With GCC 4.9 the binary is to large for the flash memory.
This is why we disabled LPE on that platform.

We propagate the bus parameters from the bus interface to the sensor
devices. Thus, the device ID of the sensor driver is set to the correct
bus id and address. Otherwise it would be zero, which is an issue if several MPU9250s
are running at the same time.

This fixes the following corner case:
1. Upload a mission.
2. Set mission mode.
3. Set loiter mode.
4. Arm.
At this point it will shoot up and go to the takeoff waypoint even
though we're not in mission but in loiter mode.

The fix makes sure that the triplet is reset to invalid (and idle) in
loiter mode if we're landed and disarmed.
It will lead to the vehcle sit in idle on the ground until you issue a
start mission (or takeoff) command.

- fix a bug where the wrong rotation order was used to compute the attitude
setpoint when using the pusher assist feature

Signed-off-by: Roman <bapstroman@gmail.com>

updated VT_FW_MOT_OFFID

Replaced individual elevator and aileron into "elevons"

-closes #5862

This makes the onboard check consistent with hmc5883_bus_option
initialization. If the current bus is the onboard bus, return !external.

It fixes the onboard mag (HMC5883) for AeroFC. Its priority is now 100,
previously it was 255, so that if an external mag is attached, it will be
preferred.

This reverts commit 126172d7.

This commit is going in the wrong direction: it's actually the same as
it was before, but allows changing the autostart configuration to
something other than 4070, which means people could wrongly assume it
would work in other configurations.

Needed for log streaming

- fixed a bug where tilt compensation was done also when the horizontal
velocity was controlled. This is not needed because in this case
the controller outputs a 3D thrust vector.

Signed-off-by: Roman <bapstroman@gmail.com>

* Fixes for sitl gazebo ground truth.

* Switch ekf2 to new logging module for sitl.