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

#ifndef _ASM_WORD_AT_A_TIME_H
#define _ASM_WORD_AT_A_TIME_H

/*
 * Word-at-a-time interfaces for PowerPC.
 */

#include <linux/kernel.h>
#include <asm/asm-compat.h>

#ifdef __BIG_ENDIAN__

struct word_at_a_time {
	const unsigned long high_bits, low_bits;
};

#define WORD_AT_A_TIME_CONSTANTS { REPEAT_BYTE(0xfe) + 1, REPEAT_BYTE(0x7f) }

/* Bit set in the bytes that have a zero */
static inline long prep_zero_mask(unsigned long val, unsigned long rhs, const struct word_at_a_time *c)
{
	unsigned long mask = (val & c->low_bits) + c->low_bits;
	return ~(mask | rhs);
}

#define create_zero_mask(mask) (mask)

static inline long find_zero(unsigned long mask)
{
	long leading_zero_bits;

	asm (PPC_CNTLZL "%0,%1" : "=r" (leading_zero_bits) : "r" (mask));
	return leading_zero_bits >> 3;
}

static inline bool has_zero(unsigned long val, unsigned long *data, const struct word_at_a_time *c)
{
	unsigned long rhs = val | c->low_bits;
	*data = rhs;
	return (val + c->high_bits) & ~rhs;
}

static inline unsigned long zero_bytemask(unsigned long mask)
{
	return ~1ul << __fls(mask);
}

#else

#ifdef CONFIG_64BIT

/* unused */
struct word_at_a_time {
};

#define WORD_AT_A_TIME_CONSTANTS { }

/* This will give us 0xff for a NULL char and 0x00 elsewhere */
static inline unsigned long has_zero(unsigned long a, unsigned long *bits, const struct word_at_a_time *c)
{
	unsigned long ret;
	unsigned long zero = 0;

	asm("cmpb %0,%1,%2" : "=r" (ret) : "r" (a), "r" (zero));
	*bits = ret;

	return ret;
}

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

/* Alan Modra's little-endian strlen tail for 64-bit */
static inline unsigned long create_zero_mask(unsigned long bits)
{
	unsigned long leading_zero_bits;
	long trailing_zero_bit_mask;

	asm("addi	%1,%2,-1\n\t"
	    "andc	%1,%1,%2\n\t"
	    "popcntd	%0,%1"
		: "=r" (leading_zero_bits), "=&r" (trailing_zero_bit_mask)
		: "b" (bits));

	return leading_zero_bits;
}

static inline unsigned long find_zero(unsigned long mask)
{
	return mask >> 3;
}

/* This assumes that we never ask for an all 1s bitmask */
static inline unsigned long zero_bytemask(unsigned long mask)
{
	return (1UL << mask) - 1;
}

#else	/* 32-bit case */

struct word_at_a_time {
	const unsigned long one_bits, high_bits;
};

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

/*
 * 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.
 */

/* 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;
}

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

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

/* 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;
}

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

#endif /* CONFIG_64BIT */

#endif /* __BIG_ENDIAN__ */

/*
 * We use load_unaligned_zero() in a selftest, which builds a userspace
 * program. Some linker scripts seem to discard the .fixup section, so allow
 * the test code to use a different section name.
 */
#ifndef FIXUP_SECTION
#define FIXUP_SECTION ".fixup"
#endif

static inline unsigned long load_unaligned_zeropad(const void *addr)
{
	unsigned long ret, offset, tmp;

	asm(
	"1:	" PPC_LL "%[ret], 0(%[addr])\n"
	"2:\n"
	".section " FIXUP_SECTION ",\"ax\"\n"
	"3:	"
#ifdef __powerpc64__
	"clrrdi		%[tmp], %[addr], 3\n\t"
	"clrlsldi	%[offset], %[addr], 61, 3\n\t"
	"ld		%[ret], 0(%[tmp])\n\t"
#ifdef __BIG_ENDIAN__
	"sld		%[ret], %[ret], %[offset]\n\t"
#else
	"srd		%[ret], %[ret], %[offset]\n\t"
#endif
#else
	"clrrwi		%[tmp], %[addr], 2\n\t"
	"clrlslwi	%[offset], %[addr], 30, 3\n\t"
	"lwz		%[ret], 0(%[tmp])\n\t"
#ifdef __BIG_ENDIAN__
	"slw		%[ret], %[ret], %[offset]\n\t"
#else
	"srw		%[ret], %[ret], %[offset]\n\t"
#endif
#endif
	"b	2b\n"
	".previous\n"
	".section __ex_table,\"a\"\n\t"
		PPC_LONG_ALIGN "\n\t"
		PPC_LONG "1b,3b\n"
	".previous"
	: [tmp] "=&b" (tmp), [offset] "=&r" (offset), [ret] "=&r" (ret)
	: [addr] "b" (addr), "m" (*(unsigned long *)addr));

	return ret;
}

#undef FIXUP_SECTION

#endif /* _ASM_WORD_AT_A_TIME_H */
Commit	Line	Data
1629372c PM	1	#ifndef _ASM_WORD_AT_A_TIME_H
	2	#define _ASM_WORD_AT_A_TIME_H
	3
	4	/*
	5	* Word-at-a-time interfaces for PowerPC.
	6	*/
	7
	8	#include <linux/kernel.h>
	9	#include <asm/asm-compat.h>
	10
4c74c330 AB	11	#ifdef __BIG_ENDIAN__
4c74c330 AB	12
1629372c PM	13	struct word_at_a_time {
	14	const unsigned long high_bits, low_bits;
	15	};
	16
	17	#define WORD_AT_A_TIME_CONSTANTS { REPEAT_BYTE(0xfe) + 1, REPEAT_BYTE(0x7f) }
	18
	19	/* Bit set in the bytes that have a zero */
	20	static inline long prep_zero_mask(unsigned long val, unsigned long rhs, const struct word_at_a_time *c)
	21	{
	22	unsigned long mask = (val & c->low_bits) + c->low_bits;
	23	return ~(mask \| rhs);
	24	}
	25
	26	#define create_zero_mask(mask) (mask)
	27
	28	static inline long find_zero(unsigned long mask)
	29	{
	30	long leading_zero_bits;
	31
	32	asm (PPC_CNTLZL "%0,%1" : "=r" (leading_zero_bits) : "r" (mask));
	33	return leading_zero_bits >> 3;
	34	}
	35
	36	static inline bool has_zero(unsigned long val, unsigned long data, const struct word_at_a_time c)
	37	{
	38	unsigned long rhs = val \| c->low_bits;
	39	*data = rhs;
	40	return (val + c->high_bits) & ~rhs;
	41	}
	42
7a5692e6 CM	43	static inline unsigned long zero_bytemask(unsigned long mask)
	44	{
	45	return ~1ul << __fls(mask);
	46	}
	47
4c74c330 AB	48	#else
4c74c330 AB	49
8989aa4a AB	50	#ifdef CONFIG_64BIT
	51
	52	/* unused */
4c74c330	53	struct word_at_a_time {
4c74c330 AB	54	};
4c74c330 AB	55
8989aa4a	56	#define WORD_AT_A_TIME_CONSTANTS { }
4c74c330	57
8989aa4a AB	58	/* This will give us 0xff for a NULL char and 0x00 elsewhere */
	59	static inline unsigned long has_zero(unsigned long a, unsigned long bits, const struct word_at_a_time c)
	60	{
	61	unsigned long ret;
	62	unsigned long zero = 0;
4c74c330	63
8989aa4a AB	64	asm("cmpb %0,%1,%2" : "=r" (ret) : "r" (a), "r" (zero));
8989aa4a AB	65	*bits = ret;
d0cebfa6	66
8989aa4a AB	67	return ret;
	68	}
	69
	70	static inline unsigned long prep_zero_mask(unsigned long a, unsigned long bits, const struct word_at_a_time *c)
	71	{
	72	return bits;
	73	}
	74
	75	/* Alan Modra's little-endian strlen tail for 64-bit */
	76	static inline unsigned long create_zero_mask(unsigned long bits)
4c74c330	77	{
d0cebfa6 PB	78	unsigned long leading_zero_bits;
	79	long trailing_zero_bit_mask;
	80
8989aa4a AB	81	asm("addi %1,%2,-1\n\t"
	82	"andc %1,%1,%2\n\t"
	83	"popcntd %0,%1"
	84	: "=r" (leading_zero_bits), "=&r" (trailing_zero_bit_mask)
b4c11211	85	: "b" (bits));
8989aa4a AB	86
	87	return leading_zero_bits;
	88	}
	89
	90	static inline unsigned long find_zero(unsigned long mask)
	91	{
	92	return mask >> 3;
	93	}
	94
	95	/* This assumes that we never ask for an all 1s bitmask */
	96	static inline unsigned long zero_bytemask(unsigned long mask)
	97	{
	98	return (1UL << mask) - 1;
4c74c330 AB	99	}
	100
	101	#else /* 32-bit case */
	102
8989aa4a AB	103	struct word_at_a_time {
	104	const unsigned long one_bits, high_bits;
	105	};
	106
	107	#define WORD_AT_A_TIME_CONSTANTS { REPEAT_BYTE(0x01), REPEAT_BYTE(0x80) }
	108
d0cebfa6 PB	109	/*
	110	* This is largely generic for little-endian machines, but the
	111	* optimal byte mask counting is probably going to be something
	112	* that is architecture-specific. If you have a reliably fast
	113	* bit count instruction, that might be better than the multiply
	114	* and shift, for example.
	115	*/
	116
4c74c330 AB	117	/* Carl Chatfield / Jan Achrenius G+ version for 32-bit */
	118	static inline long count_masked_bytes(long mask)
	119	{
	120	/* (000000 0000ff 00ffff ffffff) -> ( 1 1 2 3 ) */
	121	long a = (0x0ff0001+mask) >> 23;
	122	/* Fix the 1 for 00 case */
	123	return a & mask;
	124	}
	125
d0cebfa6 PB	126	static inline unsigned long create_zero_mask(unsigned long bits)
	127	{
	128	bits = (bits - 1) & ~bits;
	129	return bits >> 7;
	130	}
	131
	132	static inline unsigned long find_zero(unsigned long mask)
	133	{
	134	return count_masked_bytes(mask);
	135	}
	136
4c74c330 AB	137	/* Return nonzero if it has a zero */
	138	static inline unsigned long has_zero(unsigned long a, unsigned long bits, const struct word_at_a_time c)
	139	{
	140	unsigned long mask = ((a - c->one_bits) & ~a) & c->high_bits;
	141	*bits = mask;
	142	return mask;
	143	}
	144
	145	static inline unsigned long prep_zero_mask(unsigned long a, unsigned long bits, const struct word_at_a_time *c)
	146	{
	147	return bits;
	148	}
	149
4c74c330 AB	150	/* The mask we created is directly usable as a bytemask */
	151	#define zero_bytemask(mask) (mask)
	152
8989aa4a AB	153	#endif /* CONFIG_64BIT */
	154
	155	#endif /* __BIG_ENDIAN__ */
4c74c330	156
fe2a1bb1 ME	157	/*
	158	* We use load_unaligned_zero() in a selftest, which builds a userspace
	159	* program. Some linker scripts seem to discard the .fixup section, so allow
	160	* the test code to use a different section name.
	161	*/
	162	#ifndef FIXUP_SECTION
	163	#define FIXUP_SECTION ".fixup"
	164	#endif
	165
de5946c0 AB	166	static inline unsigned long load_unaligned_zeropad(const void *addr)
	167	{
	168	unsigned long ret, offset, tmp;
	169
	170	asm(
	171	"1: " PPC_LL "%[ret], 0(%[addr])\n"
	172	"2:\n"
fe2a1bb1	173	".section " FIXUP_SECTION ",\"ax\"\n"
de5946c0 AB	174	"3: "
	175	#ifdef __powerpc64__
	176	"clrrdi %[tmp], %[addr], 3\n\t"
	177	"clrlsldi %[offset], %[addr], 61, 3\n\t"
	178	"ld %[ret], 0(%[tmp])\n\t"
	179	#ifdef __BIG_ENDIAN__
	180	"sld %[ret], %[ret], %[offset]\n\t"
	181	#else
	182	"srd %[ret], %[ret], %[offset]\n\t"
	183	#endif
	184	#else
	185	"clrrwi %[tmp], %[addr], 2\n\t"
	186	"clrlslwi %[offset], %[addr], 30, 3\n\t"
	187	"lwz %[ret], 0(%[tmp])\n\t"
	188	#ifdef __BIG_ENDIAN__
	189	"slw %[ret], %[ret], %[offset]\n\t"
	190	#else
	191	"srw %[ret], %[ret], %[offset]\n\t"
	192	#endif
	193	#endif
	194	"b 2b\n"
	195	".previous\n"
	196	".section __ex_table,\"a\"\n\t"
	197	PPC_LONG_ALIGN "\n\t"
	198	PPC_LONG "1b,3b\n"
	199	".previous"
	200	: [tmp] "=&b" (tmp), [offset] "=&r" (offset), [ret] "=&r" (ret)
	201	: [addr] "b" (addr), "m" ((unsigned long )addr));
	202
	203	return ret;
	204	}
	205
fe2a1bb1 ME	206	#undef FIXUP_SECTION
fe2a1bb1 ME	207
1629372c	208	#endif /* _ASM_WORD_AT_A_TIME_H */