What is the process to port system software or languages to Apple's M1?

Question

Hi there,I searched around but could not find any documentation how to port software to Apple M1 systems.For example: How did Golang got ported? As far as I know Golang is implemented in Golang and just until some months ago it could not create M1 executables. If Golang would be language implemented in C or C++ I could somewhat guess that you need "just" run the sources through the Apple M1 clang/llvm compiler and patch the errors/warnings until it works.Or another example: How do they port Linux to M1? I mean you need binutils-glibc-gcc to get Linux and other neccessary tools like grub running. Does binutils have a port to M1? Does gcc create M1 executables? Or does this project not only port the kernel but also the toolchain?On one hand I look forward to the possibility to start one of the very long discussions you can find in the HN community - these are fun to read -. On the other hand I would also be thankful to get a link to a document I was not able to find.

wmf · Accepted Answer

M1 uses the ARM64[1] instruction set and a lot of compilers started working on ARM64 years ago so in many cases compilers were already ready before M1 was announced. Creating a macOS/ARM64 toolchain is mostly a combination of existing features from Linux/ARM64 and macOS/x86-64. For example, Linux uses ELF but macOS needs a Mach-O linker but most languages already had Mach-O support anyway. Once you have the toolchain it's a matter of recompiling and fixing any non-portable assumptions in the code. Fortunately x86-64 and ARM64 are both 64-bit little-endian so porting is fairly easy but the memory consistency models are different which can expose bugs in poorly-written code.[1] The official name is something else but it's dumb