Consider adding LBT_86 support to box64 about box64

[ptitSeb]

That can be implemented only in the dynarec. But loongson dynarec is not done yet.

[tempdragon]

I am new to this project so this question may sound stupid. The src/emu seems to be related to similar functions. Why is modification to that part discouraged here?

[ptitSeb]

I am new to this project so this question may sound stupid. The src/emu seems to be related to similar functions. Why is modification to that part discouraged here?

The src/emu is generic, and is used on all architecture. There is no custom code per arch. This is also slow, being an interpreter. That feature would be used on the dynarec, but this (the dynarc), is not really started yet (actually, I have a very old POC of a dynarec in a private repo, but I still need to merge the code in current box64, and it was only an embryo of it anyway).

So, once a Loogarch dynarec is there, sure, this feature will be added. But on the interpreter, it make little sense to do it (because using that feature needs low level controls of operations, wich the interpreter does not provide).

[tempdragon]

If I want to contribute to this part, should I wait for the "basic" version(may be the updated one of your POC) to finish and get published or should I try to write one based on the RISC-V version?

[ptitSeb]

If I want to contribute to this part, should I wait for the "basic" version(may be the updated one of your POC) to finish and get published or should I try to write one based on the RISC-V version?

It would be better to wait for the basic part yes. The infrastructure can be a bit tricky to setup.

[ksco]

I’m also interested in a loongarch backend for box64, and I think great progress will be made this year.

[tempdragon]

It may be a good idea to publish that POC to allow others to contribute, and potentially accelerate the development. I am actually working on a scratch adapted from RV64(using search & replace), but further research should still be necessary.

[ptitSeb]

It may be a good idea to publish that POC to allow others to contribute, and potentially accelerate the development. I am actually working on a scratch adapted from RV64(using search & replace), but further research should still be necessary.

No, it's separate codebase, based on a older box64. The code needs to be merged with current box64, no need to multiply the codebase, and I prefer to merge while there are still not much to merge (it's basiacaly the infrastructure a few basic opcodes)