[PATCH] ipsum_calc_block: Optimize size and speed

[Joakim Tjernlund]

Joakim Tjernlund <Joakim.Tjernlund at transmode.se> wrote on 2010/04/23 10:02:54:
>
> This is a much simpler and efficent impl. of the IP checksum.
> It is a dry port from Quagga and I have not tested it.
> ---
>  lib/checksum.c |   44 +++++++++++---------------------------------
>  1 files changed, 11 insertions(+), 33 deletions(-)
>
> diff --git a/lib/checksum.c b/lib/checksum.c
> index 33cb386..b836bdb 100644
> --- a/lib/checksum.c
> +++ b/lib/checksum.c
> @@ -15,25 +15,10 @@
>  #include "nest/bird.h"
>  #include "checksum.h"
>
> -static u16            /* One-complement addition */
> -add16(u16 sum, u16 x)
> -{
> -  u16 z = sum + x;
> -  return z + (z < sum);
> -}
> -
> -static u32
> -add32(u32 sum, u32 x)
> -{
> -  u32 z = sum + x;
> -  return z + (z < sum);
> -}
> -
>  static u16
>  ipsum_calc_block(u16 *x, unsigned len, u16 sum)
>  {
> -  int rest;
> -  u32 tmp, *xx;
> +  u32 tmp;
>
>    /*
>     *  A few simple facts about the IP checksum (see RFC 1071 for detailed
> @@ -51,23 +36,16 @@ ipsum_calc_block(u16 *x, unsigned len, u16 sum)
>    if (!len)
>      return sum;
>    len >>= 1;
> -  if ((unsigned long) x & 2)      /* Align to 32-bit boundary */
> -    {
> -      sum = add16(sum, *x++);
> -      len--;
> -    }
> -  rest = len & 1;
> -  len >>= 1;
> -  tmp = 0;
> -  xx = (u32 *) x;
> -  while (len)
> -    {
> -      tmp = add32(tmp, *xx++);
> -      len--;
> -    }
> -  sum = add16(sum, add16(tmp & 0xffff, tmp >> 16U));
> -  if (rest)
> -    sum = add16(sum, *(u16 *) xx);
> +  tmp = sum;
> +  for(x--; len; --len)
> +      tmp += *++x;
> +  /*
> +   * Add back carry outs from top 16 bits to low 16 bits.
> +   */
> +  tmp  = (tmp >> 16) + (tmp & 0xffff);    /* add high-16 to low-16 */
> +  tmp += (tmp >> 16);                     /* add carry */
> +  sum = ~tmp;          /* ones-complement, then truncate to 16 bits */

Possibly the "sum = ~tmp;" line should be removed. Looks like it is
performed by the calling routines.

    Jocke