Doxygen for automatically generated headers in TMS570

[Tan Gemicioglu]

The header files in the TMS570 BSP are automatically generated from .json files. I've added Doxygen headers to 20 of the files but Joel was unsure if it should be added since the files are auto-generated.

[Tan Gemicioglu]

Since Trac isn't letting me attach the patch, I've put it on Dropbox: ​https://www.dropbox.com/s/k88jl5bz17853g4/0001-Added-doxygen-headers-to-20-files-in-tms570.patch

[Pavel Pisa]

Headers would be overwritten when files they are regenerated.
Manual editing is time consuming as well. So we need to add this to the Python script.

​https://github.com/AoLaD/rtems-tms570-utils/blob/master/headers/python/make_header.py

If update is committed to mainline then it can be used as reminder for future script change. We compare what changed after each script run against previous git state. But I would leave that for future headers update.

On the other hand, there are pending massive additions to TMS570 BSP code which are required for RTEMS self contained BSP startup. To have some idea, look at tms570-bsp-with-hwinit branch. The branch includes startup alternative with integrated HalCoGen? code and initial work on minimal RTEMS better matching solution.

​https://github.com/AoLaD/rtems/commits/tms570-bsp-with-hwinit

I have even done some proof of feasibility to use same PinMux? implementation for tms570lc4357. We communicate about requirements for support of future Hercules family chips with Ti expert and he expressed that he likes our approach taken for RTEMS and considers it better than HalCoGen? taken one at least in PinMux? area.

So we are far form full chip support and changes on RTEMS master which are not coordinated with WIP branches, would conflict during integration.