Re: [Lightning] [Fwd: Re: lightning for x86-64???]

[Paolo Bonzini]

> You probably saw the "README" file in mzscheme/src/lightning. The
> addition of "short jump" mode versus "big jump" mode is probably a bad
> idea; it's probably better to have different lightning opcodes.
Actually I like the idea of a sort of PIC mode.
> Also,
> the room for three locals left by PPC jit_prolog() is completely
> MzScheme-specific.
>   
Yes, I'm only caring about the x86_64 merge for now.
> Some x86_64 notes at the top of MzScheme's "jit.c" actually belong in
> the "README" file, now that I think about it:
>
>   3) Immediate operands must be 32-bit values on x86_64, except with
>      jit_movi, jit_sti, jit_ld, jit_bXi, jit_calli, and jit_finishi.
>   
Ouch.  Need something like the RISC jit_big... later though, or we can 
just document it as a current restriction -- it's the first 64-bit port 
anyway.
>   4) Function calls are limited to 3 arguments (i.e., jit_prepare()
>      must never be called with a number greater than 3). This limit
>      is related to the way the x86_64 port shuffles arguments into
>      temporary registers.
>   
We'll fix it.  Do you have a pointer to a quick description of the ABI?  
Informatiion that is enough should be: ordering of registers to pass the 
parameters, caller/callee saved registers.
>   5) On x86_64, arguments are delivered in JIT_V2, JIT_V3, and JIT_R2,
>      in that order. So don't set JIT_R2 before getting the third
>      argument, etc.
>   
Hmm... that's bad...  we should move JIT_Vn/JIT_Rn on x86_64.

Paolo