ptlrpcd_check() always scan all requests on ptlrpc_request_set::set_requests and try to finish completed requests, this is low efficiency. Even worse, l_wait_event() always checks condition for twice before sleeping and one more time after waking up, which means it will call ptlrpcd_check() for three times in each loop.

The solution is quite simple, ptlrpc_check_set() can move completed requests at head of ptlrpc_request_set::set_requests, so ptlrpcd_check doesn't need to scan all requests on a rqset anymore.

This is an improvement for LU-5053.