Sorry for the late answer. At some later point, if we still use that MPU, we also use the J120/J140 boards from Auvidea, I would look into backporting the driver form here:
(or any newer version).
It looks like 9150 works basically exactly the same (commitwise), but since the ID is different it would be considered unsupported by the driver.
Also I have not check the interrupt support or if it does interrupt timestamping (would be nice), but at least it supports device-tree (the 4.4 version does not and is missing the more recent INV MPUs). I have not worked with the IIO/IMU subsystem so far, so I don’t know how easy it would be to us, but of course better latency and a common interface is always better ;)
But If I were you I would begin there, since the driver is in mainline I would not use any other.