RE: Re: [PATCHv5] atomic: add *_dec_not_zero

["David Laight" <David.Laight@ACULAB.COM>]

Looking at this:

> #ifndef atomic_inc_unless_negative
> static inline int atomic_inc_unless_negative(atomic_t *p)
> {
>         int v, v1;
>         for (v = 0; v >= 0; v = v1) {
>                 v1 = atomic_cmpxchg(p, v, v + 1);
>                 if (likely(v1 == v))
>                         return 1;
>         }
>         return 0;
> }
> #endif

why is it optimised for '*p' being zero??
I'd have though the initial assignment to 'v' should
be made by reading '*p' without any memory barriers (etc).

        David