Re: -fasm=intel not always suppported

[Mike Stump <mrs at apple dot com>]

On Nov 15, 2005, at 12:40 PM, Andrew Pinski wrote:
> Also x86 inline-asm is harder than what you put it.

?  You lost me here.

> It is not a matter of geting the constraints right but a matter of  
> saying which register gets clobbered and which register gets placed  
> where.

Yes.  This is trivial and most of the work has already been done.

> An example of this is rdtsc which takes no operands but places its  
> result in
> eax:edx, how would you write that in MS style asm?

Surely there are tons of ways, all of them work, I might do it this way:

long long foo() {
  union {
    struct {
      int h;
      int l;
    } i;
    long long ll;
  } v;

  asm {
    rdtsc
    mov v.i.h, edx
    mov v.i.l, eax
  }
  return v.ll;
}

but don't flame me for it if you like another spelling.

> You would have to use an extra mov instruction like

Yup, that works too (s/ebx/edx/ of course).

> what happens if you write it as:
> long long f(void)
> {
>   asm { rdtsc }
>   return 0;
> }

This has the same answer as if you had written:
    asm ("rdtsc" : : "eax", "edx");
    return 0;

Seems confused to me to write that, but, one can.  Hope I got the  
spelling close enough for you to know what I meant.  I still find gcc  
asm syntax hard to write at times.

> Does your code mark eax and ebx as being clobbered

Yes, of course.