[linux-2.6-block.git] / arch / x86 / include / asm / word-at-a-time.h

/* SPDX-License-Identifier: GPL-2.0 */
#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
b2441318	1	/* SPDX-License-Identifier: GPL-2.0 */
f68e556e LT	2	#ifndef _ASM_WORD_AT_A_TIME_H
	3	#define _ASM_WORD_AT_A_TIME_H
	4
44696908 DM	5	#include <linux/kernel.h>
44696908 DM	6
f68e556e LT	7	/*
	8	* This is largely generic for little-endian machines, but the
	9	* optimal byte mask counting is probably going to be something
	10	* that is architecture-specific. If you have a reliably fast
	11	* bit count instruction, that might be better than the multiply
	12	* and shift, for example.
	13	*/
36126f8f LT	14	struct word_at_a_time {
	15	const unsigned long one_bits, high_bits;
	16	};
	17
	18	#define WORD_AT_A_TIME_CONSTANTS { REPEAT_BYTE(0x01), REPEAT_BYTE(0x80) }
f68e556e LT	19
	20	#ifdef CONFIG_64BIT
	21
	22	/*
	23	* Jan Achrenius on G+: microoptimized version of
	24	* the simpler "(mask & ONEBYTES) * ONEBYTES >> 56"
	25	* that works for the bytemasks without having to
	26	* mask them first.
	27	*/
	28	static inline long count_masked_bytes(unsigned long mask)
	29	{
	30	return mask*0x0001020304050608ul >> 56;
	31	}
	32
	33	#else /* 32-bit case */
	34
	35	/* Carl Chatfield / Jan Achrenius G+ version for 32-bit */
	36	static inline long count_masked_bytes(long mask)
	37	{
	38	/* (000000 0000ff 00ffff ffffff) -> ( 1 1 2 3 ) */
	39	long a = (0x0ff0001+mask) >> 23;
	40	/* Fix the 1 for 00 case */
	41	return a & mask;
	42	}
	43
	44	#endif
	45
36126f8f LT	46	/* Return nonzero if it has a zero */
	47	static inline unsigned long has_zero(unsigned long a, unsigned long bits, const struct word_at_a_time c)
	48	{
	49	unsigned long mask = ((a - c->one_bits) & ~a) & c->high_bits;
	50	*bits = mask;
	51	return mask;
	52	}
	53
	54	static inline unsigned long prep_zero_mask(unsigned long a, unsigned long bits, const struct word_at_a_time *c)
	55	{
	56	return bits;
	57	}
	58
	59	static inline unsigned long create_zero_mask(unsigned long bits)
	60	{
	61	bits = (bits - 1) & ~bits;
	62	return bits >> 7;
	63	}
	64
	65	/* The mask we created is directly usable as a bytemask */
	66	#define zero_bytemask(mask) (mask)
	67
	68	static inline unsigned long find_zero(unsigned long mask)
f68e556e	69	{
36126f8f	70	return count_masked_bytes(mask);
f68e556e LT	71	}
f68e556e LT	72
e419b4cc LT	73	/*
	74	* Load an unaligned word from kernel space.
	75	*
	76	* In the (very unlikely) case of the word being a page-crosser
	77	* and the next page not being mapped, take the exception and
	78	* return zeroes in the non-existing part.
	79	*/
	80	static inline unsigned long load_unaligned_zeropad(const void *addr)
	81	{
	82	unsigned long ret, dummy;
	83
	84	asm(
	85	"1:\tmov %2,%0\n"
	86	"2:\n"
	87	".section .fixup,\"ax\"\n"
	88	"3:\t"
	89	"lea %2,%1\n\t"
	90	"and %3,%1\n\t"
	91	"mov (%1),%0\n\t"
	92	"leal %2,%%ecx\n\t"
	93	"andl %4,%%ecx\n\t"
	94	"shll $3,%%ecx\n\t"
	95	"shr %%cl,%0\n\t"
	96	"jmp 2b\n"
	97	".previous\n"
	98	_ASM_EXTABLE(1b, 3b)
	99	:"=&r" (ret),"=&c" (dummy)
	100	:"m" ((unsigned long )addr),
	101	"i" (-sizeof(unsigned long)),
	102	"i" (sizeof(unsigned long)-1));
	103	return ret;
	104	}
	105
f68e556e	106	#endif /* _ASM_WORD_AT_A_TIME_H */