md: avoid use of broken kzalloc mempool

The kzalloc mempool does not re-zero items that have been used and then
returned to the pool.  Manually zero the allocated multipath_bh instead.
Acked-by: NNeil Brown <neilb@suse.de>
Signed-off-by: NSage Weil <sage@newdream.net>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

drivers/md/multipath.c

@@ -150,6 +150,7 @@ static int multipath_make_request (struct request_queue *q, struct bio * bio)
 	}
 
 	mp_bh = mempool_alloc(conf->pool, GFP_NOIO);
+	memset(mp_bh, 0, sizeof(*mp_bh));
 
 	mp_bh->master_bio = bio;
 	mp_bh->mddev = mddev;
@@ -493,7 +494,7 @@ static int multipath_run (mddev_t *mddev)
 	}
 	mddev->degraded = conf->raid_disks - conf->working_disks;
 
-	conf->pool = mempool_create_kzalloc_pool(NR_RESERVED_BUFS,
+	conf->pool = mempool_create_kmalloc_pool(NR_RESERVED_BUFS,
 						 sizeof(struct multipath_bh));
 	if (conf->pool == NULL) {
 		printk(KERN_ERR