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
|
/* -*- mode: c tab-width: 2; c-basic-indent: 2; indent-tabs-mode: nil -*-
*
* The GIMP -- an image manipulation program
* Copyright (C) 1995 Spencer Kimball and Peter Mattis
*
*
* 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.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
*/
#if __GNUC__ >= 3
/*
* Convert the low 8bit byte of the src to 16bit words in dst.
*/
#define mmx_low_bytes_to_words(src,dst,zero) \
"\tmovq %%"#src", %%"#dst"; " \
"\tpunpcklbw %%"#zero", %%"#dst"\n"
/*
* Convert the high 8bit byte of the src to 16bit words in dst.
*/
#define mmx_high_bytes_to_words(src,dst,zero) \
"\tmovq %%"#src", %%"#dst"; " \
"\tpunpckhbw %%"#zero", %%"#dst"\n"
#define xmm_low_bytes_to_words(src,dst,zero) \
"\tmovdqu %%"#src", %%"#dst"; " \
"\tpunpcklbw %%"#zero", %%"#dst"\n"
#define xmm_high_bytes_to_words(src,dst,zero) \
"\tmovdqu %%"#src", %%"#dst"; " \
"\tpunpckhbw %%"#zero", %%"#dst"\n"
/* a = INT_MULT(a,b) */
#define mmx_int_mult(a,b,w128) \
"\tpmullw %%"#b", %%"#a"; " \
"\tpaddw %%"#w128", %%"#a"; " \
"\tmovq %%"#a", %%"#b"; " \
"\tpsrlw $8, %%"#b"; " \
"\tpaddw %%"#a", %%"#b"; " \
"\tpsrlw $8, %%"#b"\n"
#define sse2_int_mult(a,b,w128) \
"\tpmullw %%"#b", %%"#a"; " \
"\tpaddw %%"#w128", %%"#a"; " \
"\tmovdqu %%"#a", %%"#b"; " \
"\tpsrlw $8, %%"#b"; " \
"\tpaddw %%"#a", %%"#b"; " \
"\tpsrlw $8, %%"#b"\n"
/*
* Double-word divide. Adjusted for subsequent unsigned packing
* (high-order bit of each word is cleared)
* Clobbers eax, ecx edx
*/
#define pdivwX(dividend,divisor,quotient) "movd %%" #dividend ",%%eax; " \
"movd %%" #divisor ",%%ecx; " \
"xorl %%edx,%%edx; " \
"divw %%cx; " \
"roll $16, %%eax; " \
"roll $16, %%ecx; " \
"xorl %%edx,%%edx; " \
"divw %%cx; " \
"btr $15, %%eax; " \
"roll $16, %%eax; " \
"btr $15, %%eax; " \
"movd %%eax,%%" #quotient ";"
/*
* Quadword divide. No adjustment for subsequent unsigned packing
* (high-order bit of each word is left alone)
* clobber list must include: "%eax", "%ecx", "%edx", divisor quotient
*/
#define pdivwqX(dividend,divisor,quotient) "movd %%" #dividend ",%%eax; " \
"movd %%" #divisor ",%%ecx; " \
"xorl %%edx,%%edx; " \
"divw %%cx; " \
"roll $16, %%eax; " \
"roll $16, %%ecx; " \
"xorl %%edx,%%edx; " \
"divw %%cx; " \
"roll $16, %%eax; " \
"movd %%eax,%%" #quotient "; " \
"psrlq $32,%%" #dividend ";" \
"psrlq $32,%%" #divisor ";" \
"movd %%" #dividend ",%%eax; " \
"movd %%" #divisor ",%%ecx; " \
"xorl %%edx,%%edx; " \
"divw %%cx; " \
"roll $16, %%eax; " \
"roll $16, %%ecx; " \
"xorl %%edx,%%edx; " \
"divw %%cx; " \
"roll $16, %%eax; " \
"movd %%eax,%%" #divisor ";" \
"psllq $32,%%" #divisor ";" \
"por %%" #divisor ",%%" #quotient ";"
#define pdivwqX_clobber "%eax", "%ecx", "%edx", "%cc"
/*
* Quadword divide. Adjusted for subsequent unsigned packing
* (high-order bit of each word is cleared)
*/
#define pdivwuqX(dividend,divisor,quotient) \
pdivwX(dividend,divisor,quotient) \
"psrlq $32,%%" #dividend ";" \
"psrlq $32,%%" #divisor ";" \
pdivwX(dividend,divisor,quotient) \
"movd %%eax,%%" #divisor ";" \
"psllq $32,%%" #divisor ";" \
"por %%" #divisor ",%%" #quotient ";"
#define pdivwuqX_clobber pdivwqX_clobber
#define xmm_pdivwqX(dividend,divisor,quotient,scratch) "movd %%" #dividend ",%%eax; " \
"movd %%" #divisor ",%%ecx; " \
"xorl %%edx,%%edx; " \
"divw %%cx; " \
"roll $16, %%eax; " \
"roll $16, %%ecx; " \
"xorl %%edx,%%edx; " \
"divw %%cx; " \
"roll $16, %%eax; " \
"movd %%eax,%%" #quotient "; " \
"psrlq $32,%%" #divisor ";" \
"psrlq $32,%%" #dividend ";" \
"movd %%" #dividend ",%%eax; " \
"movd %%" #divisor ",%%ecx; " \
"xorl %%edx,%%edx; " \
"divw %%cx; " \
"roll $16, %%eax; " \
"roll $16, %%ecx; " \
"xorl %%edx,%%edx; " \
"divw %%cx; " \
"roll $16, %%eax; " \
"movd %%eax,%%" #scratch ";" \
"psllq $32,%%" #scratch ";" \
"psrlq $32,%%" #divisor ";" \
"psrlq $32,%%" #dividend ";" \
"movd %%" #dividend ",%%eax; " \
"movd %%" #divisor ",%%ecx; " \
"xorl %%edx,%%edx; " \
"divw %%cx; " \
"roll $16, %%eax; " \
"roll $16, %%ecx; " \
"xorl %%edx,%%edx; " \
"divw %%cx; " \
"roll $16, %%eax; " \
"movd %%eax,%%" #scratch ";" \
"psllq $64,%%" #scratch ";" \
"psrlq $32,%%" #divisor ";" \
"psrlq $32,%%" #dividend ";" \
"movd %%" #dividend ",%%eax; " \
"movd %%" #divisor ",%%ecx; " \
"xorl %%edx,%%edx; " \
"divw %%cx; " \
"roll $16, %%eax; " \
"roll $16, %%ecx; " \
"xorl %%edx,%%edx; " \
"divw %%cx; " \
"roll $16, %%eax; " \
"movd %%eax,%%" #scratch ";" \
"psllq $96,%%" #scratch ";" \
"por %%" #scratch ",%%" #quotient ";"
#define xmm_pdivwX(dividend,divisor,quotient) "movd %%" #dividend ",%%eax; " \
"movd %%" #divisor ",%%ecx; " \
"xorl %%edx,%%edx; " \
"divw %%cx; " \
"roll $16, %%eax; " \
"roll $16, %%ecx; " \
"xorl %%edx,%%edx; " \
"divw %%cx; " \
"btr $15, %%eax; " \
"roll $16, %%eax; " \
"btr $15, %%eax; " \
"movd %%eax,%%" #quotient ";"
#define xmm_pdivwuqX(dividend,divisor,quotient,scratch) \
xmm_pdivwX(dividend,divisor,scratch) \
"movd %%"#scratch ",%%"#quotient ";" \
"psrlq $32,%%"#dividend ";" \
"psrlq $32,%%"#divisor ";" \
xmm_pdivwX(dividend,divisor,scratch) \
"psllq $32,%%"#scratch ";" \
"por %%"#scratch ",%%"#quotient ";" \
"psrlq $32,%%"#dividend ";" \
"psrlq $32,%%"#divisor ";" \
xmm_pdivwX(dividend,divisor,scratch) \
"psllq $64,%%"#scratch ";" \
"por %%"#scratch ",%%"#quotient ";" \
"psrlq $32,%%"#dividend ";" \
"psrlq $32,%%"#divisor ";" \
xmm_pdivwX(dividend,divisor,scratch) \
"psllq $96,%%"#scratch ";" \
"por %%"#scratch ",%%"#quotient
/* equivalent to the INT_MULT() macro in gimp-composite-generic.c */
/*
* opr2 = INT_MULT(opr1, opr2, t)
*
* Operates across quad-words using x86 word (16bit) value.
* Result is left in opr2
*
* opr1 = opr1 * opr2 + w128
* opr2 = opr1
* opr2 = ((opr2 >> 8) + opr1) >> 8
*/
#define pmulwX(opr1,opr2,w128) \
"\tpmullw %%"#opr2", %%"#opr1"; " \
"\tpaddw %%"#w128", %%"#opr1"; " \
"\tmovq %%"#opr1", %%"#opr2"; " \
"\tpsrlw $8, %%"#opr2"; " \
"\tpaddw %%"#opr1", %%"#opr2"; " \
"\tpsrlw $8, %%"#opr2"\n"
#define xmm_pmulwX(opr1,opr2,w128) \
"\tpmullw %%"#opr2", %%"#opr1"; " \
"\tpaddw %%"#w128", %%"#opr1"; " \
"\tmovdqu %%"#opr1", %%"#opr2"; " \
"\tpsrlw $8, %%"#opr2"; " \
"\tpaddw %%"#opr1", %%"#opr2"; " \
"\tpsrlw $8, %%"#opr2"\n"
typedef unsigned char uint8;
typedef unsigned short uint16;
typedef unsigned long uint32;
typedef unsigned long long uint64;
typedef struct { uint64 __uint64[2]; } uint128;
extern const guint32 va8_alpha_mask[2];
extern const guint32 va8_b255[2];
extern const guint32 va8_w1[2];
extern const guint32 va8_w255[2];
#endif /* __GNUC__ >= 3 */
|
