|
1
|
1 /*
|
|
|
2 * mmx.h
|
|
36
|
3 * Copyright (C) 1997-2001 H. Dietz and R. Fisher
|
|
1
|
4 *
|
|
|
5 * This file is part of mpeg2dec, a free MPEG-2 video stream decoder.
|
|
|
6 *
|
|
|
7 * mpeg2dec is free software; you can redistribute it and/or modify
|
|
|
8 * it under the terms of the GNU General Public License as published by
|
|
|
9 * the Free Software Foundation; either version 2 of the License, or
|
|
|
10 * (at your option) any later version.
|
|
|
11 *
|
|
|
12 * mpeg2dec is distributed in the hope that it will be useful,
|
|
|
13 * but WITHOUT ANY WARRANTY; without even the implied warranty of
|
|
|
14 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
|
|
15 * GNU General Public License for more details.
|
|
|
16 *
|
|
|
17 * You should have received a copy of the GNU General Public License
|
|
|
18 * along with this program; if not, write to the Free Software
|
|
|
19 * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
|
|
|
20 */
|
|
|
21
|
|
|
22 /*
|
|
|
23 * The type of an value that fits in an MMX register (note that long
|
|
|
24 * long constant values MUST be suffixed by LL and unsigned long long
|
|
|
25 * values by ULL, lest they be truncated by the compiler)
|
|
|
26 */
|
|
|
27
|
|
|
28 typedef union {
|
|
|
29 long long q; /* Quadword (64-bit) value */
|
|
|
30 unsigned long long uq; /* Unsigned Quadword */
|
|
|
31 int d[2]; /* 2 Doubleword (32-bit) values */
|
|
|
32 unsigned int ud[2]; /* 2 Unsigned Doubleword */
|
|
|
33 short w[4]; /* 4 Word (16-bit) values */
|
|
|
34 unsigned short uw[4]; /* 4 Unsigned Word */
|
|
|
35 char b[8]; /* 8 Byte (8-bit) values */
|
|
|
36 unsigned char ub[8]; /* 8 Unsigned Byte */
|
|
|
37 float s[2]; /* Single-precision (32-bit) value */
|
|
|
38 } ATTR_ALIGN(8) mmx_t; /* On an 8-byte (64-bit) boundary */
|
|
|
39
|
|
|
40
|
|
|
41 #define mmx_i2r(op,imm,reg) \
|
|
|
42 __asm__ __volatile__ (#op " %0, %%" #reg \
|
|
|
43 : /* nothing */ \
|
|
36
|
44 : "i" (imm) )
|
|
1
|
45
|
|
|
46 #define mmx_m2r(op,mem,reg) \
|
|
|
47 __asm__ __volatile__ (#op " %0, %%" #reg \
|
|
|
48 : /* nothing */ \
|
|
36
|
49 : "m" (mem))
|
|
1
|
50
|
|
|
51 #define mmx_r2m(op,reg,mem) \
|
|
|
52 __asm__ __volatile__ (#op " %%" #reg ", %0" \
|
|
36
|
53 : "=m" (mem) \
|
|
1
|
54 : /* nothing */ )
|
|
|
55
|
|
|
56 #define mmx_r2r(op,regs,regd) \
|
|
|
57 __asm__ __volatile__ (#op " %" #regs ", %" #regd)
|
|
|
58
|
|
|
59
|
|
|
60 #define emms() __asm__ __volatile__ ("emms")
|
|
|
61
|
|
|
62 #define movd_m2r(var,reg) mmx_m2r (movd, var, reg)
|
|
|
63 #define movd_r2m(reg,var) mmx_r2m (movd, reg, var)
|
|
|
64 #define movd_r2r(regs,regd) mmx_r2r (movd, regs, regd)
|
|
|
65
|
|
|
66 #define movq_m2r(var,reg) mmx_m2r (movq, var, reg)
|
|
|
67 #define movq_r2m(reg,var) mmx_r2m (movq, reg, var)
|
|
|
68 #define movq_r2r(regs,regd) mmx_r2r (movq, regs, regd)
|
|
|
69
|
|
|
70 #define packssdw_m2r(var,reg) mmx_m2r (packssdw, var, reg)
|
|
|
71 #define packssdw_r2r(regs,regd) mmx_r2r (packssdw, regs, regd)
|
|
|
72 #define packsswb_m2r(var,reg) mmx_m2r (packsswb, var, reg)
|
|
|
73 #define packsswb_r2r(regs,regd) mmx_r2r (packsswb, regs, regd)
|
|
|
74
|
|
|
75 #define packuswb_m2r(var,reg) mmx_m2r (packuswb, var, reg)
|
|
|
76 #define packuswb_r2r(regs,regd) mmx_r2r (packuswb, regs, regd)
|
|
|
77
|
|
|
78 #define paddb_m2r(var,reg) mmx_m2r (paddb, var, reg)
|
|
|
79 #define paddb_r2r(regs,regd) mmx_r2r (paddb, regs, regd)
|
|
|
80 #define paddd_m2r(var,reg) mmx_m2r (paddd, var, reg)
|
|
|
81 #define paddd_r2r(regs,regd) mmx_r2r (paddd, regs, regd)
|
|
|
82 #define paddw_m2r(var,reg) mmx_m2r (paddw, var, reg)
|
|
|
83 #define paddw_r2r(regs,regd) mmx_r2r (paddw, regs, regd)
|
|
|
84
|
|
|
85 #define paddsb_m2r(var,reg) mmx_m2r (paddsb, var, reg)
|
|
|
86 #define paddsb_r2r(regs,regd) mmx_r2r (paddsb, regs, regd)
|
|
|
87 #define paddsw_m2r(var,reg) mmx_m2r (paddsw, var, reg)
|
|
|
88 #define paddsw_r2r(regs,regd) mmx_r2r (paddsw, regs, regd)
|
|
|
89
|
|
|
90 #define paddusb_m2r(var,reg) mmx_m2r (paddusb, var, reg)
|
|
|
91 #define paddusb_r2r(regs,regd) mmx_r2r (paddusb, regs, regd)
|
|
|
92 #define paddusw_m2r(var,reg) mmx_m2r (paddusw, var, reg)
|
|
|
93 #define paddusw_r2r(regs,regd) mmx_r2r (paddusw, regs, regd)
|
|
|
94
|
|
|
95 #define pand_m2r(var,reg) mmx_m2r (pand, var, reg)
|
|
|
96 #define pand_r2r(regs,regd) mmx_r2r (pand, regs, regd)
|
|
|
97
|
|
|
98 #define pandn_m2r(var,reg) mmx_m2r (pandn, var, reg)
|
|
|
99 #define pandn_r2r(regs,regd) mmx_r2r (pandn, regs, regd)
|
|
|
100
|
|
|
101 #define pcmpeqb_m2r(var,reg) mmx_m2r (pcmpeqb, var, reg)
|
|
|
102 #define pcmpeqb_r2r(regs,regd) mmx_r2r (pcmpeqb, regs, regd)
|
|
|
103 #define pcmpeqd_m2r(var,reg) mmx_m2r (pcmpeqd, var, reg)
|
|
|
104 #define pcmpeqd_r2r(regs,regd) mmx_r2r (pcmpeqd, regs, regd)
|
|
|
105 #define pcmpeqw_m2r(var,reg) mmx_m2r (pcmpeqw, var, reg)
|
|
|
106 #define pcmpeqw_r2r(regs,regd) mmx_r2r (pcmpeqw, regs, regd)
|
|
|
107
|
|
|
108 #define pcmpgtb_m2r(var,reg) mmx_m2r (pcmpgtb, var, reg)
|
|
|
109 #define pcmpgtb_r2r(regs,regd) mmx_r2r (pcmpgtb, regs, regd)
|
|
|
110 #define pcmpgtd_m2r(var,reg) mmx_m2r (pcmpgtd, var, reg)
|
|
|
111 #define pcmpgtd_r2r(regs,regd) mmx_r2r (pcmpgtd, regs, regd)
|
|
|
112 #define pcmpgtw_m2r(var,reg) mmx_m2r (pcmpgtw, var, reg)
|
|
|
113 #define pcmpgtw_r2r(regs,regd) mmx_r2r (pcmpgtw, regs, regd)
|
|
|
114
|
|
|
115 #define pmaddwd_m2r(var,reg) mmx_m2r (pmaddwd, var, reg)
|
|
|
116 #define pmaddwd_r2r(regs,regd) mmx_r2r (pmaddwd, regs, regd)
|
|
|
117
|
|
|
118 #define pmulhw_m2r(var,reg) mmx_m2r (pmulhw, var, reg)
|
|
|
119 #define pmulhw_r2r(regs,regd) mmx_r2r (pmulhw, regs, regd)
|
|
|
120
|
|
|
121 #define pmullw_m2r(var,reg) mmx_m2r (pmullw, var, reg)
|
|
|
122 #define pmullw_r2r(regs,regd) mmx_r2r (pmullw, regs, regd)
|
|
|
123
|
|
|
124 #define por_m2r(var,reg) mmx_m2r (por, var, reg)
|
|
|
125 #define por_r2r(regs,regd) mmx_r2r (por, regs, regd)
|
|
|
126
|
|
|
127 #define pslld_i2r(imm,reg) mmx_i2r (pslld, imm, reg)
|
|
|
128 #define pslld_m2r(var,reg) mmx_m2r (pslld, var, reg)
|
|
|
129 #define pslld_r2r(regs,regd) mmx_r2r (pslld, regs, regd)
|
|
|
130 #define psllq_i2r(imm,reg) mmx_i2r (psllq, imm, reg)
|
|
|
131 #define psllq_m2r(var,reg) mmx_m2r (psllq, var, reg)
|
|
|
132 #define psllq_r2r(regs,regd) mmx_r2r (psllq, regs, regd)
|
|
|
133 #define psllw_i2r(imm,reg) mmx_i2r (psllw, imm, reg)
|
|
|
134 #define psllw_m2r(var,reg) mmx_m2r (psllw, var, reg)
|
|
|
135 #define psllw_r2r(regs,regd) mmx_r2r (psllw, regs, regd)
|
|
|
136
|
|
|
137 #define psrad_i2r(imm,reg) mmx_i2r (psrad, imm, reg)
|
|
|
138 #define psrad_m2r(var,reg) mmx_m2r (psrad, var, reg)
|
|
|
139 #define psrad_r2r(regs,regd) mmx_r2r (psrad, regs, regd)
|
|
|
140 #define psraw_i2r(imm,reg) mmx_i2r (psraw, imm, reg)
|
|
|
141 #define psraw_m2r(var,reg) mmx_m2r (psraw, var, reg)
|
|
|
142 #define psraw_r2r(regs,regd) mmx_r2r (psraw, regs, regd)
|
|
|
143
|
|
|
144 #define psrld_i2r(imm,reg) mmx_i2r (psrld, imm, reg)
|
|
|
145 #define psrld_m2r(var,reg) mmx_m2r (psrld, var, reg)
|
|
|
146 #define psrld_r2r(regs,regd) mmx_r2r (psrld, regs, regd)
|
|
|
147 #define psrlq_i2r(imm,reg) mmx_i2r (psrlq, imm, reg)
|
|
|
148 #define psrlq_m2r(var,reg) mmx_m2r (psrlq, var, reg)
|
|
|
149 #define psrlq_r2r(regs,regd) mmx_r2r (psrlq, regs, regd)
|
|
|
150 #define psrlw_i2r(imm,reg) mmx_i2r (psrlw, imm, reg)
|
|
|
151 #define psrlw_m2r(var,reg) mmx_m2r (psrlw, var, reg)
|
|
|
152 #define psrlw_r2r(regs,regd) mmx_r2r (psrlw, regs, regd)
|
|
|
153
|
|
|
154 #define psubb_m2r(var,reg) mmx_m2r (psubb, var, reg)
|
|
|
155 #define psubb_r2r(regs,regd) mmx_r2r (psubb, regs, regd)
|
|
|
156 #define psubd_m2r(var,reg) mmx_m2r (psubd, var, reg)
|
|
|
157 #define psubd_r2r(regs,regd) mmx_r2r (psubd, regs, regd)
|
|
|
158 #define psubw_m2r(var,reg) mmx_m2r (psubw, var, reg)
|
|
|
159 #define psubw_r2r(regs,regd) mmx_r2r (psubw, regs, regd)
|
|
|
160
|
|
|
161 #define psubsb_m2r(var,reg) mmx_m2r (psubsb, var, reg)
|
|
|
162 #define psubsb_r2r(regs,regd) mmx_r2r (psubsb, regs, regd)
|
|
|
163 #define psubsw_m2r(var,reg) mmx_m2r (psubsw, var, reg)
|
|
|
164 #define psubsw_r2r(regs,regd) mmx_r2r (psubsw, regs, regd)
|
|
|
165
|
|
|
166 #define psubusb_m2r(var,reg) mmx_m2r (psubusb, var, reg)
|
|
|
167 #define psubusb_r2r(regs,regd) mmx_r2r (psubusb, regs, regd)
|
|
|
168 #define psubusw_m2r(var,reg) mmx_m2r (psubusw, var, reg)
|
|
|
169 #define psubusw_r2r(regs,regd) mmx_r2r (psubusw, regs, regd)
|
|
|
170
|
|
|
171 #define punpckhbw_m2r(var,reg) mmx_m2r (punpckhbw, var, reg)
|
|
|
172 #define punpckhbw_r2r(regs,regd) mmx_r2r (punpckhbw, regs, regd)
|
|
|
173 #define punpckhdq_m2r(var,reg) mmx_m2r (punpckhdq, var, reg)
|
|
|
174 #define punpckhdq_r2r(regs,regd) mmx_r2r (punpckhdq, regs, regd)
|
|
|
175 #define punpckhwd_m2r(var,reg) mmx_m2r (punpckhwd, var, reg)
|
|
|
176 #define punpckhwd_r2r(regs,regd) mmx_r2r (punpckhwd, regs, regd)
|
|
|
177
|
|
|
178 #define punpcklbw_m2r(var,reg) mmx_m2r (punpcklbw, var, reg)
|
|
|
179 #define punpcklbw_r2r(regs,regd) mmx_r2r (punpcklbw, regs, regd)
|
|
|
180 #define punpckldq_m2r(var,reg) mmx_m2r (punpckldq, var, reg)
|
|
|
181 #define punpckldq_r2r(regs,regd) mmx_r2r (punpckldq, regs, regd)
|
|
|
182 #define punpcklwd_m2r(var,reg) mmx_m2r (punpcklwd, var, reg)
|
|
|
183 #define punpcklwd_r2r(regs,regd) mmx_r2r (punpcklwd, regs, regd)
|
|
|
184
|
|
|
185 #define pxor_m2r(var,reg) mmx_m2r (pxor, var, reg)
|
|
|
186 #define pxor_r2r(regs,regd) mmx_r2r (pxor, regs, regd)
|
|
|
187
|
|
|
188
|
|
|
189 /* 3DNOW extensions */
|
|
|
190
|
|
|
191 #define pavgusb_m2r(var,reg) mmx_m2r (pavgusb, var, reg)
|
|
|
192 #define pavgusb_r2r(regs,regd) mmx_r2r (pavgusb, regs, regd)
|
|
|
193
|
|
|
194
|
|
|
195 /* AMD MMX extensions - also available in intel SSE */
|
|
|
196
|
|
|
197
|
|
|
198 #define mmx_m2ri(op,mem,reg,imm) \
|
|
|
199 __asm__ __volatile__ (#op " %1, %0, %%" #reg \
|
|
|
200 : /* nothing */ \
|
|
|
201 : "X" (mem), "X" (imm))
|
|
|
202 #define mmx_r2ri(op,regs,regd,imm) \
|
|
|
203 __asm__ __volatile__ (#op " %0, %%" #regs ", %%" #regd \
|
|
|
204 : /* nothing */ \
|
|
|
205 : "X" (imm) )
|
|
|
206
|
|
|
207 #define mmx_fetch(mem,hint) \
|
|
|
208 __asm__ __volatile__ ("prefetch" #hint " %0" \
|
|
|
209 : /* nothing */ \
|
|
|
210 : "X" (mem))
|
|
|
211
|
|
|
212
|
|
|
213 #define maskmovq(regs,maskreg) mmx_r2ri (maskmovq, regs, maskreg)
|
|
|
214
|
|
|
215 #define movntq_r2m(mmreg,var) mmx_r2m (movntq, mmreg, var)
|
|
|
216
|
|
|
217 #define pavgb_m2r(var,reg) mmx_m2r (pavgb, var, reg)
|
|
|
218 #define pavgb_r2r(regs,regd) mmx_r2r (pavgb, regs, regd)
|
|
|
219 #define pavgw_m2r(var,reg) mmx_m2r (pavgw, var, reg)
|
|
|
220 #define pavgw_r2r(regs,regd) mmx_r2r (pavgw, regs, regd)
|
|
|
221
|
|
|
222 #define pextrw_r2r(mmreg,reg,imm) mmx_r2ri (pextrw, mmreg, reg, imm)
|
|
|
223
|
|
|
224 #define pinsrw_r2r(reg,mmreg,imm) mmx_r2ri (pinsrw, reg, mmreg, imm)
|
|
|
225
|
|
|
226 #define pmaxsw_m2r(var,reg) mmx_m2r (pmaxsw, var, reg)
|
|
|
227 #define pmaxsw_r2r(regs,regd) mmx_r2r (pmaxsw, regs, regd)
|
|
|
228
|
|
|
229 #define pmaxub_m2r(var,reg) mmx_m2r (pmaxub, var, reg)
|
|
|
230 #define pmaxub_r2r(regs,regd) mmx_r2r (pmaxub, regs, regd)
|
|
|
231
|
|
|
232 #define pminsw_m2r(var,reg) mmx_m2r (pminsw, var, reg)
|
|
|
233 #define pminsw_r2r(regs,regd) mmx_r2r (pminsw, regs, regd)
|
|
|
234
|
|
|
235 #define pminub_m2r(var,reg) mmx_m2r (pminub, var, reg)
|
|
|
236 #define pminub_r2r(regs,regd) mmx_r2r (pminub, regs, regd)
|
|
|
237
|
|
|
238 #define pmovmskb(mmreg,reg) \
|
|
|
239 __asm__ __volatile__ ("movmskps %" #mmreg ", %" #reg)
|
|
|
240
|
|
|
241 #define pmulhuw_m2r(var,reg) mmx_m2r (pmulhuw, var, reg)
|
|
|
242 #define pmulhuw_r2r(regs,regd) mmx_r2r (pmulhuw, regs, regd)
|
|
|
243
|
|
|
244 #define prefetcht0(mem) mmx_fetch (mem, t0)
|
|
|
245 #define prefetcht1(mem) mmx_fetch (mem, t1)
|
|
|
246 #define prefetcht2(mem) mmx_fetch (mem, t2)
|
|
|
247 #define prefetchnta(mem) mmx_fetch (mem, nta)
|
|
|
248
|
|
|
249 #define psadbw_m2r(var,reg) mmx_m2r (psadbw, var, reg)
|
|
|
250 #define psadbw_r2r(regs,regd) mmx_r2r (psadbw, regs, regd)
|
|
|
251
|
|
|
252 #define pshufw_m2r(var,reg,imm) mmx_m2ri(pshufw, var, reg, imm)
|
|
|
253 #define pshufw_r2r(regs,regd,imm) mmx_r2ri(pshufw, regs, regd, imm)
|
|
|
254
|
|
|
255 #define sfence() __asm__ __volatile__ ("sfence\n\t")
|
