Clang + ARM64: how to pass X8 as an argument? [duplicate]

In the ARM64 architecture, if the size of the result exceeds 16 bytes, it is transmitted in the X8 register. I need to call the function in such a way as to pass this register.

typedef struct{char a[16 + 1];} retptr;
typedef retptr (*Func)(void* x0, void* x1, void* x2, void* x3);

void call(void* x0, void* x1, void* x2, void* x3, void* x8, void* address)
{
    retptr* r = (retptr*)x8;
    *r = ((Func)(address))(x0, x1, x2, x3);
}

The problem is that the value is not passed to the register - a pointer to the stack is passed, and then the data is copied. https://godbolt.org/z/D-zAnt

call: // @call
  sub sp, sp, #64 // =64
  str x19, [sp, #32] // 8-byte Folded Spill
  stp x29, x30, [sp, #48] // 16-byte Folded Spill
  add x29, sp, #48 // =48
  add x8, sp, #8 // =8
  mov x19, x4
  blr x5
  ldrb w8, [sp, #24]
  ldur q0, [sp, #8]
  strb w8, [x19, #16]
  str q0, [x19]
  ldp x29, x30, [sp, #48] // 16-byte Folded Reload
  ldr x19, [sp, #32] // 8-byte Folded Reload
  add sp, sp, #64 // =64
  ret