Re: [RFC] div64_64 support

[Jan Engelhardt <jengelh@linux01.gwdg.de>]

On Feb 26 2007 13:28, Stephen Hemminger wrote:
>> 
>> ./arch/arm26/lib/udivdi3.c
>> ./arch/sh/lib/udivdi3.c
>> ./arch/sparc/lib/udivdi3.S
>> 
>> should not this be consolidated too?
>
>Hmm. Those are the GCC internal versions, that are picked up but
>doing divide in place. Do we want to allow general 64 bit in kernel to
>be easily used? It could cause sloppy slow code, but it would look
>cleaner.

Then our reviewers should catch it, and if not, the janitors will
(/me winks at R.P.J.Day and trivial@).

>@@ -134,7 +112,7 @@
> 	 */
> 	do {
> 		x1 = x;
>-		x = (2 * x + (uint32_t) div64_64(a, x*x)) / 3;
>+		x = (2 * x + (u32) (a / x*x)) / 3;

Eye see a bug.

Previously there was div64_64(a, x*x) which is equivalent to
(a)/(x*x), or just: a/(x^2). But now you do a/x*x, which is
equivalent to a*x/x (in the domain of real numbers). Furthermore,
a/x*x is a-(a%x), which does not even remotely match a/(x^2).

Please keep the math intact, thank you ;-)



Jan
--