[mirror_ubuntu-artful-kernel.git] / arch / x86 / include / asm / arch_hweight.h

#ifndef _ASM_X86_HWEIGHT_H
#define _ASM_X86_HWEIGHT_H

#include <asm/cpufeatures.h>

#ifdef CONFIG_64BIT
/* popcnt %edi, %eax -- redundant REX prefix for alignment */
#define POPCNT32 ".byte 0xf3,0x40,0x0f,0xb8,0xc7"
/* popcnt %rdi, %rax */
#define POPCNT64 ".byte 0xf3,0x48,0x0f,0xb8,0xc7"
#define REG_IN "D"
#define REG_OUT "a"
#else
/* popcnt %eax, %eax */
#define POPCNT32 ".byte 0xf3,0x0f,0xb8,0xc0"
#define REG_IN "a"
#define REG_OUT "a"
#endif

/*
 * __sw_hweightXX are called from within the alternatives below
 * and callee-clobbered registers need to be taken care of. See
 * ARCH_HWEIGHT_CFLAGS in <arch/x86/Kconfig> for the respective
 * compiler switches.
 */
static __always_inline unsigned int __arch_hweight32(unsigned int w)
{
	unsigned int res = 0;

	asm (ALTERNATIVE("call __sw_hweight32", POPCNT32, X86_FEATURE_POPCNT)
		     : "="REG_OUT (res)
		     : REG_IN (w));

	return res;
}

static inline unsigned int __arch_hweight16(unsigned int w)
{
	return __arch_hweight32(w & 0xffff);
}

static inline unsigned int __arch_hweight8(unsigned int w)
{
	return __arch_hweight32(w & 0xff);
}

#ifdef CONFIG_X86_32
static inline unsigned long __arch_hweight64(__u64 w)
{
	return  __arch_hweight32((u32)w) +
		__arch_hweight32((u32)(w >> 32));
}
#else
static __always_inline unsigned long __arch_hweight64(__u64 w)
{
	unsigned long res = 0;

	asm (ALTERNATIVE("call __sw_hweight64", POPCNT64, X86_FEATURE_POPCNT)
		     : "="REG_OUT (res)
		     : REG_IN (w));

	return res;
}
#endif /* CONFIG_X86_32 */

#endif
Commit	Line	Data
d61931d8 BP	1	#ifndef _ASM_X86_HWEIGHT_H
	2	#define _ASM_X86_HWEIGHT_H
	3
cd4d09ec BP	4	#include <asm/cpufeatures.h>
cd4d09ec BP	5
d61931d8	6	#ifdef CONFIG_64BIT
c59bd568 PA	7	/* popcnt %edi, %eax -- redundant REX prefix for alignment */
c59bd568 PA	8	#define POPCNT32 ".byte 0xf3,0x40,0x0f,0xb8,0xc7"
d61931d8	9	/* popcnt %rdi, %rax */
c59bd568	10	#define POPCNT64 ".byte 0xf3,0x48,0x0f,0xb8,0xc7"
d61931d8 BP	11	#define REG_IN "D"
	12	#define REG_OUT "a"
	13	#else
	14	/* popcnt %eax, %eax */
c59bd568	15	#define POPCNT32 ".byte 0xf3,0x0f,0xb8,0xc0"
d61931d8 BP	16	#define REG_IN "a"
	17	#define REG_OUT "a"
	18	#endif
	19
	20	/*
	21	* __sw_hweightXX are called from within the alternatives below
	22	* and callee-clobbered registers need to be taken care of. See
	23	* ARCH_HWEIGHT_CFLAGS in <arch/x86/Kconfig> for the respective
	24	* compiler switches.
	25	*/
d14edb16	26	static __always_inline unsigned int __arch_hweight32(unsigned int w)
d61931d8 BP	27	{
	28	unsigned int res = 0;
	29
c59bd568	30	asm (ALTERNATIVE("call __sw_hweight32", POPCNT32, X86_FEATURE_POPCNT)
d61931d8 BP	31	: "="REG_OUT (res)
	32	: REG_IN (w));
	33
	34	return res;
	35	}
	36
	37	static inline unsigned int __arch_hweight16(unsigned int w)
	38	{
	39	return __arch_hweight32(w & 0xffff);
	40	}
	41
	42	static inline unsigned int __arch_hweight8(unsigned int w)
	43	{
	44	return __arch_hweight32(w & 0xff);
	45	}
	46
d14edb16	47	#ifdef CONFIG_X86_32
d61931d8 BP	48	static inline unsigned long __arch_hweight64(__u64 w)
d61931d8 BP	49	{
d61931d8 BP	50	return __arch_hweight32((u32)w) +
d61931d8 BP	51	__arch_hweight32((u32)(w >> 32));
d14edb16	52	}
d61931d8	53	#else
d14edb16 DV	54	static __always_inline unsigned long __arch_hweight64(__u64 w)
	55	{
	56	unsigned long res = 0;
	57
c59bd568	58	asm (ALTERNATIVE("call __sw_hweight64", POPCNT64, X86_FEATURE_POPCNT)
d61931d8 BP	59	: "="REG_OUT (res)
d61931d8 BP	60	: REG_IN (w));
d61931d8 BP	61
	62	return res;
	63	}
d14edb16	64	#endif /* CONFIG_X86_32 */
d61931d8 BP	65
d61931d8 BP	66	#endif