mm: compaction: do not compact pgdat for order-0
authorMel Gorman <mgorman@suse.de>
Wed, 11 Sep 2013 21:22:19 +0000 (14:22 -0700)
committerLinus Torvalds <torvalds@linux-foundation.org>
Wed, 11 Sep 2013 22:57:55 +0000 (15:57 -0700)
If kswapd was reclaiming for a high order and resets it to 0 due to
fragmentation it will still call compact_pgdat.  For the most part, this
will fail a compaction_suitable() test and not compact but it is
unnecessarily sloppy.  It could be fixed in the caller but fix it in the
API instead.

[dhillf@gmail.com: pointed out that it was a potential problem]
Signed-off-by: Mel Gorman <mgorman@suse.de>
Cc: Hillf Danton <dhillf@gmail.com>
Acked-by: Minchan Kim <minchan@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
mm/compaction.c

index 05ccb4cc0bdb984dc2613461703587ec199804cd..c43789388cd8667536bfd9644a5bbe2cd0d35f3e 100644 (file)
@@ -1131,6 +1131,9 @@ void compact_pgdat(pg_data_t *pgdat, int order)
                .sync = false,
        };
 
+       if (!order)
+               return;
+
        __compact_pgdat(pgdat, &cc);
 }