Btrfs: barrier before waitqueue_active

We need a barrir before calling waitqueue_active otherwise we will miss
wakeups.  So in places that do atomic_dec(); then atomic_read() use
atomic_dec_return() which imply a memory barrier (see memory-barriers.txt)
and then add an explicit memory barrier everywhere else that need them.
Thanks,

Signed-off-by: Josef Bacik <jbacik@fusionio.com>
diff --git a/fs/btrfs/delayed-inode.c b/fs/btrfs/delayed-inode.c
index 00deed4e..07d5eeb 100644
--- a/fs/btrfs/delayed-inode.c
+++ b/fs/btrfs/delayed-inode.c
@@ -512,8 +512,8 @@
 
 	rb_erase(&delayed_item->rb_node, root);
 	delayed_item->delayed_node->count--;
-	atomic_dec(&delayed_root->items);
-	if (atomic_read(&delayed_root->items) < BTRFS_DELAYED_BACKGROUND &&
+	if (atomic_dec_return(&delayed_root->items) <
+	    BTRFS_DELAYED_BACKGROUND &&
 	    waitqueue_active(&delayed_root->wait))
 		wake_up(&delayed_root->wait);
 }
@@ -1056,8 +1056,7 @@
 		delayed_node->count--;
 
 		delayed_root = delayed_node->root->fs_info->delayed_root;
-		atomic_dec(&delayed_root->items);
-		if (atomic_read(&delayed_root->items) <
+		if (atomic_dec_return(&delayed_root->items) <
 		    BTRFS_DELAYED_BACKGROUND &&
 		    waitqueue_active(&delayed_root->wait))
 			wake_up(&delayed_root->wait);