Re: + kernelh-make-abs-work-with-64-bit-types.patch added to -mm tree

[Andrew Morton <akpm@linux-foundation.org>]

On Wed, 23 Sep 2015 16:44:29 +0300 Alexey Dobriyan <adobriyan@gmail.com> wrote:

> On Fri, Sep 18, 2015 at 12:36 AM,  <akpm@linux-foundation.org> wrote:
> 
> 
> > -#define abs(x) ({                                              \
> > -               long ret;                                       \
> > -               if (sizeof(x) == sizeof(long)) {                \
> > -                       long __x = (x);                         \
> > -                       ret = (__x < 0) ? -__x : __x;           \
> > -               } else {                                        \
> > -                       int __x = (x);                          \
> > -                       ret = (__x < 0) ? -__x : __x;           \
> > -               }                                               \
> > -               ret;                                            \
> > -       })
> > +#define abs(x) __builtin_choose_expr(sizeof(x) == sizeof(s64), ({      \
> > +               s64 __x = (x);                                          \
> > +               (__x < 0) ? -__x : __x;                                 \
> > +       }), ({                                                          \
> > +               long ret;                                               \
> > +               if (sizeof(x) == sizeof(long)) {                        \
> > +                       long __x = (x);                                 \
> > +                       ret = (__x < 0) ? -__x : __x;                   \
> > +               } else {                                                \
> > +                       int __x = (x);                                  \
> > +                       ret = (__x < 0) ? -__x : __x;                   \
> > +               }                                                       \
> > +               ret;                                                    \
> > +       }))
> 
> 1. "char" should be banned, because it's signedness is not well defined.
> 2. there is unnecessary expansion to long.
> 
> I've sent kabs() before which didn't go in because it didn't work for
> INT_MAX et al
> (don't worry, this abs() doens't as well) but it is nicer that this
> version in other aspects
> (hopefully).
> 
> [PATCH v2] Add kabs()
> http://marc.info/?l=linux-kernel&m=133518745522740&w=4

Looks nice.  Care to send along a patch which updates the abs() in
linux-next?

> As for INT_MIN, it pretty impossible to make compiler to not generate a branch
> despite generating correct result without a branch.