Fix: Factor out integer-only stdio into a common header

[ALTracer]

Detailed description

• The previous PR Fix: Decouple headers general.h/platform.h #1530 reduced visibility of "newlib integer-only stdio" redefinitions to only those translation units which #include platform.h This lead to flash size increase for 7 platforms (all not using newlib-nano) by 37 KiB because of linker pulling float/double capable stdio implementations (and softfp routines for double etc).
• This PR ties these redefinitions back to platform_support.h visible via general.h to every platform except hosted.

Your checklist for this pull request

• I've read the Code of Conduct
• I've read the guidelines for contributing to this repository
• It builds for hardware native (make PROBE_HOST=native)
• It builds as BMDA (make PROBE_HOST=hosted)
• I've tested it to the best of my ability
• My commit messages provide a useful short description of what the commits do

Closing issues

Testing

To reproduce/verify, make all_platforms and compare binary sizes of files under src/artifacts/v1.9.0-... for relevant revisions.
Also Puncover to check for functions pulled from libc.
I don't know yet how to easily automatically perform checks for ENABLE_DEBUG=1 builds without encountering build failures or invoking make per-platform manually, but these binaries are Big anyways.

Miscellaneous

If there is a better header to place this in than platform_support.h, then I'll edit this branch.
I could also edit the copyright string, author and filename for "stdio_newlib.h" as I'm only reorganizing the header contents and not actually writing implementation code in this PR. The touched platform.h headers contain at least three authors in "Written by" strings.

[dragonmux]

LGTM. We were contemplating if "stdio_newlib.h" is necessarily the best name for this new header - perhaps "newlib_stdio_defs.h" or something might be better.. but as this fixes both a huge amount of duplication and a bug, we're going to get this merged so main gets unbroken.