Re: [PATCH 1/3] X86: Optimise fls(), ffs() and fls64()

[Geert Uytterhoeven <geert@linux-m68k.org>]

On Wed, Jan 13, 2010 at 20:39, David Howells <dhowells@redhat.com> wrote:
> diff --git a/include/asm-generic/bitops/fls64.h b/include/asm-generic/bitops/fls64.h
> index b097cf8..f81e04c 100644
> --- a/include/asm-generic/bitops/fls64.h
> +++ b/include/asm-generic/bitops/fls64.h
> @@ -18,16 +18,25 @@
>  static __always_inline int fls64(__u64 x)
>  {
>        __u32 h = x >> 32;
> -       if (h)
> -               return fls(h) + 32;
> -       return fls(x);
> +       int bitpos = -1;
> +
> +       asm("bsrl       %1,%0   \n"
> +           "subl       %2,%0   \n"
> +           "bsrl       %3,%0   \n"
> +           : "+r" (bitpos)
> +           : "rm" (x), "i"(32), "rm" (h));
> +
> +       return bitpos + 33;
>  }
>  #elif BITS_PER_LONG == 64
>  static __always_inline int fls64(__u64 x)
>  {
> -       if (x == 0)
> -               return 0;
> -       return __fls(x) + 1;
> +       long bitpos = -1;
> +
> +       asm("bsrq %1,%0"
> +           : "+r" (bitpos)
> +           : "rm" (x));
> +       return bitpos + 1;
>  }
>  #else
>  #error BITS_PER_LONG not 32 or 64

Why do you put x86 assembler in asm-generic? Those files are supposed
to be platform-agnostic.

Gr{oetje,eeting}s,

						Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
							    -- Linus Torvalds