[deliverable/linux.git] / arch / x86 / include / asm / word-at-a-time.h

#ifndef _ASM_WORD_AT_A_TIME_H
#define _ASM_WORD_AT_A_TIME_H

#include <linux/kernel.h>

/*
 * This is largely generic for little-endian machines, but the
 * optimal byte mask counting is probably going to be something
 * that is architecture-specific. If you have a reliably fast
 * bit count instruction, that might be better than the multiply
 * and shift, for example.
 */
struct word_at_a_time {
	const unsigned long one_bits, high_bits;
};

#define WORD_AT_A_TIME_CONSTANTS { REPEAT_BYTE(0x01), REPEAT_BYTE(0x80) }

#ifdef CONFIG_64BIT

/*
 * Jan Achrenius on G+: microoptimized version of
 * the simpler "(mask & ONEBYTES) * ONEBYTES >> 56"
 * that works for the bytemasks without having to
 * mask them first.
 */
static inline long count_masked_bytes(unsigned long mask)
{
	return mask*0x0001020304050608ul >> 56;
}

#else	/* 32-bit case */

/* Carl Chatfield / Jan Achrenius G+ version for 32-bit */
static inline long count_masked_bytes(long mask)
{
	/* (000000 0000ff 00ffff ffffff) -> ( 1 1 2 3 ) */
	long a = (0x0ff0001+mask) >> 23;
	/* Fix the 1 for 00 case */
	return a & mask;
}

#endif

/* Return nonzero if it has a zero */
static inline unsigned long has_zero(unsigned long a, unsigned long *bits, const struct word_at_a_time *c)
{
	unsigned long mask = ((a - c->one_bits) & ~a) & c->high_bits;
	*bits = mask;
	return mask;
}

static inline unsigned long prep_zero_mask(unsigned long a, unsigned long bits, const struct word_at_a_time *c)
{
	return bits;
}

static inline unsigned long create_zero_mask(unsigned long bits)
{
	bits = (bits - 1) & ~bits;
	return bits >> 7;
}

/* The mask we created is directly usable as a bytemask */
#define zero_bytemask(mask) (mask)

static inline unsigned long find_zero(unsigned long mask)
{
	return count_masked_bytes(mask);
}

/*
 * Load an unaligned word from kernel space.
 *
 * In the (very unlikely) case of the word being a page-crosser
 * and the next page not being mapped, take the exception and
 * return zeroes in the non-existing part.
 */
static inline unsigned long load_unaligned_zeropad(const void *addr)
{
	unsigned long ret, dummy;

	asm(
		"1:\tmov %2,%0\n"
		"2:\n"
		".section .fixup,\"ax\"\n"
		"3:\t"
		"lea %2,%1\n\t"
		"and %3,%1\n\t"
		"mov (%1),%0\n\t"
		"leal %2,%%ecx\n\t"
		"andl %4,%%ecx\n\t"
		"shll $3,%%ecx\n\t"
		"shr %%cl,%0\n\t"
		"jmp 2b\n"
		".previous\n"
		_ASM_EXTABLE(1b, 3b)
		:"=&r" (ret),"=&c" (dummy)
		:"m" (*(unsigned long *)addr),
		 "i" (-sizeof(unsigned long)),
		 "i" (sizeof(unsigned long)-1));
	return ret;
}

#endif /* _ASM_WORD_AT_A_TIME_H */
Commit	Line	Data
f68e556e LT	1	#ifndef _ASM_WORD_AT_A_TIME_H
	2	#define _ASM_WORD_AT_A_TIME_H
	3
44696908 DM	4	#include <linux/kernel.h>
44696908 DM	5
f68e556e LT	6	/*
	7	* This is largely generic for little-endian machines, but the
	8	* optimal byte mask counting is probably going to be something
	9	* that is architecture-specific. If you have a reliably fast
	10	* bit count instruction, that might be better than the multiply
	11	* and shift, for example.
	12	*/
36126f8f LT	13	struct word_at_a_time {
	14	const unsigned long one_bits, high_bits;
	15	};
	16
	17	#define WORD_AT_A_TIME_CONSTANTS { REPEAT_BYTE(0x01), REPEAT_BYTE(0x80) }
f68e556e LT	18
	19	#ifdef CONFIG_64BIT
	20
	21	/*
	22	* Jan Achrenius on G+: microoptimized version of
	23	* the simpler "(mask & ONEBYTES) * ONEBYTES >> 56"
	24	* that works for the bytemasks without having to
	25	* mask them first.
	26	*/
	27	static inline long count_masked_bytes(unsigned long mask)
	28	{
	29	return mask*0x0001020304050608ul >> 56;
	30	}
	31
	32	#else /* 32-bit case */
	33
	34	/* Carl Chatfield / Jan Achrenius G+ version for 32-bit */
	35	static inline long count_masked_bytes(long mask)
	36	{
	37	/* (000000 0000ff 00ffff ffffff) -> ( 1 1 2 3 ) */
	38	long a = (0x0ff0001+mask) >> 23;
	39	/* Fix the 1 for 00 case */
	40	return a & mask;
	41	}
	42
	43	#endif
	44
36126f8f LT	45	/* Return nonzero if it has a zero */
	46	static inline unsigned long has_zero(unsigned long a, unsigned long bits, const struct word_at_a_time c)
	47	{
	48	unsigned long mask = ((a - c->one_bits) & ~a) & c->high_bits;
	49	*bits = mask;
	50	return mask;
	51	}
	52
	53	static inline unsigned long prep_zero_mask(unsigned long a, unsigned long bits, const struct word_at_a_time *c)
	54	{
	55	return bits;
	56	}
	57
	58	static inline unsigned long create_zero_mask(unsigned long bits)
	59	{
	60	bits = (bits - 1) & ~bits;
	61	return bits >> 7;
	62	}
	63
	64	/* The mask we created is directly usable as a bytemask */
	65	#define zero_bytemask(mask) (mask)
	66
	67	static inline unsigned long find_zero(unsigned long mask)
f68e556e	68	{
36126f8f	69	return count_masked_bytes(mask);
f68e556e LT	70	}
f68e556e LT	71
e419b4cc LT	72	/*
	73	* Load an unaligned word from kernel space.
	74	*
	75	* In the (very unlikely) case of the word being a page-crosser
	76	* and the next page not being mapped, take the exception and
	77	* return zeroes in the non-existing part.
	78	*/
	79	static inline unsigned long load_unaligned_zeropad(const void *addr)
	80	{
	81	unsigned long ret, dummy;
	82
	83	asm(
	84	"1:\tmov %2,%0\n"
	85	"2:\n"
	86	".section .fixup,\"ax\"\n"
	87	"3:\t"
	88	"lea %2,%1\n\t"
	89	"and %3,%1\n\t"
	90	"mov (%1),%0\n\t"
	91	"leal %2,%%ecx\n\t"
	92	"andl %4,%%ecx\n\t"
	93	"shll $3,%%ecx\n\t"
	94	"shr %%cl,%0\n\t"
	95	"jmp 2b\n"
	96	".previous\n"
	97	_ASM_EXTABLE(1b, 3b)
	98	:"=&r" (ret),"=&c" (dummy)
	99	:"m" ((unsigned long )addr),
	100	"i" (-sizeof(unsigned long)),
	101	"i" (sizeof(unsigned long)-1));
	102	return ret;
	103	}
	104
f68e556e	105	#endif /* _ASM_WORD_AT_A_TIME_H */