[PATCH] mm: stop kswapd's infinite loop at high order allocation

[KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>]

ok, wassim confirmed this patch works well.


==
From: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Subject: [PATCH] mm: kswapd stop infinite loop at high order allocation

Wassim Dagash reported following kswapd infinite loop problem.

  kswapd runs in some infinite loop trying to swap until order 10 of zone
  highmem is OK, While zone higmem (as I understand) has nothing to do
  with contiguous memory (cause there is no 1-1 mapping) which means
  kswapd will continue to try to balance order 10 of zone highmem
  forever (or until someone release a very large chunk of highmem).

He proposed remove contenious checking on highmem at all.
However hugepage on highmem need contenious highmem page.

To add infinite loop stopper is simple and good.



Reported-by: wassim dagash <wassim.dagash@gmail.com>
Signed-off-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
---
 mm/vmscan.c |   11 +++++++++++
 1 file changed, 11 insertions(+)

Index: b/mm/vmscan.c
===================================================================
--- a/mm/vmscan.c
+++ b/mm/vmscan.c
@@ -1872,6 +1872,17 @@ out:
 
 		try_to_freeze();
 
+		/*
+		 * When highmem is very fragmented,
+		 * alloc_pages(GFP_KERNEL, very-high-order) can cause
+		 * infinite loop because zone_watermark_ok(highmem) failed.
+		 * However, alloc_pages(GFP_KERNEL..) indicate highmem memory
+		 * continuousness isn't necessary.
+		 * Therefore we don't want contenious check at 2nd loop.
+		 */
+		if (nr_reclaimed < SWAP_CLUSTER_MAX)
+			order = sc.order = 0;
+
 		goto loop_again;
 	}
 


--
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>