[PATCH 7/9] try harder on large allocations

[Joel Schopp <jschopp@austin.ibm.com>]

Fragmentation avoidance patches increase our chances of satisfying high order
allocations.  So this patch takes more than one iteration at trying to fulfill
those allocations because unlike before the extra iterations are often useful.

Signed-off-by: Mel Gorman <mel@csn.ul.ie>
Signed-off-by: Joel Schopp <jschopp@austin.ibm.com>

Index: 2.6.13-joel2/mm/page_alloc.c
===================================================================
--- 2.6.13-joel2.orig/mm/page_alloc.c	2005-09-21 11:13:14.%N -0500
+++ 2.6.13-joel2/mm/page_alloc.c	2005-09-21 11:14:49.%N -0500
@@ -944,7 +944,8 @@ __alloc_pages(unsigned int __nocast gfp_
 	int can_try_harder;
 	int did_some_progress;
 	int alloctype;
- 
+	int highorder_retry = 3;
+
 	alloctype = (gfp_mask & __GFP_RCLM_BITS);
 	might_sleep_if(wait);
 
@@ -1090,7 +1091,14 @@ rebalance:
 				goto got_pg;
 		}
 
-		out_of_memory(gfp_mask, order);
+		if (order < MAX_ORDER/2) out_of_memory(gfp_mask, order);
+		/*
+		 * Due to low fragmentation efforts, we should try a little
+		 * harder to satisfy high order allocations
+		 */
+		if (order >= MAX_ORDER/2 && --highorder_retry > 0)
+			goto rebalance;
+
 		goto restart;
 	}
 
@@ -1107,6 +1115,8 @@ rebalance:
 			do_retry = 1;
 		if (gfp_mask & __GFP_NOFAIL)
 			do_retry = 1;
+		if (order >= MAX_ORDER/2 && --highorder_retry > 0)
+			do_retry = 1;
 	}
 	if (do_retry) {
 		blk_congestion_wait(WRITE, HZ/50);