libceph: no more kick_requests() race

Since we no longer drop the request mutex between registering and
mapping an osd request in ceph_osdc_start_request(), there is no
chance of a race with kick_requests().

We can now therefore map and send the new request unconditionally
(but we'll issue a warning should it ever occur).
Signed-off-by: Alex Elder <elder@inktank.com>
Reviewed-off-by: Sage Weil <sage@inktank.com>

net/ceph/osd_client.c

@@ -1779,13 +1779,8 @@ int ceph_osdc_start_request(struct ceph_osd_client *osdc,
 
 	down_read(&osdc->map_sem);
 	mutex_lock(&osdc->request_mutex);
-	/*
-	 * a racing kick_requests() may have sent the message for us
-	 * while we dropped request_mutex above, so only send now if
-	 * the request still han't been touched yet.
-	 */
 	__register_request(osdc, req);
-	if (req->r_sent == 0) {
+	WARN_ON(req->r_sent);
 	rc = __map_request(osdc, req, 0);
 	if (rc < 0) {
 		if (nofail) {
@@ -1802,8 +1797,6 @@ int ceph_osdc_start_request(struct ceph_osd_client *osdc,
 		__send_request(osdc, req);
 	}
 	rc = 0;
-	}
-
 out_unlock:
 	mutex_unlock(&osdc->request_mutex);
 	up_read(&osdc->map_sem);