[deliverable/linux.git] / arch / arm / include / asm / xor.h

/*
 *  arch/arm/include/asm/xor.h
 *
 *  Copyright (C) 2001 Russell King
 *
 * This program is free software; you can redistribute it and/or modify
 * it under the terms of the GNU General Public License version 2 as
 * published by the Free Software Foundation.
 */
#include <linux/hardirq.h>
#include <asm-generic/xor.h>
#include <asm/hwcap.h>
#include <asm/neon.h>

#define __XOR(a1, a2) a1 ^= a2

#define GET_BLOCK_2(dst) \
	__asm__("ldmia	%0, {%1, %2}" \
		: "=r" (dst), "=r" (a1), "=r" (a2) \
		: "0" (dst))

#define GET_BLOCK_4(dst) \
	__asm__("ldmia	%0, {%1, %2, %3, %4}" \
		: "=r" (dst), "=r" (a1), "=r" (a2), "=r" (a3), "=r" (a4) \
		: "0" (dst))

#define XOR_BLOCK_2(src) \
	__asm__("ldmia	%0!, {%1, %2}" \
		: "=r" (src), "=r" (b1), "=r" (b2) \
		: "0" (src)); \
	__XOR(a1, b1); __XOR(a2, b2);

#define XOR_BLOCK_4(src) \
	__asm__("ldmia	%0!, {%1, %2, %3, %4}" \
		: "=r" (src), "=r" (b1), "=r" (b2), "=r" (b3), "=r" (b4) \
		: "0" (src)); \
	__XOR(a1, b1); __XOR(a2, b2); __XOR(a3, b3); __XOR(a4, b4)

#define PUT_BLOCK_2(dst) \
	__asm__ __volatile__("stmia	%0!, {%2, %3}" \
		: "=r" (dst) \
		: "0" (dst), "r" (a1), "r" (a2))

#define PUT_BLOCK_4(dst) \
	__asm__ __volatile__("stmia	%0!, {%2, %3, %4, %5}" \
		: "=r" (dst) \
		: "0" (dst), "r" (a1), "r" (a2), "r" (a3), "r" (a4))

static void
xor_arm4regs_2(unsigned long bytes, unsigned long *p1, unsigned long *p2)
{
	unsigned int lines = bytes / sizeof(unsigned long) / 4;
	register unsigned int a1 __asm__("r4");
	register unsigned int a2 __asm__("r5");
	register unsigned int a3 __asm__("r6");
	register unsigned int a4 __asm__("r7");
	register unsigned int b1 __asm__("r8");
	register unsigned int b2 __asm__("r9");
	register unsigned int b3 __asm__("ip");
	register unsigned int b4 __asm__("lr");

	do {
		GET_BLOCK_4(p1);
		XOR_BLOCK_4(p2);
		PUT_BLOCK_4(p1);
	} while (--lines);
}

static void
xor_arm4regs_3(unsigned long bytes, unsigned long *p1, unsigned long *p2,
		unsigned long *p3)
{
	unsigned int lines = bytes / sizeof(unsigned long) / 4;
	register unsigned int a1 __asm__("r4");
	register unsigned int a2 __asm__("r5");
	register unsigned int a3 __asm__("r6");
	register unsigned int a4 __asm__("r7");
	register unsigned int b1 __asm__("r8");
	register unsigned int b2 __asm__("r9");
	register unsigned int b3 __asm__("ip");
	register unsigned int b4 __asm__("lr");

	do {
		GET_BLOCK_4(p1);
		XOR_BLOCK_4(p2);
		XOR_BLOCK_4(p3);
		PUT_BLOCK_4(p1);
	} while (--lines);
}

static void
xor_arm4regs_4(unsigned long bytes, unsigned long *p1, unsigned long *p2,
		unsigned long *p3, unsigned long *p4)
{
	unsigned int lines = bytes / sizeof(unsigned long) / 2;
	register unsigned int a1 __asm__("r8");
	register unsigned int a2 __asm__("r9");
	register unsigned int b1 __asm__("ip");
	register unsigned int b2 __asm__("lr");

	do {
		GET_BLOCK_2(p1);
		XOR_BLOCK_2(p2);
		XOR_BLOCK_2(p3);
		XOR_BLOCK_2(p4);
		PUT_BLOCK_2(p1);
	} while (--lines);
}

static void
xor_arm4regs_5(unsigned long bytes, unsigned long *p1, unsigned long *p2,
		unsigned long *p3, unsigned long *p4, unsigned long *p5)
{
	unsigned int lines = bytes / sizeof(unsigned long) / 2;
	register unsigned int a1 __asm__("r8");
	register unsigned int a2 __asm__("r9");
	register unsigned int b1 __asm__("ip");
	register unsigned int b2 __asm__("lr");

	do {
		GET_BLOCK_2(p1);
		XOR_BLOCK_2(p2);
		XOR_BLOCK_2(p3);
		XOR_BLOCK_2(p4);
		XOR_BLOCK_2(p5);
		PUT_BLOCK_2(p1);
	} while (--lines);
}

static struct xor_block_template xor_block_arm4regs = {
	.name	= "arm4regs",
	.do_2	= xor_arm4regs_2,
	.do_3	= xor_arm4regs_3,
	.do_4	= xor_arm4regs_4,
	.do_5	= xor_arm4regs_5,
};

#undef XOR_TRY_TEMPLATES
#define XOR_TRY_TEMPLATES			\
	do {					\
		xor_speed(&xor_block_arm4regs);	\
		xor_speed(&xor_block_8regs);	\
		xor_speed(&xor_block_32regs);	\
		NEON_TEMPLATES;			\
	} while (0)

#ifdef CONFIG_KERNEL_MODE_NEON

extern struct xor_block_template const xor_block_neon_inner;

static void
xor_neon_2(unsigned long bytes, unsigned long *p1, unsigned long *p2)
{
	if (in_interrupt()) {
		xor_arm4regs_2(bytes, p1, p2);
	} else {
		kernel_neon_begin();
		xor_block_neon_inner.do_2(bytes, p1, p2);
		kernel_neon_end();
	}
}

static void
xor_neon_3(unsigned long bytes, unsigned long *p1, unsigned long *p2,
		unsigned long *p3)
{
	if (in_interrupt()) {
		xor_arm4regs_3(bytes, p1, p2, p3);
	} else {
		kernel_neon_begin();
		xor_block_neon_inner.do_3(bytes, p1, p2, p3);
		kernel_neon_end();
	}
}

static void
xor_neon_4(unsigned long bytes, unsigned long *p1, unsigned long *p2,
		unsigned long *p3, unsigned long *p4)
{
	if (in_interrupt()) {
		xor_arm4regs_4(bytes, p1, p2, p3, p4);
	} else {
		kernel_neon_begin();
		xor_block_neon_inner.do_4(bytes, p1, p2, p3, p4);
		kernel_neon_end();
	}
}

static void
xor_neon_5(unsigned long bytes, unsigned long *p1, unsigned long *p2,
		unsigned long *p3, unsigned long *p4, unsigned long *p5)
{
	if (in_interrupt()) {
		xor_arm4regs_5(bytes, p1, p2, p3, p4, p5);
	} else {
		kernel_neon_begin();
		xor_block_neon_inner.do_5(bytes, p1, p2, p3, p4, p5);
		kernel_neon_end();
	}
}

static struct xor_block_template xor_block_neon = {
	.name	= "neon",
	.do_2	= xor_neon_2,
	.do_3	= xor_neon_3,
	.do_4	= xor_neon_4,
	.do_5	= xor_neon_5
};

#define NEON_TEMPLATES	\
	do { if (cpu_has_neon()) xor_speed(&xor_block_neon); } while (0)
#else
#define NEON_TEMPLATES
#endif
Commit	Line	Data
1da177e4	1	/*
4baa9922	2	* arch/arm/include/asm/xor.h
1da177e4 LT	3	*
	4	* Copyright (C) 2001 Russell King
	5	*
	6	* This program is free software; you can redistribute it and/or modify
	7	* it under the terms of the GNU General Public License version 2 as
	8	* published by the Free Software Foundation.
	9	*/
01956597	10	#include <linux/hardirq.h>
1da177e4	11	#include <asm-generic/xor.h>
01956597 AB	12	#include <asm/hwcap.h>
01956597 AB	13	#include <asm/neon.h>
1da177e4 LT	14
	15	#define __XOR(a1, a2) a1 ^= a2
	16
	17	#define GET_BLOCK_2(dst) \
	18	__asm__("ldmia %0, {%1, %2}" \
	19	: "=r" (dst), "=r" (a1), "=r" (a2) \
	20	: "0" (dst))
	21
	22	#define GET_BLOCK_4(dst) \
	23	__asm__("ldmia %0, {%1, %2, %3, %4}" \
	24	: "=r" (dst), "=r" (a1), "=r" (a2), "=r" (a3), "=r" (a4) \
	25	: "0" (dst))
	26
	27	#define XOR_BLOCK_2(src) \
	28	__asm__("ldmia %0!, {%1, %2}" \
	29	: "=r" (src), "=r" (b1), "=r" (b2) \
	30	: "0" (src)); \
	31	__XOR(a1, b1); __XOR(a2, b2);
	32
	33	#define XOR_BLOCK_4(src) \
	34	__asm__("ldmia %0!, {%1, %2, %3, %4}" \
	35	: "=r" (src), "=r" (b1), "=r" (b2), "=r" (b3), "=r" (b4) \
	36	: "0" (src)); \
	37	__XOR(a1, b1); __XOR(a2, b2); __XOR(a3, b3); __XOR(a4, b4)
	38
	39	#define PUT_BLOCK_2(dst) \
	40	__asm__ __volatile__("stmia %0!, {%2, %3}" \
	41	: "=r" (dst) \
	42	: "0" (dst), "r" (a1), "r" (a2))
	43
	44	#define PUT_BLOCK_4(dst) \
	45	__asm__ __volatile__("stmia %0!, {%2, %3, %4, %5}" \
	46	: "=r" (dst) \
	47	: "0" (dst), "r" (a1), "r" (a2), "r" (a3), "r" (a4))
	48
	49	static void
	50	xor_arm4regs_2(unsigned long bytes, unsigned long p1, unsigned long p2)
	51	{
	52	unsigned int lines = bytes / sizeof(unsigned long) / 4;
	53	register unsigned int a1 __asm__("r4");
	54	register unsigned int a2 __asm__("r5");
	55	register unsigned int a3 __asm__("r6");
	56	register unsigned int a4 __asm__("r7");
	57	register unsigned int b1 __asm__("r8");
	58	register unsigned int b2 __asm__("r9");
	59	register unsigned int b3 __asm__("ip");
	60	register unsigned int b4 __asm__("lr");
	61
	62	do {
	63	GET_BLOCK_4(p1);
	64	XOR_BLOCK_4(p2);
	65	PUT_BLOCK_4(p1);
	66	} while (--lines);
	67	}
	68
	69	static void
	70	xor_arm4regs_3(unsigned long bytes, unsigned long p1, unsigned long p2,
	71	unsigned long *p3)
	72	{
	73	unsigned int lines = bytes / sizeof(unsigned long) / 4;
	74	register unsigned int a1 __asm__("r4");
	75	register unsigned int a2 __asm__("r5");
	76	register unsigned int a3 __asm__("r6");
	77	register unsigned int a4 __asm__("r7");
78	register unsigned int b1 __asm__("r8");
79	register unsigned int b2 __asm__("r9");
80	register unsigned int b3 __asm__("ip");
81	register unsigned int b4 __asm__("lr");
82
83	do {
84	GET_BLOCK_4(p1);
85	XOR_BLOCK_4(p2);
86	XOR_BLOCK_4(p3);
87	PUT_BLOCK_4(p1);
88	} while (--lines);
89	}
90
91	static void
92	xor_arm4regs_4(unsigned long bytes, unsigned long p1, unsigned long p2,
93	unsigned long p3, unsigned long p4)
94	{
95	unsigned int lines = bytes / sizeof(unsigned long) / 2;
96	register unsigned int a1 __asm__("r8");
97	register unsigned int a2 __asm__("r9");
98	register unsigned int b1 __asm__("ip");
99	register unsigned int b2 __asm__("lr");
100
101	do {
102	GET_BLOCK_2(p1);
103	XOR_BLOCK_2(p2);
104	XOR_BLOCK_2(p3);
105	XOR_BLOCK_2(p4);
106	PUT_BLOCK_2(p1);
107	} while (--lines);
108	}
109
110	static void
111	xor_arm4regs_5(unsigned long bytes, unsigned long p1, unsigned long p2,
112	unsigned long p3, unsigned long p4, unsigned long *p5)
113	{
114	unsigned int lines = bytes / sizeof(unsigned long) / 2;
115	register unsigned int a1 __asm__("r8");
116	register unsigned int a2 __asm__("r9");
117	register unsigned int b1 __asm__("ip");
118	register unsigned int b2 __asm__("lr");
119
120	do {
121	GET_BLOCK_2(p1);
122	XOR_BLOCK_2(p2);
123	XOR_BLOCK_2(p3);
124	XOR_BLOCK_2(p4);
125	XOR_BLOCK_2(p5);
126	PUT_BLOCK_2(p1);
127	} while (--lines);
128	}
129
130	static struct xor_block_template xor_block_arm4regs = {
131	.name = "arm4regs",
132	.do_2 = xor_arm4regs_2,
133	.do_3 = xor_arm4regs_3,
134	.do_4 = xor_arm4regs_4,
135	.do_5 = xor_arm4regs_5,
136	};
137
138	#undef XOR_TRY_TEMPLATES
139	#define XOR_TRY_TEMPLATES \
140	do { \
141	xor_speed(&xor_block_arm4regs); \
142	xor_speed(&xor_block_8regs); \
143	xor_speed(&xor_block_32regs); \
01956597	144	NEON_TEMPLATES; \
1da177e4	145	} while (0)
01956597 AB	146
	147	#ifdef CONFIG_KERNEL_MODE_NEON
	148
	149	extern struct xor_block_template const xor_block_neon_inner;
	150
	151	static void
	152	xor_neon_2(unsigned long bytes, unsigned long p1, unsigned long p2)
	153	{
	154	if (in_interrupt()) {
	155	xor_arm4regs_2(bytes, p1, p2);
	156	} else {
	157	kernel_neon_begin();
	158	xor_block_neon_inner.do_2(bytes, p1, p2);
	159	kernel_neon_end();
	160	}
	161	}
	162
	163	static void
	164	xor_neon_3(unsigned long bytes, unsigned long p1, unsigned long p2,
	165	unsigned long *p3)
	166	{
	167	if (in_interrupt()) {
	168	xor_arm4regs_3(bytes, p1, p2, p3);
	169	} else {
	170	kernel_neon_begin();
	171	xor_block_neon_inner.do_3(bytes, p1, p2, p3);
	172	kernel_neon_end();
	173	}
	174	}
	175
	176	static void
	177	xor_neon_4(unsigned long bytes, unsigned long p1, unsigned long p2,
	178	unsigned long p3, unsigned long p4)
	179	{
	180	if (in_interrupt()) {
	181	xor_arm4regs_4(bytes, p1, p2, p3, p4);
	182	} else {
	183	kernel_neon_begin();
	184	xor_block_neon_inner.do_4(bytes, p1, p2, p3, p4);
	185	kernel_neon_end();
	186	}
	187	}
	188
	189	static void
	190	xor_neon_5(unsigned long bytes, unsigned long p1, unsigned long p2,
	191	unsigned long p3, unsigned long p4, unsigned long *p5)
	192	{
	193	if (in_interrupt()) {
	194	xor_arm4regs_5(bytes, p1, p2, p3, p4, p5);
	195	} else {
	196	kernel_neon_begin();
	197	xor_block_neon_inner.do_5(bytes, p1, p2, p3, p4, p5);
	198	kernel_neon_end();
	199	}
	200	}
	201
	202	static struct xor_block_template xor_block_neon = {
	203	.name = "neon",
	204	.do_2 = xor_neon_2,
	205	.do_3 = xor_neon_3,
	206	.do_4 = xor_neon_4,
	207	.do_5 = xor_neon_5
	208	};
	209
210	#define NEON_TEMPLATES \
211	do { if (cpu_has_neon()) xor_speed(&xor_block_neon); } while (0)
212	#else
213	#define NEON_TEMPLATES
214	#endif