Re: [PATCH] document optimizing macro for translating PROT_ to VM_ bits

[Thomas Schlichter <schlicht@uni-mannheim.de>]

[-- Attachment #1: Type: text/plain, Size: 1391 bytes --]

Hi,

here just my 2 cents...

On Tuesday 30 September 2003 11:24, Muli Ben-Yehuda wrote:

  ~~ snip ~~

> /*
>  * assert that only a single bit is on in 'bit'
>  */
> #define assert_single_bit(bit) do { \
>         if (__builtin_constant_p(bit)) {			        \
> 		if ((bit & (bit -1)))					\
> 			__assert_single_bit_failed_dont_exist();	\
> 	}  else								\
> 		BUG_ON(!(bit & (bit - 1)));				\
> 	} while(0)

In the BUG_ON statement the "!" looks wrong to me...

> /*
>  * Optimisation function.  It is equivalent to:
>  *      (x & bit1) ? bit2 : 0
>  * but this version is faster.
>  * ("bit1" and "bit2" must be single bits).
>  */
> static inline unsigned long
> inline_calc_vm_trans(unsigned long x, unsigned long bit1, unsigned long
> bit2) {
> 	assert_single_bit(bit1);
> 	assert_single_bit(bit2);
>
> 	return ((bit1) <= (bit2) ? ((x) & (bit1)) * ((bit2) / (bit1))
>
> 		: ((x) & (bit1)) / ((bit1) / (bit2)));
>
> }

Why don't we do:

static inline unsigned long calc_vm_trans(const unsigned long x,
		const unsigned long bit1, const unsigned long bit2) {
	assert_single_bit(bit1);
	assert_single_bit(bit2);

	/* Optimisation function */
	if (__builtin_constant_p(bit1) && __builtin_constant_p(bit2)) {
		return ((bit1) <= (bit2) ? ((x) & (bit1)) * ((bit2) / (bit1))
			: ((x) & (bit1)) / ((bit1) / (bit2)));
	}

	return (x & bit1) ? bit2 : 0;
}

Best regards
   Thomas Schlichter

[-- Attachment #2: signature --]
[-- Type: application/pgp-signature, Size: 189 bytes --]