Fix: Buffer utilities codegen

[dragonmux]

Detailed description

In this PR we address a codegen/optimisation issue that has been identified in the buffer utilities functions. These functions include the read_{l,b}e{2,4,8}() routines which do an endian-aware read of a buffer to (re)construct an integer.

The way these functions were expressed is somewhat naïve and results in extremely poor codegen (a sled of loads, shifts and adds) due to the need to deal with an offset into the buffer. It was quickly identified, however, that moving the data to a function-internal buffer with memcpy() and then doing the endian-dependant read produced the correct code output - one load, and depending on host endian, one swap.

This greatly improves the performance of these functions and their call footprint as they get inlined where they're called.

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 (see Building the firmware)
• It builds as BMDA (see Building the BMDA)
• 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

[esden]

LGTM