[mirror_ubuntu-zesty-kernel.git] / arch / powerpc / include / asm / bitops.h

/*
 * PowerPC atomic bit operations.
 *
 * Merged version by David Gibson <david@gibson.dropbear.id.au>.
 * Based on ppc64 versions by: Dave Engebretsen, Todd Inglett, Don
 * Reed, Pat McCarthy, Peter Bergner, Anton Blanchard.  They
 * originally took it from the ppc32 code.
 *
 * Within a word, bits are numbered LSB first.  Lot's of places make
 * this assumption by directly testing bits with (val & (1<<nr)).
 * This can cause confusion for large (> 1 word) bitmaps on a
 * big-endian system because, unlike little endian, the number of each
 * bit depends on the word size.
 *
 * The bitop functions are defined to work on unsigned longs, so for a
 * ppc64 system the bits end up numbered:
 *   |63..............0|127............64|191...........128|255...........192|
 * and on ppc32:
 *   |31.....0|63....32|95....64|127...96|159..128|191..160|223..192|255..224|
 *
 * There are a few little-endian macros used mostly for filesystem
 * bitmaps, these work on similar bit arrays layouts, but
 * byte-oriented:
 *   |7...0|15...8|23...16|31...24|39...32|47...40|55...48|63...56|
 *
 * The main difference is that bit 3-5 (64b) or 3-4 (32b) in the bit
 * number field needs to be reversed compared to the big-endian bit
 * fields. This can be achieved by XOR with 0x38 (64b) or 0x18 (32b).
 *
 * This program is free software; you can redistribute it and/or
 * modify it under the terms of the GNU General Public License
 * as published by the Free Software Foundation; either version
 * 2 of the License, or (at your option) any later version.
 */

#ifndef _ASM_POWERPC_BITOPS_H
#define _ASM_POWERPC_BITOPS_H

#ifdef __KERNEL__

#ifndef _LINUX_BITOPS_H
#error only <linux/bitops.h> can be included directly
#endif

#include <linux/compiler.h>
#include <asm/asm-compat.h>
#include <asm/synch.h>

/* PPC bit number conversion */
#define PPC_BITLSHIFT(be)	(BITS_PER_LONG - 1 - (be))
#define PPC_BIT(bit)		(1UL << PPC_BITLSHIFT(bit))
#define PPC_BITMASK(bs, be)	((PPC_BIT(bs) - PPC_BIT(be)) | PPC_BIT(bs))

#include <asm/barrier.h>

/* Macro for generating the ***_bits() functions */
#define DEFINE_BITOP(fn, op, prefix)		\
static __inline__ void fn(unsigned long mask,	\
		volatile unsigned long *_p)	\
{						\
	unsigned long old;			\
	unsigned long *p = (unsigned long *)_p;	\
	__asm__ __volatile__ (			\
	prefix					\
"1:"	PPC_LLARX(%0,0,%3,0) "\n"		\
	stringify_in_c(op) "%0,%0,%2\n"		\
	PPC405_ERR77(0,%3)			\
	PPC_STLCX "%0,0,%3\n"			\
	"bne- 1b\n"				\
	: "=&r" (old), "+m" (*p)		\
	: "r" (mask), "r" (p)			\
	: "cc", "memory");			\
}

DEFINE_BITOP(set_bits, or, "")
DEFINE_BITOP(clear_bits, andc, "")
DEFINE_BITOP(clear_bits_unlock, andc, PPC_RELEASE_BARRIER)
DEFINE_BITOP(change_bits, xor, "")

static __inline__ void set_bit(int nr, volatile unsigned long *addr)
{
	set_bits(BIT_MASK(nr), addr + BIT_WORD(nr));
}

static __inline__ void clear_bit(int nr, volatile unsigned long *addr)
{
	clear_bits(BIT_MASK(nr), addr + BIT_WORD(nr));
}

static __inline__ void clear_bit_unlock(int nr, volatile unsigned long *addr)
{
	clear_bits_unlock(BIT_MASK(nr), addr + BIT_WORD(nr));
}

static __inline__ void change_bit(int nr, volatile unsigned long *addr)
{
	change_bits(BIT_MASK(nr), addr + BIT_WORD(nr));
}

/* Like DEFINE_BITOP(), with changes to the arguments to 'op' and the output
 * operands. */
#define DEFINE_TESTOP(fn, op, prefix, postfix, eh)	\
static __inline__ unsigned long fn(			\
		unsigned long mask,			\
		volatile unsigned long *_p)		\
{							\
	unsigned long old, t;				\
	unsigned long *p = (unsigned long *)_p;		\
	__asm__ __volatile__ (				\
	prefix						\
"1:"	PPC_LLARX(%0,0,%3,eh) "\n"			\
	stringify_in_c(op) "%1,%0,%2\n"			\
	PPC405_ERR77(0,%3)				\
	PPC_STLCX "%1,0,%3\n"				\
	"bne- 1b\n"					\
	postfix						\
	: "=&r" (old), "=&r" (t)			\
	: "r" (mask), "r" (p)				\
	: "cc", "memory");				\
	return (old & mask);				\
}

DEFINE_TESTOP(test_and_set_bits, or, PPC_ATOMIC_ENTRY_BARRIER,
	      PPC_ATOMIC_EXIT_BARRIER, 0)
DEFINE_TESTOP(test_and_set_bits_lock, or, "",
	      PPC_ACQUIRE_BARRIER, 1)
DEFINE_TESTOP(test_and_clear_bits, andc, PPC_ATOMIC_ENTRY_BARRIER,
	      PPC_ATOMIC_EXIT_BARRIER, 0)
DEFINE_TESTOP(test_and_change_bits, xor, PPC_ATOMIC_ENTRY_BARRIER,
	      PPC_ATOMIC_EXIT_BARRIER, 0)

static __inline__ int test_and_set_bit(unsigned long nr,
				       volatile unsigned long *addr)
{
	return test_and_set_bits(BIT_MASK(nr), addr + BIT_WORD(nr)) != 0;
}

static __inline__ int test_and_set_bit_lock(unsigned long nr,
				       volatile unsigned long *addr)
{
	return test_and_set_bits_lock(BIT_MASK(nr),
				addr + BIT_WORD(nr)) != 0;
}

static __inline__ int test_and_clear_bit(unsigned long nr,
					 volatile unsigned long *addr)
{
	return test_and_clear_bits(BIT_MASK(nr), addr + BIT_WORD(nr)) != 0;
}

static __inline__ int test_and_change_bit(unsigned long nr,
					  volatile unsigned long *addr)
{
	return test_and_change_bits(BIT_MASK(nr), addr + BIT_WORD(nr)) != 0;
}

#include <asm-generic/bitops/non-atomic.h>

static __inline__ void __clear_bit_unlock(int nr, volatile unsigned long *addr)
{
	__asm__ __volatile__(PPC_RELEASE_BARRIER "" ::: "memory");
	__clear_bit(nr, addr);
}

/*
 * Return the zero-based bit position (LE, not IBM bit numbering) of
 * the most significant 1-bit in a double word.
 */
static __inline__ __attribute__((const))
int __ilog2(unsigned long x)
{
	int lz;

	asm (PPC_CNTLZL "%0,%1" : "=r" (lz) : "r" (x));
	return BITS_PER_LONG - 1 - lz;
}

static inline __attribute__((const))
int __ilog2_u32(u32 n)
{
	int bit;
	asm ("cntlzw %0,%1" : "=r" (bit) : "r" (n));
	return 31 - bit;
}

#ifdef __powerpc64__
static inline __attribute__((const))
int __ilog2_u64(u64 n)
{
	int bit;
	asm ("cntlzd %0,%1" : "=r" (bit) : "r" (n));
	return 63 - bit;
}
#endif

/*
 * Determines the bit position of the least significant 0 bit in the
 * specified double word. The returned bit position will be
 * zero-based, starting from the right side (63/31 - 0).
 */
static __inline__ unsigned long ffz(unsigned long x)
{
	/* no zero exists anywhere in the 8 byte area. */
	if ((x = ~x) == 0)
		return BITS_PER_LONG;

	/*
	 * Calculate the bit position of the least significant '1' bit in x
	 * (since x has been changed this will actually be the least significant
	 * '0' bit in * the original x).  Note: (x & -x) gives us a mask that
	 * is the least significant * (RIGHT-most) 1-bit of the value in x.
	 */
	return __ilog2(x & -x);
}

static __inline__ unsigned long __ffs(unsigned long x)
{
	return __ilog2(x & -x);
}

/*
 * ffs: find first bit set. This is defined the same way as
 * the libc and compiler builtin ffs routines, therefore
 * differs in spirit from the above ffz (man ffs).
 */
static __inline__ int ffs(int x)
{
	unsigned long i = (unsigned long)x;
	return __ilog2(i & -i) + 1;
}

/*
 * fls: find last (most-significant) bit set.
 * Note fls(0) = 0, fls(1) = 1, fls(0x80000000) = 32.
 */
static __inline__ int fls(unsigned int x)
{
	int lz;

	asm ("cntlzw %0,%1" : "=r" (lz) : "r" (x));
	return 32 - lz;
}

static __inline__ unsigned long __fls(unsigned long x)
{
	return __ilog2(x);
}

/*
 * 64-bit can do this using one cntlzd (count leading zeroes doubleword)
 * instruction; for 32-bit we use the generic version, which does two
 * 32-bit fls calls.
 */
#ifdef __powerpc64__
static __inline__ int fls64(__u64 x)
{
	int lz;

	asm ("cntlzd %0,%1" : "=r" (lz) : "r" (x));
	return 64 - lz;
}
#else
#include <asm-generic/bitops/fls64.h>
#endif /* __powerpc64__ */

#ifdef CONFIG_PPC64
unsigned int __arch_hweight8(unsigned int w);
unsigned int __arch_hweight16(unsigned int w);
unsigned int __arch_hweight32(unsigned int w);
unsigned long __arch_hweight64(__u64 w);
#include <asm-generic/bitops/const_hweight.h>
#else
#include <asm-generic/bitops/hweight.h>
#endif

#include <asm-generic/bitops/find.h>

/* Little-endian versions */
#include <asm-generic/bitops/le.h>

/* Bitmap functions for the ext2 filesystem */

#include <asm-generic/bitops/ext2-atomic-setbit.h>

#include <asm-generic/bitops/sched.h>

#endif /* __KERNEL__ */

#endif /* _ASM_POWERPC_BITOPS_H */
Commit	Line	Data
a0e60b20 DG	1	/*
	2	* PowerPC atomic bit operations.
	3	*
	4	* Merged version by David Gibson <david@gibson.dropbear.id.au>.
	5	* Based on ppc64 versions by: Dave Engebretsen, Todd Inglett, Don
	6	* Reed, Pat McCarthy, Peter Bergner, Anton Blanchard. They
	7	* originally took it from the ppc32 code.
	8	*
	9	* Within a word, bits are numbered LSB first. Lot's of places make
	10	* this assumption by directly testing bits with (val & (1<<nr)).
	11	* This can cause confusion for large (> 1 word) bitmaps on a
	12	* big-endian system because, unlike little endian, the number of each
	13	* bit depends on the word size.
	14	*
	15	* The bitop functions are defined to work on unsigned longs, so for a
	16	* ppc64 system the bits end up numbered:
e7a7a65e	17	* \|63..............0\|127............64\|191...........128\|255...........192\|
a0e60b20	18	* and on ppc32:
e7a7a65e	19	* \|31.....0\|63....32\|95....64\|127...96\|159..128\|191..160\|223..192\|255..224\|
a0e60b20 DG	20	*
	21	* There are a few little-endian macros used mostly for filesystem
	22	* bitmaps, these work on similar bit arrays layouts, but
	23	* byte-oriented:
	24	* \|7...0\|15...8\|23...16\|31...24\|39...32\|47...40\|55...48\|63...56\|
	25	*
	26	* The main difference is that bit 3-5 (64b) or 3-4 (32b) in the bit
	27	* number field needs to be reversed compared to the big-endian bit
	28	* fields. This can be achieved by XOR with 0x38 (64b) or 0x18 (32b).
	29	*
	30	* This program is free software; you can redistribute it and/or
	31	* modify it under the terms of the GNU General Public License
	32	* as published by the Free Software Foundation; either version
	33	* 2 of the License, or (at your option) any later version.
	34	*/
	35
	36	#ifndef _ASM_POWERPC_BITOPS_H
	37	#define _ASM_POWERPC_BITOPS_H
	38
	39	#ifdef __KERNEL__
	40
0624517d JS	41	#ifndef _LINUX_BITOPS_H
	42	#error only <linux/bitops.h> can be included directly
	43	#endif
	44
a0e60b20	45	#include <linux/compiler.h>
3ddfbcf1	46	#include <asm/asm-compat.h>
a0e60b20 DG	47	#include <asm/synch.h>
a0e60b20 DG	48
e22a2274 MS	49	/* PPC bit number conversion */
	50	#define PPC_BITLSHIFT(be) (BITS_PER_LONG - 1 - (be))
	51	#define PPC_BIT(bit) (1UL << PPC_BITLSHIFT(bit))
	52	#define PPC_BITMASK(bs, be) ((PPC_BIT(bs) - PPC_BIT(be)) \| PPC_BIT(bs))
	53
c645073f	54	#include <asm/barrier.h>
a0e60b20	55
0d2d3e38	56	/* Macro for generating the **_bits() functions /
576be130	57	#define DEFINE_BITOP(fn, op, prefix) \
0d2d3e38 GT	58	static __inline__ void fn(unsigned long mask, \
	59	volatile unsigned long *_p) \
	60	{ \
	61	unsigned long old; \
	62	unsigned long p = (unsigned long )_p; \
	63	__asm__ __volatile__ ( \
	64	prefix \
864b9e6f	65	"1:" PPC_LLARX(%0,0,%3,0) "\n" \
0d2d3e38 GT	66	stringify_in_c(op) "%0,%0,%2\n" \
	67	PPC405_ERR77(0,%3) \
	68	PPC_STLCX "%0,0,%3\n" \
	69	"bne- 1b\n" \
0d2d3e38 GT	70	: "=&r" (old), "+m" (*p) \
	71	: "r" (mask), "r" (p) \
	72	: "cc", "memory"); \
	73	}
	74
576be130 ME	75	DEFINE_BITOP(set_bits, or, "")
	76	DEFINE_BITOP(clear_bits, andc, "")
	77	DEFINE_BITOP(clear_bits_unlock, andc, PPC_RELEASE_BARRIER)
	78	DEFINE_BITOP(change_bits, xor, "")
0d2d3e38	79
a0e60b20 DG	80	static __inline__ void set_bit(int nr, volatile unsigned long *addr)
a0e60b20 DG	81	{
2237f4f4	82	set_bits(BIT_MASK(nr), addr + BIT_WORD(nr));
a0e60b20 DG	83	}
	84
	85	static __inline__ void clear_bit(int nr, volatile unsigned long *addr)
	86	{
2237f4f4	87	clear_bits(BIT_MASK(nr), addr + BIT_WORD(nr));
a0e60b20 DG	88	}
a0e60b20 DG	89
66ffb04c NP	90	static __inline__ void clear_bit_unlock(int nr, volatile unsigned long *addr)
66ffb04c NP	91	{
2237f4f4	92	clear_bits_unlock(BIT_MASK(nr), addr + BIT_WORD(nr));
66ffb04c NP	93	}
66ffb04c NP	94
a0e60b20 DG	95	static __inline__ void change_bit(int nr, volatile unsigned long *addr)
a0e60b20 DG	96	{
2237f4f4	97	change_bits(BIT_MASK(nr), addr + BIT_WORD(nr));
0d2d3e38	98	}
a0e60b20	99
0d2d3e38 GT	100	/* Like DEFINE_BITOP(), with changes to the arguments to 'op' and the output
0d2d3e38 GT	101	* operands. */
864b9e6f AB	102	#define DEFINE_TESTOP(fn, op, prefix, postfix, eh) \
	103	static __inline__ unsigned long fn( \
	104	unsigned long mask, \
	105	volatile unsigned long *_p) \
	106	{ \
	107	unsigned long old, t; \
	108	unsigned long p = (unsigned long )_p; \
	109	__asm__ __volatile__ ( \
	110	prefix \
	111	"1:" PPC_LLARX(%0,0,%3,eh) "\n" \
	112	stringify_in_c(op) "%1,%0,%2\n" \
	113	PPC405_ERR77(0,%3) \
	114	PPC_STLCX "%1,0,%3\n" \
	115	"bne- 1b\n" \
	116	postfix \
	117	: "=&r" (old), "=&r" (t) \
	118	: "r" (mask), "r" (p) \
	119	: "cc", "memory"); \
	120	return (old & mask); \
a0e60b20 DG	121	}
a0e60b20 DG	122
b97021f8 BH	123	DEFINE_TESTOP(test_and_set_bits, or, PPC_ATOMIC_ENTRY_BARRIER,
b97021f8 BH	124	PPC_ATOMIC_EXIT_BARRIER, 0)
f10e2e5b AB	125	DEFINE_TESTOP(test_and_set_bits_lock, or, "",
f10e2e5b AB	126	PPC_ACQUIRE_BARRIER, 1)
b97021f8 BH	127	DEFINE_TESTOP(test_and_clear_bits, andc, PPC_ATOMIC_ENTRY_BARRIER,
	128	PPC_ATOMIC_EXIT_BARRIER, 0)
	129	DEFINE_TESTOP(test_and_change_bits, xor, PPC_ATOMIC_ENTRY_BARRIER,
	130	PPC_ATOMIC_EXIT_BARRIER, 0)
0d2d3e38	131
a0e60b20 DG	132	static __inline__ int test_and_set_bit(unsigned long nr,
	133	volatile unsigned long *addr)
	134	{
2237f4f4	135	return test_and_set_bits(BIT_MASK(nr), addr + BIT_WORD(nr)) != 0;
a0e60b20 DG	136	}
a0e60b20 DG	137
66ffb04c NP	138	static __inline__ int test_and_set_bit_lock(unsigned long nr,
	139	volatile unsigned long *addr)
	140	{
2237f4f4 AM	141	return test_and_set_bits_lock(BIT_MASK(nr),
2237f4f4 AM	142	addr + BIT_WORD(nr)) != 0;
66ffb04c NP	143	}
66ffb04c NP	144
a0e60b20 DG	145	static __inline__ int test_and_clear_bit(unsigned long nr,
	146	volatile unsigned long *addr)
	147	{
2237f4f4	148	return test_and_clear_bits(BIT_MASK(nr), addr + BIT_WORD(nr)) != 0;
a0e60b20 DG	149	}
	150
	151	static __inline__ int test_and_change_bit(unsigned long nr,
	152	volatile unsigned long *addr)
	153	{
2237f4f4	154	return test_and_change_bits(BIT_MASK(nr), addr + BIT_WORD(nr)) != 0;
a0e60b20 DG	155	}
a0e60b20 DG	156
e779b2f9	157	#include <asm-generic/bitops/non-atomic.h>
a0e60b20	158
66ffb04c NP	159	static __inline__ void __clear_bit_unlock(int nr, volatile unsigned long *addr)
66ffb04c NP	160	{
f10e2e5b	161	__asm__ __volatile__(PPC_RELEASE_BARRIER "" ::: "memory");
66ffb04c NP	162	__clear_bit(nr, addr);
	163	}
	164
a0e60b20 DG	165	/*
	166	* Return the zero-based bit position (LE, not IBM bit numbering) of
	167	* the most significant 1-bit in a double word.
	168	*/
ef55d53c DH	169	static __inline__ __attribute__((const))
ef55d53c DH	170	int __ilog2(unsigned long x)
a0e60b20 DG	171	{
	172	int lz;
	173
3ddfbcf1	174	asm (PPC_CNTLZL "%0,%1" : "=r" (lz) : "r" (x));
a0e60b20 DG	175	return BITS_PER_LONG - 1 - lz;
	176	}
	177
ef55d53c DH	178	static inline __attribute__((const))
	179	int __ilog2_u32(u32 n)
	180	{
	181	int bit;
	182	asm ("cntlzw %0,%1" : "=r" (bit) : "r" (n));
	183	return 31 - bit;
	184	}
	185
	186	#ifdef __powerpc64__
	187	static inline __attribute__((const))
02241696	188	int __ilog2_u64(u64 n)
ef55d53c DH	189	{
	190	int bit;
	191	asm ("cntlzd %0,%1" : "=r" (bit) : "r" (n));
	192	return 63 - bit;
	193	}
	194	#endif
	195
a0e60b20 DG	196	/*
	197	* Determines the bit position of the least significant 0 bit in the
	198	* specified double word. The returned bit position will be
	199	* zero-based, starting from the right side (63/31 - 0).
	200	*/
	201	static __inline__ unsigned long ffz(unsigned long x)
	202	{
	203	/* no zero exists anywhere in the 8 byte area. */
	204	if ((x = ~x) == 0)
	205	return BITS_PER_LONG;
	206
	207	/*
25985edc LDM	208	* Calculate the bit position of the least significant '1' bit in x
25985edc LDM	209	* (since x has been changed this will actually be the least significant
a0e60b20 DG	210	* '0' bit in * the original x). Note: (x & -x) gives us a mask that
	211	* is the least significant * (RIGHT-most) 1-bit of the value in x.
	212	*/
	213	return __ilog2(x & -x);
	214	}
	215
6e4c632c	216	static __inline__ unsigned long __ffs(unsigned long x)
a0e60b20 DG	217	{
	218	return __ilog2(x & -x);
	219	}
	220
	221	/*
	222	* ffs: find first bit set. This is defined the same way as
	223	* the libc and compiler builtin ffs routines, therefore
	224	* differs in spirit from the above ffz (man ffs).
	225	*/
	226	static __inline__ int ffs(int x)
	227	{
	228	unsigned long i = (unsigned long)x;
	229	return __ilog2(i & -i) + 1;
	230	}
	231
	232	/*
	233	* fls: find last (most-significant) bit set.
	234	* Note fls(0) = 0, fls(1) = 1, fls(0x80000000) = 32.
	235	*/
	236	static __inline__ int fls(unsigned int x)
	237	{
	238	int lz;
	239
	240	asm ("cntlzw %0,%1" : "=r" (lz) : "r" (x));
	241	return 32 - lz;
	242	}
9f264be6	243
56a6b1eb AH	244	static __inline__ unsigned long __fls(unsigned long x)
	245	{
	246	return __ilog2(x);
	247	}
	248
9f264be6 PM	249	/*
	250	* 64-bit can do this using one cntlzd (count leading zeroes doubleword)
	251	* instruction; for 32-bit we use the generic version, which does two
	252	* 32-bit fls calls.
	253	*/
	254	#ifdef __powerpc64__
	255	static __inline__ int fls64(__u64 x)
	256	{
	257	int lz;
	258
	259	asm ("cntlzd %0,%1" : "=r" (lz) : "r" (x));
	260	return 64 - lz;
	261	}
	262	#else
e779b2f9	263	#include <asm-generic/bitops/fls64.h>
9f264be6 PM	264	#endif /* __powerpc64__ */
9f264be6 PM	265
64ff3128 AB	266	#ifdef CONFIG_PPC64
	267	unsigned int __arch_hweight8(unsigned int w);
	268	unsigned int __arch_hweight16(unsigned int w);
	269	unsigned int __arch_hweight32(unsigned int w);
	270	unsigned long __arch_hweight64(__u64 w);
	271	#include <asm-generic/bitops/const_hweight.h>
	272	#else
e779b2f9	273	#include <asm-generic/bitops/hweight.h>
64ff3128 AB	274	#endif
64ff3128 AB	275
47b9d9bd	276	#include <asm-generic/bitops/find.h>
a0e60b20 DG	277
a0e60b20 DG	278	/* Little-endian versions */
79597be9	279	#include <asm-generic/bitops/le.h>
a0e60b20	280
a0e60b20 DG	281	/* Bitmap functions for the ext2 filesystem */
a0e60b20 DG	282
148817ba	283	#include <asm-generic/bitops/ext2-atomic-setbit.h>
a0e60b20	284
e779b2f9	285	#include <asm-generic/bitops/sched.h>
a0e60b20 DG	286
	287	#endif /* __KERNEL__ */
	288
	289	#endif /* _ASM_POWERPC_BITOPS_H */