this was broken in 2.61

this makes it easier to diagnose elevon mixing issues

thanks Alexey!

allow disabling of the use of the barometer for vertical acceleration
compensation

this drives the speed of the 50Hz loop by the number of samples
accumulated in the IMU. This should give much more consistent timing
in DCM.

Thanks to Randy for introducing this scheme in ArduCopter!

this fixes using GPS_PROTOCOL to specify a specific GPS with a GPS
that takes a few updates before it works (eg. needing baud rate
changes).

This makes it easier to use an APM1-1280 with more features enabled

this allows for a larger amount of steering control at low speeds
without causing osciallation after takeoff

AP_InertialSensor: use AP_TimerProcess's queue_process to run read from MPU6000 after any currently running processes complete

Also restored suspend_timer to void return type

This reduces the maximum delay between when new sensor data arrives and when it is actually used by about 5ms.

ArduCopter: added NUM_IMU_SAMPLES_FOR_XYZHZ definitions for MPU6000 and Oilpan to allow more syncing of the mainloop with the arrival of data from the IMU.

AP_ADC: added num_samples_available to allow main loop to be synced with arrival of new sensor data for Oilpan.

AP_IMU: added num_samples_available to allow main loop timing to be synced with arrival of new data from IMU.

This reduces the delay between when data arrives and when it is used
by up to 1ms.

Added num_samples_available method to all InertialSensors to allow
main loop timing to be synced with sensors.

this allows for remote reboot of APM

when we fetch relative altitude waypoints from APM to a file, we need
to preserve the altitude

This reduces the number of rubbish pulses on the APM2 PWM output pins
on startup, which should help with booting some cheaper ESCs.

Thanks to Randy for the idea!

Changed verify_nav_roi command to attempt to point at the target if our camera mount type is not pan_tilt nor pan_tilt_roll

this removes the duplicate code. Throttle trim is not changed.

The nudge value depended on RC3_TRIM, which is not a reliable
value. It tried to only take effect when the throttle stick was above
50%, but if RC3_TRIM was high for some reason (say 1500) then it would
actually depress the throttle by a large amount, which could cause the
plane to stall.

This also adds a boolean option THROTTLE_NUDGE to allow disabling of
throttle nudging