staging: lustre: fix bug with LL_MRF_RETURN in loop_make_request

Turns out that LL_MRF_RETURN is "void" so loop_make_request() would fall
through to the error path for a successful call.  Either this never
matters, or the error path really doesn't do anything.

Fix it up to do what the code is written to look like it is doing, if
that really is what it should be doing is another story...

Cc: Peng Tao <tao.peng@emc.com>
Cc: Zhao Hongjiang <zhaohongjiang@huawei.com>
Cc: Andreas Dilger <andreas.dilger@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

drivers/staging/lustre/lustre/include/linux/lustre_compat25.h

@@ -232,8 +232,6 @@ static inline int ll_namei_to_lookup_intent_flag(int flag)
 	return flag;
 }
 
-# define LL_MRF_RETURN(rc)
-
 #include <linux/fs.h>
 
 # define ll_umode_t	umode_t

drivers/staging/lustre/lustre/llite/lloop.c

@@ -365,10 +365,9 @@ static void loop_make_request(struct request_queue *q, struct bio *old_bio)
 		goto err;
 	}
 	loop_add_bio(lo, old_bio);
-	LL_MRF_RETURN(0);
+	return;
 err:
 	cfs_bio_io_error(old_bio, old_bio->bi_size);
-	LL_MRF_RETURN(0);
 }