When stacking request-based dm device on non blk-mq device and
device-mapper target could not map the request (error target is used,
multipath target with all paths down, etc), the WARN_ON_ONCE() in
free_rq_clone() will trigger when it shouldn't.
The warning was added by commit 
aa6df8d ("dm: fix free_rq_clone() NULL
pointer when requeueing unmapped request").  But free_rq_clone() with
clone->q == NULL is valid usage for the case where
dm_kill_unmapped_request() initiates request cleanup.
Fix this false warning by just removing the WARN_ON -- it only generated
false positives and was never useful in catching the intended case
(completing clone request not being mapped e.g. clone->q being NULL).
Fixes: aa6df8d ("dm: fix free_rq_clone() NULL pointer when requeueing unmapped request")
Reported-by: Bart Van Assche <bart.vanassche@sandisk.com>
Reported-by: Jun'ichi Nomura <j-nomura@ce.jp.nec.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
 
        dm_put(md);
 }
 
-static void free_rq_clone(struct request *clone, bool must_be_mapped)
+static void free_rq_clone(struct request *clone)
 {
        struct dm_rq_target_io *tio = clone->end_io_data;
        struct mapped_device *md = tio->md;
 
-       WARN_ON_ONCE(must_be_mapped && !clone->q);
-
        blk_rq_unprep_clone(clone);
 
        if (md->type == DM_TYPE_MQ_REQUEST_BASED)
                        rq->sense_len = clone->sense_len;
        }
 
-       free_rq_clone(clone, true);
+       free_rq_clone(clone);
        if (!rq->q->mq_ops)
                blk_end_request_all(rq, error);
        else
        }
 
        if (clone)
-               free_rq_clone(clone, false);
+               free_rq_clone(clone);
 }
 
 /*