Use x8 register for result struct address on aarch64

[tyfkda]

Unlike x64 or RISCV64, aarch64 uses %x8 (instead of first argument register) to pass the function result struct address:

sub.h:

typedef struct {
    long long x, y, z;
} FooStruct;

extern FooStruct return_struct(long long x, long long y, long long z);

sub.c:

#include "sub.h"

FooStruct return_struct(long long x, long long y, long long z) {
    FooStruct s = {.x = x, .y = y, .z = z};
    return s;
}

$ gcc -S -o- -O2 -fno-asynchronous-unwind-tables sub.c
        .section        __TEXT,__text,regular,pure_instructions
        .build_version macos, 14, 0     sdk_version 15, 2
        .globl  _return_struct                  ; -- Begin function return_struct
        .p2align        2
_return_struct:                         ; @return_struct
        .cfi_startproc
; %bb.0:
        stp     x0, x1, [x8]
        str     x2, [x8, #16]
        ret
        .cfi_endproc
                                        ; -- End function
.subsections_via_symbols

main.c:

#include "sub.h"

FooStruct return_struct(long long x, long long y, long long z) {
    FooStruct s = {.x = x, .y = y, .z = z};
    return s;
}

$ gcc -S -o- -O2 -fno-asynchronous-unwind-tables main.c
        .section        __TEXT,__text,regular,pure_instructions
        .build_version macos, 14, 0     sdk_version 15, 2
        .globl  _main                           ; -- Begin function main
        .p2align        2
_main:                                  ; @main
        .cfi_startproc
; %bb.0:
        sub     sp, sp, #48
        stp     x29, x30, [sp, #32]             ; 16-byte Folded Spill
        add     x29, sp, #32
        .cfi_def_cfa w29, 16
        .cfi_offset w30, -8
        .cfi_offset w29, -16
        add     x8, sp, #8                           ;########## FooStruct s
        mov     w0, #1                          ; =0x1
        mov     w1, #2                          ; =0x2
        mov     w2, #3                          ; =0x3
        bl      _return_struct
        ldr     w8, [sp, #8]
        ldr     w9, [sp, #16]
        ldr     w10, [sp, #24]
        add     w8, w9, w8
        add     w0, w8, w10
        ldp     x29, x30, [sp, #32]             ; 16-byte Folded Reload
        add     sp, sp, #48
        ret
        .cfi_endproc
                                        ; -- End function
.subsections_via_symbols