However, __rmqueue_fallback() is only called when all of the start_migratetype queue is empty. That said, __rmqueue_fallback always puts back memory to the wrong queue except try_to_steal_freepages() changed pageblock type (i.e. requested size is smaller than half of page block). The end result is that the antifragmentation framework increases fragmenation instead of decreasing it. Mel's original anti fragmentation does the right thing. But commit 47118af076f6 ("mm: mmzone: MIGRATE_CMA migration type added") broke it. This patch restores sane and old behavior. It also removes an incorrect comment which was introduced by commit fef903efcf0c ("mm/page_alloc.c: restructure free-page stealing code and fix a bug"). Signed-off-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Cc: Mel Gorman <mgorman@suse.de> Cc: Michal Nazarewicz <mina86@mina86.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

mm: __rmqueue_fallback() should respect pageblock type
When __rmqueue_fallback() doesn't find a free block with the required size it splits a larger page and puts the rest of the page onto the free list. But it has one serious mistake. When putting back, __rmqueue_fallback() always use start_migratetype if type is not CMA. However, __rmqueue_fallback() is only called when all of the start_migratetype queue is empty. That said, __rmqueue_fallback always puts back memory to the wrong queue except try_to_steal_freepages() changed pageblock type (i.e. requested size is smaller than half of page block). The end result is that the antifragmentation framework increases fragmenation instead of decreasing it. Mel's original anti fragmentation does the right thing. But commit 47118af076f6 ("mm: mmzone: MIGRATE_CMA migration type added") broke it. This patch restores sane and old behavior. It also removes an incorrect comment which was introduced by commit fef903efcf0c ("mm/page_alloc.c: restructure free-page stealing code and fix a bug"). Signed-off-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Cc: Mel Gorman <mgorman@suse.de> Cc: Michal Nazarewicz <mina86@mina86.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
KOSAKI Motohiro · Linus Torvalds
1 parent 52c8f6a5ae
Showing 1 changed file with 5 additions and 10 deletions Side-by-side Diff
mm/page_alloc.c
@@ -1027,6 +1027,10 @@
 {
 	int current_order = page_order(page);
  
+	/*
+	 * When borrowing from MIGRATE_CMA, we need to release the excess
+	 * buddy pages to CMA itself.
+	 */
 	if (is_migrate_cma(fallback_type))
 		return fallback_type;
  
  
@@ -1091,17 +1095,8 @@
 			list_del(&page->lru);
 			rmv_page_order(page);
  
-			/*
-			 * Borrow the excess buddy pages as well, irrespective
-			 * of whether we stole freepages, or took ownership of
-			 * the pageblock or not.
-			 *
-			 * Exception: When borrowing from MIGRATE_CMA, release
-			 * the excess buddy pages to CMA itself.
-			 */
 			expand(zone, page, order, current_order, area,
-			       is_migrate_cma(migratetype)
-			     ? migratetype : start_migratetype);
+			       new_type);
  
 			trace_mm_page_alloc_extfrag(page, order, current_order,
 				start_migratetype, migratetype, new_type);
...	...	@@ -1027,6 +1027,10 @@
1027	1027	{
1028	1028	int current_order = page_order(page);
1029	1029
	1030	+ /*
	1031	+ * When borrowing from MIGRATE_CMA, we need to release the excess
	1032	+ * buddy pages to CMA itself.
	1033	+ */
1030	1034	if (is_migrate_cma(fallback_type))
1031	1035	return fallback_type;
1032	1036
1033	1037
...	...	@@ -1091,17 +1095,8 @@
1091	1095	list_del(&page->lru);
1092	1096	rmv_page_order(page);
1093	1097
1094		- /*
1095		- * Borrow the excess buddy pages as well, irrespective
1096		- * of whether we stole freepages, or took ownership of
1097		- * the pageblock or not.
1098		- *
1099		- * Exception: When borrowing from MIGRATE_CMA, release
1100		- * the excess buddy pages to CMA itself.
1101		- */
1102	1098	expand(zone, page, order, current_order, area,
1103		- is_migrate_cma(migratetype)
1104		- ? migratetype : start_migratetype);
	1099	+ new_type);
1105	1100
1106	1101	trace_mm_page_alloc_extfrag(page, order, current_order,
1107	1102	start_migratetype, migratetype, new_type);