MPU9250 is mostly an MPU6500 with a mag (AK8963) in the same package.
Support driving MPU6500 with the MPU9250 driver. The id of the driver
isn't set differently since this way it allows to force a recalibration.

Ideally MPU9250 driver could even not exist and the support for these
sensors be merged back in the MPU6000 that's more complete. This is an
intermediate step in that direction.

Move the addresses to .cpp where they are used. This should help
noticing discrepancies like the OSR we are using and the sampling
interval.

  - There's no I2C2, so remove it
  - Fix alignments everywhere in board_config.h
  - Clarify STM32 ADC (not available) and the ADC on the FGPA
  - Fix format on some comments

All the other drivers use 16G, do the same for MPU6500.

We are getting way to much transfer errors with high baudrate

MC_YAWRATE_MAX 20  => 50

Signed-off-by: bresch <brescianimathieu@gmail.com>

Seems that the work queue is not available for the tests.

These configs do not have the work queue enabled.

Some NuttX configs do not have the work queues enabled.

not needed anymore because of autosave

Not needed anymore due to autosave

to protect against concurrent writes to the same file

- add a saving delay of 300ms
- save at most once every 2 seconds

This got introduced the the addition of the perf counter to the log.

Fixes the error:
unknown log message type 73