Ensure xPortGetFreeHeapSize reports DRAM

[mhightower83]

Create dedicated function for xPortGetFreeHeapSize() that only reports on DRAM.
NONOS SDK API system_get_free_heap_size() relies on xPortGetFreeHeapSize() for the free Heap size.

Possible breaking change for multiple Heap Sketches calling system_get_free_heap_size(); it will now always report free DRAM Heap size.

Update and export umm_free_heap_size_lw() to report the free Heap size of the current Heap.
Updated ESP.getFreeHeap() to use umm_free_heap_size_lw().

Updated build options to supply exported umm_free_heap_size_lw() via either UMM_STATS or UMM_INFO.

Improved build option support via the SketchName.ino.globals.h method for Heap options: UMM_INFO, UMM_INLINE_METRICS, UMM_STATS, UMM_STATS_FULL, UMM_BEST_FIT, and UMM_FIRST_FIT. While uncommon to change from the defaults, you can review umm_malloc_cfgport.h for more details, which may help reduce your Sketch's size in dire situations. Assuming you are willing to give up some functionality.
For debugging UMM_STATS_FULL can offer additional stats, like Heap low water mark (umm_free_heap_size_min()).

See umm_malloc/Notes.h entry Sep 26, 2022 for overview.

uncrustified umm_poison.c

Edited: revised to track PR changes

[mcspr]

Why not 'alias' for both variants?
Wouldn't there be less preprocessor if-def in both umm_info.c and umm_local.c if we have different functionality of stat retrieval?

Specifically, function name change in _info.c where we have it either be umm_free_heap_size or umm_free_heap_size_info. And the _local.c at the same time may or may not define it.

[mhightower83]

@mcspr Okay, I have made some changes I'll push them tomorrow after I review them again. I did some build option variations and found some issues. Another function needs to be broken out. Also, improved support for comment block build options.

[mcspr]

Ran a couple of tests switching build opts, everything works as described.

Only (minor?) issue is we still could make getHeapStats do nothing when umm info / stats become empty macro calls. This was already happening before the PR, though.

[mhightower83]

Only (minor?) issue is we still could make getHeapStats do nothing when umm info / stats become empty macro calls. This was already happening before the PR, though.

@mcspr what do you think it should do? The only info I could return would be free heap size; however,
I thought that it would be confusing to have a crippled function better to completely not work. If they were looking for free heap size they could use the other method. I suppose a build conditional with deprecated attributes might be a good way to go.

[mcspr]

Only (minor?) issue is we still could make getHeapStats do nothing when umm info / stats become empty macro calls. This was already happening before the PR, though.

@mcspr what do you think it should do? The only info I could return would be free heap size; however, I thought that it would be confusing to have a crippled function better to completely not work. If they were looking for free heap size they could use the other method. I suppose a build conditional with deprecated attributes might be a good way to go.

Fail on link, since we have a clear dependency? Empty macro inside UMM still makes sense to skip some actions, but external one probably doesn't.

If we don't use ESP.getHeapStats, we don't care whether anything related to stats is implemented.
If we do use it but mess up umm flags, linker won't find anything and show undefined reference.

[mcspr]

Something about that umm double include does not feel right... cfgport may come either first, or last. Kind of a weird dependency.
But, lgtm otherwise, if this is done with updates

[mhightower83]

When I make changes I migrate closer to upstream usage where umm_malloc_cfg.h is the upstream library version and all local port defines/changes go in umm_malloc_cfgport.h. Before, port changes were mixed in with umm_malloc_cfg.h.
It is weird to put umm_malloc_cfgport.h first in Arduino.h; however, it seems to have all the options defined (and normalized) for the port in one place which is what was needed for the build. Having Esp.h reference umm_malloc build options made things complicated.

I have more changes for aligning with umm_malloc upstream plus a few enhancements, I now need to work through those and see what I have broken (if any) with this shuffle.

I think I am done with this one. It was a lot more change than I wanted for a quick fix.