Baro data arriving too soon after the last measurement due to a high sampling rate or timing jitter is rejected inside the ecl EKF to prevent the data buffer overflowing.
This patch checks the timestamp difference from the last measurement, and if to small, the data is accumulated and the average sent to the EKF when the time delta
is acceptable.

Fixes:

1) Invalid data with a zero time stamp could be the EKF ends up in the data buffers and result in loss of 'good' data from the buffers

2) Magnetometer data was arriving at a rate faster than the data buffers could handle resulting in loss of data.

If the replay data for the baro or mag data has a zero time stamp, then the corresponding relative timestamp published over the combined sensor topic must be se to RELATIVE_TIMESTAMP_INVALID so that the ekf2 module does not try to use this data.

* Fix jmavsim HITL simulation of MAV_CMD_DO_REPOSITION in the case where you have no radio attached to the PX4 and so you have disabled RC link loss for that reason (set NAV_RCL_ACT = 0) but you still want the jmavsimulation to work.  The line of code changed here causes failsafe RTL to kick in without this fix.

* Add altitude hold option using Z position control, while doing velocity control on vx and vy.

* Fix style and rebase issues

Reformatting makes it look worse than it is, but it really only adds {}

This saves ~1.1KB of RAM for systems with only 1 sensor of each type.

If there are more sensors, they will be dynamically added, such that
failover still works.

reduces the stack depth

* param: fix bug when param value is changed to 'zero' for the first time, it won't be saved.

* param: revert incorrectly removed code.

This fixes the case where more than 10 args were needed for instance
when starting mavlink with all its options.

Add GitHub links to maintenance team list.

Listing existing maintainers explicitly.

Current implementation checks against geofence only if it is in polygon form.
When it's created via params, it accepts all the waypoints as the number of vertices = 0.
Thus, changed the function to the one that is used to check whether geofence is breached in flight.

   Update the comment, to explain how to achive a different perescale
   value.

   Added PX4_IO_TIMER_ALTERNATE_RATE one board agnostic ifdef
   PX4_IO_TIMER_ALTERNATE_RATE that is non board specific.

   N.B.  I would like to eliminate PX4_IO_TIMER_ALTERNATE_RATE
   as well, but I need crazyflie HW to validate that the startup
   script can just set the rate on the PWM to 3921 fast enough to
   not effect the motors.

 Add PX4_PWM_ALTERNATE_RANGES in suport of a board agnostic way
 to override the PWM_.*_{MIN|MAX} values

 Removed #ifdef in version.h

   1) Remove uneeded spi and reset code
   2) Use the Board commin for providing the BOARD_NAME
   3) Add PX4_PWM_ALTERNATE_RANGES in suport of a board agnostic
      way to override the PWM_.*_{MIN|MAX} values
   4) Remove #ifdefs from the IO timers*. Drive the config deltas
      from the crazyflie_timer_config.c and one board agnostic
      ifdef PX4_IO_TIMER_ALTERNATE_RATE that is non board specific

      *I would like to eliminate PX4_IO_TIMER_ALTERNATE_RATE
      as well, but I need HW to validate that the startup
      script can just set the rate on the PWM to 3921 fast
      enough to not effect the motors.

  CONFIG_STM32_I2CTIMEOTICKS is hard defined on PX4 master
 and defconfig defined on upstream Nuttx (nuttx_v3) it will be fixed
 there.

 Aligned Make.Defs to use the build system's defined warnigns.

- fixes #5061

in C++11, double braces are needed for std::array aggregate initialization,
or assignment with =.
see: http://en.cppreference.com/w/cpp/container/array

See discussion in https://github.com/PX4/Firmware/issues/5756

'mavlink stop-all' during a log streaming session previously led to a
resource leak, and log streaming could only be re-started by rebooting the
system.

This is helpful, so that a connected partner can switch from broadcasting
to a 'connected' state.