Re: [PATCH] Multiple & vs. && and | vs. || bugs in 2.4.20

[Werner Almesberger]

Ulrich Drepper wrote:
> > -	if (!urb->status & !acm->throttle)  {
> > +	if (!urb->status && !acm->throttle)  {
[...]
> Have you really looked at detail at all these cases?  The problem is
> that you have made the code less efficient on some platforms.  The use
> of && requires shortcut evaluation.  I.e., the compiler is forced to not

While I agree with your observation in general, this is actually
something the compiler should be able to figure out by itself:

 - there's only a side-effect if acm is NULL
 - in ACM_READY, we've already tested acm for NULL, and subsequently
   de-referenced it
 - acm is a local variable, and not aliased, so the dbg() can't
   change it

So, given the negations, || and | are equivalent in this case, and
whether a jump, conditional execution, a bit operation, or something
else yields better code is compiler, machine, and context specific.

- Werner

-- 
  _________________________________________________________________________
 / Werner Almesberger, Buenos Aires, Argentina         [e-mail] /
/_http://www.almesberger.net/____________________________________________/