AP_Param: add @SYS/defaults.parm and log default values

[IamPete1]

This adds a @SYS/defaults.parm ftp file that contains all the defaults for the board. Those default values are also logged in the PARM log at first arming. You do not get the defaults again with run-time changes.

This works, however we need to make sure everywhere is calling set_default not just set, this induces changes to AC_PID to follow that pattern so the defaults are correctly recorded. There are probably other places that doing the wrong thing where we would not get the correct value.

This does keep a linked list of all the params that have ever had set_defualt called on them, so we should probably not do that too much.... I have not calculated how many this is, probably less than 50...

The goal of this is to make it very easy to spot which params have been changed, with this patch one can grab the defaults over ftp and do a param compare.

[IamPete1]

Look like this is just enough to push Hitec Airspeed over the edge, its only getting the AP_Param changes, could aggressively define them out on periph I guess....

[IamPete1]

Functionality now removed on Periph as much as possible to minimize flash cost. Also tested with params loaded from scripting, works as expected.

[IamPete1]

Compare to tridge@065159b

[IamPete1]

Added support for packed defaults via the param ftp format. This still has @SYS/defaults.parm in plain text. We could remove that once we get GCS support, or never add it in the first place...

The is some additional work to make sure the defaults are set correctly. We need to check everywhere using the param set method and add a new set_and_default in some cases as in tridge@065159b. I don't think we need to do every case, sometimes it is useful to see where a value has been automatically set but is not the default.

[tridge]

ArduPilot/MAVProxy#1077

[IamPete1]

Just tested with ArduPilot/MAVProxy#1077 seems to work as expected. mav.parm.defaults.txt I set also set the first three servo trims in defualts.parm and they show up.

[IamPete1]

@tridge I have implemented, I think, the protocol you suggested.

[IamPete1]

rebased to resolve conflict

[tridge]

great work!