- ecl in PX4/Firmware (e4088204): https://github.com/PX4/ecl/commit/aa134d5a617cd90142d82039c4466aef2f009383
    - ecl current upstream: https://github.com/PX4/ecl/commit/56b8bb08a16f0e86f7c7dfe458f0d2d3082e92e4
    - Changes: https://github.com/PX4/ecl/compare/aa134d5a617cd90142d82039c4466aef2f009383...56b8bb08a16f0e86f7c7dfe458f0d2d3082e92e4

    56b8bb0 2018-08-23 dakejahl - added a checking for range_aid to determine if we are relying on the range finder as our primary source of height observation (#491)

tfmini: fix format and fix driver start / stop logic so it works when running PX4 as threads on Linux.

This reduces the scheduling interval to 9ms, such that the driver is always ready to read new data. Running it at exactly 100Hz is not correct since the driver and sensor measurement intervals are not "in sync", causing the driver to miss data. This causes a fill-up of the UART buffer.

    - nuttx in PX4/Firmware (72792cef): https://github.com/PX4-NuttX/nuttx/commit/bf8cce85c3f9310bcff8cdd1d979270419782a6e
    - nuttx current upstream: https://github.com/PX4-NuttX/nuttx/commit/bf8cce85c3f9310bcff8cdd1d979270419782a6e
    - Changes: https://github.com/PX4-NuttX/nuttx/compare/bf8cce85c3f9310bcff8cdd1d979270419782a6e...bf8cce85c3f9310bcff8cdd1d979270419782a6e

 - fix ocpoc line endings

      Added paramater based low pass fileter settting from
      IMU_ACCEL_CUTOFF.

      removed gotos.

  Added paramater based low pass fileter settting from
  IMU_GYRO_CUTOFF.

  Added interfaces for setting the HW low pass filter.
  Set HW LPF to 64 Hz

  Changed poll rate to ODR to 800 Hz.

  Documeted the TIMER_REDUCTION value as 20% and changed it to
  250 Us / 1250 us.

  Added standby control API, to insure chip is configured in
  standby mode.

  removed tabs.

  removed gotos.

Fixed-wing autoland: Introduce parameter to enable/disable the early activation of the landing configuration

Fixed-wing autoland: Remove the dynamic altitude acceptance calculation and just use a simple separate landing altitude acceptance radius instead. This gives users that do not use LOITER-TO-ALT waypoints more control over their landing procedure

Fixed-wing autoland: ALWAYS flare when close to the ground, independently of the horizontal distance to the land WP. This avoids that we crash into the ground at negative pitch and guarantees that the flare always starts at the same altitude above ground. However, the motor shutoff still depends on the horizontal distance to the LAND WP, thus avoiding that the motor is shut off prematurely

Fixed-wing autoland: Use a more appropriate (i.e. mostly tighter) altitude acceptance radius than just the standard altitude acceptance parameter (which may be too large to allow a precise autoland)

Fixed-wing autoland: 1) The landing configuration (flaps, different airspeed) is now already set during the loiter down instead of at the start of the landing approach. This is done to avoid any mode changes (which can cause altitude/airspeed jumps) so close to the gorund. 2) A scaling factor for the TECS throttle time constant was added which allows tighter throttle control during the landing (i.e. close to the ground) than high up in the air

  Logger was hard faulting on start due to a write to
  null pointer dereferece of errata

    NXPhlite does not have a companion link. So start
    Mavlink on ttyS4 and ttyACM only.

    - ecl in PX4/Firmware (8b6b34b8f073fc47fdce658ab5c1bccd7e1682db): https://github.com/PX4/ecl/commit/34cd1366b715d80885d71f80c72cc5d0bfde64f0
    - ecl current upstream: https://github.com/PX4/ecl/commit/aa134d5a617cd90142d82039c4466aef2f009383
    - Changes: https://github.com/PX4/ecl/compare/34cd1366b715d80885d71f80c72cc5d0bfde64f0...aa134d5a617cd90142d82039c4466aef2f009383

    aa134d5 2018-08-20 Roman - l1 position controller: make sure that a NAN roll setpoint does not keep the roll setpoint state at NAN forever

Directly set parameters in airframe config files rather than setting script variables and move the param set lines into  logic blocks, and alphabetize and tidy up param set lists.

- the timestamp is only used for logging and log analysis. For that it's
  important to have the timestamp when a setpoint becomes active.
- there was a consistent problem with the position_setpoint_triplet
  timestamp, where the timestamp was just bogus. Timesync seems to work
  correctly though. Might be a problem on the sender side?
  For example here:
  https://logs.px4.io/plot_app?log=41918a7d-4c1d-464d-9abe-aef2c0818d92

This fixes the issue where the vehicle yaws in a seemingly random direction when RC is lost.