2. Add flag INTERNAL_LIBM to mpconfigboard.mk files to signal using lib/libm.
3. Fix floating-point warnings in lib/libm
4. Turn off -finline-limit, no longer needed due to 1.
5. Add explicit flags for SPI_FLASH_FILESYSTEM and INTERNAL_FLASH_FILESYSTEM
to mpconfigboard.mk files, so they don't need to know the name of the .c file
that implements those. Makefile takes care of checking the flags and including
the right files.
gcc 6.1.1 warns when indentation is misleading, and in this case the
formatting of the code really is misleading. So adjust the formatting
to be clear of the meaning of the code.
acoshf, asinhf, atanhf were added from musl. mathsincos.c was
split up into its original, separate files (from newlibe-nano-2).
tan was added.
All of the important missing float functions are now implemented,
and pyboard now passes tests/float/math_fun.py (finally!).
