summaryrefslogtreecommitdiffstatshomepage
path: root/src/osd/eigccppc.h
blob: 7d61ee497520c017f0c9979a8a413eedf93e1ef4 (plain) (blame)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
// license:BSD-3-Clause
// copyright-holders:Vas Crabb
/***************************************************************************

    eigccppc.h

    PowerPC (32 and 64-bit) inline implementations for GCC compilers. This
    code is automatically included if appropriate by eminline.h.

***************************************************************************/

#ifndef __EIGCCPPC__
#define __EIGCCPPC__


/***************************************************************************
    INLINE MATH FUNCTIONS
***************************************************************************/

/*-------------------------------------------------
    mul_32x32 - perform a signed 32 bit x 32 bit
    multiply and return the full 64 bit result
-------------------------------------------------*/

/* GCC can do a good job of this. */


/*-------------------------------------------------
    mulu_32x32 - perform an unsigned 32 bit x
    32 bit multiply and return the full 64 bit
    result
-------------------------------------------------*/

/* GCC can do a good job of this */


/*-------------------------------------------------
    mul_32x32_hi - perform a signed 32 bit x 32 bit
    multiply and return the upper 32 bits of the
    result
-------------------------------------------------*/

#define mul_32x32_hi _mul_32x32_hi
static inline INT32 ATTR_CONST ATTR_FORCE_INLINE
_mul_32x32_hi(INT32 val1, INT32 val2)
{
	INT32 result;

	__asm__ (
		" mulhw  %[result], %[val1], %[val2] \n"
		: [result] "=r" (result)
		: [val1]   "%r" (val1)
		, [val2]   "r"  (val2)
	);

	return result;
}


/*-------------------------------------------------
    mulu_32x32_hi - perform an unsigned 32 bit x
    32 bit multiply and return the upper 32 bits
    of the result
-------------------------------------------------*/

#define mulu_32x32_hi _mulu_32x32_hi
static inline UINT32 ATTR_CONST ATTR_FORCE_INLINE
_mulu_32x32_hi(UINT32 val1, UINT32 val2)
{
	UINT32 result;

	__asm__ (
		" mulhwu  %[result], %[val1], %[val2] \n"
		: [result] "=r" (result)
		: [val1]   "%r" (val1)
		, [val2]   "r"  (val2)
	);

	return result;
}


/*-------------------------------------------------
    mul_32x32_shift - perform a signed 32 bit x
    32 bit multiply and shift the result by the
    given number of bits before truncating the
    result to 32 bits
-------------------------------------------------*/

#if !defined(__ppc64__) && !defined(__PPC64__) && !defined(_ARCH_PPC64)
#define mul_32x32_shift _mul_32x32_shift
static inline INT32 ATTR_CONST ATTR_FORCE_INLINE
_mul_32x32_shift(INT32 val1, INT32 val2, UINT8 shift)
{
	INT32 result;

	/* Valid for (0 <= shift <= 32) */
	__asm__ (
		" mullw   %[result], %[val1], %[val2]    \n"
		" mulhw   %[val1], %[val1], %[val2]      \n"
		" srw     %[result], %[result], %[shift] \n"
		" subfic  %[shift], %[shift], 0x20       \n"
		" slw     %[val1], %[val1], %[shift]     \n"
		" or      %[result], %[result], %[val1]  \n"
		: [result] "=&r" (result)
		, [shift]  "+r"  (shift)
		, [val1]   "+r"  (val1)
		: [val2]   "r"   (val2)
		: "xer"
	);

	return result;
}
#endif


/*-------------------------------------------------
    mulu_32x32_shift - perform an unsigned 32 bit x
    32 bit multiply and shift the result by the
    given number of bits before truncating the
    result to 32 bits
-------------------------------------------------*/

#if !defined(__ppc64__) && !defined(__PPC64__) && !defined(_ARCH_PPC64)
#define mulu_32x32_shift _mulu_32x32_shift
static inline UINT32 ATTR_CONST ATTR_FORCE_INLINE
_mulu_32x32_shift(UINT32 val1, UINT32 val2, UINT8 shift)
{
	UINT32 result;

	/* Valid for (0 <= shift <= 32) */
	__asm__ (
		" mullw   %[result], %[val1], %[val2]    \n"
		" mulhwu  %[val1], %[val1], %[val2]      \n"
		" srw     %[result], %[result], %[shift] \n"
		" subfic  %[shift], %[shift], 0x20       \n"
		" slw     %[val1], %[val1], %[shift]     \n"
		" or      %[result], %[result], %[val1]  \n"
		: [result] "=&r" (result)
		, [shift]  "+r"  (shift)
		, [val1]   "+r"  (val1)
		: [val2]   "r"   (val2)
		: "xer"
	);

	return result;
}
#endif


/*-------------------------------------------------
    div_64x32 - perform a signed 64 bit x 32 bit
    divide and return the 32 bit quotient
-------------------------------------------------*/

/* TBD */


/*-------------------------------------------------
    divu_64x32 - perform an unsigned 64 bit x 32 bit
    divide and return the 32 bit quotient
-------------------------------------------------*/

/* TBD */


/*-------------------------------------------------
    div_64x32_rem - perform a signed 64 bit x 32
    bit divide and return the 32 bit quotient and
    32 bit remainder
-------------------------------------------------*/

/* TBD */


/*-------------------------------------------------
    divu_64x32_rem - perform an unsigned 64 bit x
    32 bit divide and return the 32 bit quotient
    and 32 bit remainder
-------------------------------------------------*/

/* TBD */


/*-------------------------------------------------
    div_32x32_shift - perform a signed divide of
    two 32 bit values, shifting the first before
    division, and returning the 32 bit quotient
-------------------------------------------------*/

/* TBD */


/*-------------------------------------------------
    divu_32x32_shift - perform an unsigned divide of
    two 32 bit values, shifting the first before
    division, and returning the 32 bit quotient
-------------------------------------------------*/

/* TBD */


/*-------------------------------------------------
    mod_64x32 - perform a signed 64 bit x 32 bit
    divide and return the 32 bit remainder
-------------------------------------------------*/

/* TBD */


/*-------------------------------------------------
    modu_64x32 - perform an unsigned 64 bit x 32 bit
    divide and return the 32 bit remainder
-------------------------------------------------*/

/* TBD */


/*-------------------------------------------------
    recip_approx - compute an approximate floating
    point reciprocal
-------------------------------------------------*/

#define recip_approx _recip_approx
static inline float ATTR_CONST ATTR_FORCE_INLINE
_recip_approx(float value)
{
	float result;

	__asm__ (
		" fres  %[result], %[value] \n"
		: [result] "=f" (result)
		: [value]  "f"  (value)
	);

	return result;
}



/***************************************************************************
    INLINE BIT MANIPULATION FUNCTIONS
***************************************************************************/

/*-------------------------------------------------
    count_leading_zeros - return the number of
    leading zero bits in a 32-bit value
-------------------------------------------------*/

#define count_leading_zeros _count_leading_zeros
static inline UINT8 ATTR_CONST ATTR_FORCE_INLINE
_count_leading_zeros(UINT32 value)
{
	UINT32 result;

	__asm__ (
		" cntlzw  %[result], %[value] \n"
		: [result] "=r" (result)    /* result can be in any register */
		: [value]  "r"  (value)     /* 'value' can be in any register */
	);

	return result;
}


/*-------------------------------------------------
    count_leading_ones - return the number of
    leading one bits in a 32-bit value
-------------------------------------------------*/

#define count_leading_ones _count_leading_ones
static inline UINT8 ATTR_CONST ATTR_FORCE_INLINE
_count_leading_ones(UINT32 value)
{
	UINT32 result;

	__asm__ (
		" not     %[result], %[value]  \n"
		" cntlzw  %[result], %[result] \n"
		: [result] "=r" (result)    /* result can be in any register */
		: [value]  "r"  (value)     /* 'value' can be in any register */
	);

	return result;
}

#endif /* __EIGCCPPC__ */