congestion_wait(): don't use WRITE
authorKOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Mon, 26 Oct 2009 23:49:35 +0000 (16:49 -0700)
committerLinus Torvalds <torvalds@linux-foundation.org>
Thu, 29 Oct 2009 14:39:25 +0000 (07:39 -0700)
commit 8aa7e847d (Fix congestion_wait() sync/async vs read/write
confusion) replace WRITE with BLK_RW_ASYNC.  Unfortunately, concurrent mm
development made the unchanged place accidentally.

This patch fixes it too.

Signed-off-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Acked-by: Jens Axboe <jens.axboe@oracle.com>
Acked-by: Rik van Riel <riel@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
mm/vmscan.c

index 64e438898832371277e8baeb89fd5ebb1f72a14a..fbb9f6bdad6ea2d1d6c7a1d5f66300484064c98e 100644 (file)
@@ -1088,7 +1088,7 @@ static unsigned long shrink_inactive_list(unsigned long max_scan,
        int lumpy_reclaim = 0;
 
        while (unlikely(too_many_isolated(zone, file, sc))) {
-               congestion_wait(WRITE, HZ/10);
+               congestion_wait(BLK_RW_ASYNC, HZ/10);
 
                /* We are about to die and free our memory. Return now. */
                if (fatal_signal_pending(current))