This is correct. You can forego all aspects of your platform's calling conventions if you so desire. Omitting setting up stack frames is a really simple optimisation that can be done in hand-written assembly. You've already touched on what the stakes of doing such things are.

Compilers are able to do the same since a few years.

Yes, but they rarely take a whole-program approach. Reducing a complicated sequence of function calls to a few goto’s is not an easy task. Probably NP-hard?

Doing it on a whole program basis is also unlikely to give much benefit. Function calls are extremely fast, they only show their overhead in tight loops.

Yeah, reviewing the repository there’s a lot of high level language code now, so that probably removes most of the benefit of assembly. You’re still calling into the OS a lot.

Some of the other low level systems (like the Mac) had a trap system that wasn’t so far away in cycle count from user code. But in these days of needing 10,000 cycles to bridge a system call it’s best to do whatever you can to avoid calling the OS.

Function inlining can't avoid system calls

Exactly my point