On Thu, 2017-07-20 at 13:24 +0200, Peter Zijlstra wrote:
[]
> +     m = 1UL << ((fls(x) + 1) & ~1UL);

maybe 

#if BITS_PER_LONG == 64
        m = 1UL << ((fls64(x) + 1) & ~1UL);
#else
        m = 1UL << ((fls(x) + 1) & ~1UL);
#endif

Reply via email to