Yes, with asCALL_GENERIC it is pure C++ so you won't face any issues with native ABIs.
Has anyone compiled angelscript on aarch64?
I guess this applies also to M1 ARM64 native calling which is missing, ABI is specified here: https://developer.apple.com/documentation/xcode/writing-arm64-code-for-apple-platforms
But I'm definitely rusty with ASM and I don't anything of the underlying AngelScript implementation so I guess it's an hard task for me. According to ABI there are not much differences compared to standard ARM64 but I don't know how stable the current Linux implementation is.
Yes, it applies to M1 ARM64 native calling conventions too.
The Linux implementation is fully working and stable. It is definitely recommended to use it as the base for either M1 or Android implementations.
Normally it doesn't require any previous knowledge of ASM to the implement the support. Especially not when there is already an existing implementation as it will be mostly tweaks of the current implementation. Use the regression test suite, to identify the function calls that currently do not work and then use disassembly to understand the difference in how the C++ compiler generates the code and how the as_callfunc_arm64_gcc.S does it.
If you have the time available it probably won't take more than a week to get it working.
This topic is closed to new replies.
Advertisement
Popular Topics
Advertisement
