On Thu, Jun 05, 2025 at 03:55:15PM +0200, Georg-Johann Lay wrote:
> >register int *p1 asm ("r0") = …;
> >register int *p2 asm ("r1") = …;
> >int *result;
> >asm ("sysint" : "+r" (p1) : "0" (p1), "r" (p2));
> >result = p1;
>
> With "+r" (p1), the input operand "0" (p1) is not needed any more. Just:
>
> asm ("sysint" : "+r" (p1) : "r" (p2));
That's what I said. You do need to copy "p1" to "result", the code was
written to expect the result there, better keep the code correct even
when making it more elegant ;-)
Note that + is just syntactic sugar for = 0, you still have an input and
an output operand, but they are automatically tied.
> >(and by choosing better variable names, "r0" and "r1" perhaps, you don't
> >need to write a copy).
^^^
Segher
