md: fix use-after-free bug when dropping an rdev from an md array

Due to possible deadlock issues we need to use a schedule work to kobject_del
an 'rdev' object from a different thread.

A recent change means that kobject_add no longer gets a refernce, and
kobject_del doesn't put a reference.  Consequently, we need to explicitly hold
a reference to ensure that the last reference isn't dropped before the
scheduled work get a chance to call kobject_del.

Also, rename delayed_delete to md_delayed_delete to that it is more obvious in
a stack trace which code is to blame.

Cc: Al Viro <viro@ZenIV.linux.org.uk>
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
diff --git a/drivers/md/md.c b/drivers/md/md.c
index 08e0f8d..6cdd1b4 100644
--- a/drivers/md/md.c
+++ b/drivers/md/md.c
@@ -1419,10 +1419,11 @@
 	return err;
 }
 
-static void delayed_delete(struct work_struct *ws)
+static void md_delayed_delete(struct work_struct *ws)
 {
 	mdk_rdev_t *rdev = container_of(ws, mdk_rdev_t, del_work);
 	kobject_del(&rdev->kobj);
+	kobject_put(&rdev->kobj);
 }
 
 static void unbind_rdev_from_array(mdk_rdev_t * rdev)
@@ -1441,7 +1442,8 @@
 	/* We need to delay this, otherwise we can deadlock when
 	 * writing to 'remove' to "dev/state"
 	 */
-	INIT_WORK(&rdev->del_work, delayed_delete);
+	INIT_WORK(&rdev->del_work, md_delayed_delete);
+	kobject_get(&rdev->kobj);
 	schedule_work(&rdev->del_work);
 }
 
@@ -3686,7 +3688,7 @@
 				sysfs_remove_link(&mddev->kobj, nm);
 			}
 
-		/* make sure all delayed_delete calls have finished */
+		/* make sure all md_delayed_delete calls have finished */
 		flush_scheduled_work();
 
 		export_array(mddev);