Fix mpu9250 hardfaults

[dakejahl]

Problem
The mpu9250 driver will cause a system hard fault if the driver is attempted to be started on an already started bus. This is due to the introduction of extraneous bus options in #10116. Specifically the bus options for MPU_DEVICE_PATH_MPU6500_* . These options used the name exact CS, path, etc as the non 6500 option, however the dev pointer was still NULL, so it tried to start the bus.

To reproduce
This problem behavior can easily be observed by issuing an mpu9250 start via the NSH after the system has booted.

Solution
I've removed the 6500 bus options. These are not necessary, as the device type will be discovered via the WHOAMI register (besides, the only difference between the 6500 and 9250 is the mag). I've also removed the extra bus options for the ICM20948 as they do not appear to be used.

@flochir Could you please test this to make sure everything is still good on your end?

[dakejahl]

@dagar My style check is all jacked up. It forces
delete(bus.dev);
locally, but jenkins wants
delete (bus.dev);

Also make format doesn't work, still getting

Error: you're using Artistic Style Version 2.05.1
but should be using Artistic Style Version 2.06, Artistic Style Version 3.0, Artistic Style Version 3.0.1, or Artistic Style Version 3.1 instead

[LorenzMeier]

You will need to update astyle in your local install.

[dakejahl]

[dakejahl]

ROS tests are failing, not sure what's going on here.

[dagar]

This looks fine, but let's do a quick check on the various setups.

• pixracer (single mpu9250)
• TealOne (mpu9250 + mpu6500)
• external Here GPS (icm20948)
• crazyflie mpu9250 i2c?

[flochir]

@dagar No hardfault on Pixhawk 2.1 Cube, and no problems with ICM20948 on Here GNSS.