Mercurial > mplayer.hg
annotate libmpeg2/mmx.h @ 26552:806697d15e5c
Mark phony checkheaders target as such.
author | diego |
---|---|
date | Mon, 28 Apr 2008 17:57:44 +0000 |
parents | 2506f1b0bdbe |
children | da2271c341ee |
rev | line source |
---|---|
1 | 1 /* |
2 * mmx.h | |
12932 | 3 * Copyright (C) 2000-2003 Michel Lespinasse <walken@zoy.org> |
9852 | 4 * Copyright (C) 1999-2000 Aaron Holtzman <aholtzma@ess.engr.uvic.ca> |
1 | 5 * |
6 * This file is part of mpeg2dec, a free MPEG-2 video stream decoder. | |
9852 | 7 * See http://libmpeg2.sourceforge.net/ for updates. |
1 | 8 * |
9 * mpeg2dec is free software; you can redistribute it and/or modify | |
10 * it under the terms of the GNU General Public License as published by | |
11 * the Free Software Foundation; either version 2 of the License, or | |
12 * (at your option) any later version. | |
13 * | |
14 * mpeg2dec is distributed in the hope that it will be useful, | |
15 * but WITHOUT ANY WARRANTY; without even the implied warranty of | |
16 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the | |
17 * GNU General Public License for more details. | |
18 * | |
19 * You should have received a copy of the GNU General Public License | |
20 * along with this program; if not, write to the Free Software | |
21 * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA | |
22 */ | |
23 | |
24 /* | |
25 * The type of an value that fits in an MMX register (note that long | |
26 * long constant values MUST be suffixed by LL and unsigned long long | |
27 * values by ULL, lest they be truncated by the compiler) | |
28 */ | |
29 | |
30 typedef union { | |
31 long long q; /* Quadword (64-bit) value */ | |
32 unsigned long long uq; /* Unsigned Quadword */ | |
33 int d[2]; /* 2 Doubleword (32-bit) values */ | |
34 unsigned int ud[2]; /* 2 Unsigned Doubleword */ | |
35 short w[4]; /* 4 Word (16-bit) values */ | |
36 unsigned short uw[4]; /* 4 Unsigned Word */ | |
37 char b[8]; /* 8 Byte (8-bit) values */ | |
38 unsigned char ub[8]; /* 8 Unsigned Byte */ | |
39 float s[2]; /* Single-precision (32-bit) value */ | |
40 } ATTR_ALIGN(8) mmx_t; /* On an 8-byte (64-bit) boundary */ | |
41 | |
42 | |
43 #define mmx_i2r(op,imm,reg) \ | |
44 __asm__ __volatile__ (#op " %0, %%" #reg \ | |
45 : /* nothing */ \ | |
36 | 46 : "i" (imm) ) |
1 | 47 |
48 #define mmx_m2r(op,mem,reg) \ | |
49 __asm__ __volatile__ (#op " %0, %%" #reg \ | |
50 : /* nothing */ \ | |
36 | 51 : "m" (mem)) |
1 | 52 |
53 #define mmx_r2m(op,reg,mem) \ | |
54 __asm__ __volatile__ (#op " %%" #reg ", %0" \ | |
36 | 55 : "=m" (mem) \ |
1 | 56 : /* nothing */ ) |
57 | |
58 #define mmx_r2r(op,regs,regd) \ | |
59 __asm__ __volatile__ (#op " %" #regs ", %" #regd) | |
60 | |
61 | |
62 #define emms() __asm__ __volatile__ ("emms") | |
63 | |
64 #define movd_m2r(var,reg) mmx_m2r (movd, var, reg) | |
65 #define movd_r2m(reg,var) mmx_r2m (movd, reg, var) | |
9852 | 66 #define movd_v2r(var,reg) __asm__ __volatile__ ("movd %0, %%" #reg \ |
67 : /* nothing */ \ | |
68 : "rm" (var)) | |
69 #define movd_r2v(reg,var) __asm__ __volatile__ ("movd %%" #reg ", %0" \ | |
70 : "=rm" (var) \ | |
71 : /* nothing */ ) | |
1 | 72 |
73 #define movq_m2r(var,reg) mmx_m2r (movq, var, reg) | |
74 #define movq_r2m(reg,var) mmx_r2m (movq, reg, var) | |
75 #define movq_r2r(regs,regd) mmx_r2r (movq, regs, regd) | |
76 | |
77 #define packssdw_m2r(var,reg) mmx_m2r (packssdw, var, reg) | |
78 #define packssdw_r2r(regs,regd) mmx_r2r (packssdw, regs, regd) | |
79 #define packsswb_m2r(var,reg) mmx_m2r (packsswb, var, reg) | |
80 #define packsswb_r2r(regs,regd) mmx_r2r (packsswb, regs, regd) | |
81 | |
82 #define packuswb_m2r(var,reg) mmx_m2r (packuswb, var, reg) | |
83 #define packuswb_r2r(regs,regd) mmx_r2r (packuswb, regs, regd) | |
84 | |
85 #define paddb_m2r(var,reg) mmx_m2r (paddb, var, reg) | |
86 #define paddb_r2r(regs,regd) mmx_r2r (paddb, regs, regd) | |
87 #define paddd_m2r(var,reg) mmx_m2r (paddd, var, reg) | |
88 #define paddd_r2r(regs,regd) mmx_r2r (paddd, regs, regd) | |
89 #define paddw_m2r(var,reg) mmx_m2r (paddw, var, reg) | |
90 #define paddw_r2r(regs,regd) mmx_r2r (paddw, regs, regd) | |
91 | |
92 #define paddsb_m2r(var,reg) mmx_m2r (paddsb, var, reg) | |
93 #define paddsb_r2r(regs,regd) mmx_r2r (paddsb, regs, regd) | |
94 #define paddsw_m2r(var,reg) mmx_m2r (paddsw, var, reg) | |
95 #define paddsw_r2r(regs,regd) mmx_r2r (paddsw, regs, regd) | |
96 | |
97 #define paddusb_m2r(var,reg) mmx_m2r (paddusb, var, reg) | |
98 #define paddusb_r2r(regs,regd) mmx_r2r (paddusb, regs, regd) | |
99 #define paddusw_m2r(var,reg) mmx_m2r (paddusw, var, reg) | |
100 #define paddusw_r2r(regs,regd) mmx_r2r (paddusw, regs, regd) | |
101 | |
102 #define pand_m2r(var,reg) mmx_m2r (pand, var, reg) | |
103 #define pand_r2r(regs,regd) mmx_r2r (pand, regs, regd) | |
104 | |
105 #define pandn_m2r(var,reg) mmx_m2r (pandn, var, reg) | |
106 #define pandn_r2r(regs,regd) mmx_r2r (pandn, regs, regd) | |
107 | |
108 #define pcmpeqb_m2r(var,reg) mmx_m2r (pcmpeqb, var, reg) | |
109 #define pcmpeqb_r2r(regs,regd) mmx_r2r (pcmpeqb, regs, regd) | |
110 #define pcmpeqd_m2r(var,reg) mmx_m2r (pcmpeqd, var, reg) | |
111 #define pcmpeqd_r2r(regs,regd) mmx_r2r (pcmpeqd, regs, regd) | |
112 #define pcmpeqw_m2r(var,reg) mmx_m2r (pcmpeqw, var, reg) | |
113 #define pcmpeqw_r2r(regs,regd) mmx_r2r (pcmpeqw, regs, regd) | |
114 | |
115 #define pcmpgtb_m2r(var,reg) mmx_m2r (pcmpgtb, var, reg) | |
116 #define pcmpgtb_r2r(regs,regd) mmx_r2r (pcmpgtb, regs, regd) | |
117 #define pcmpgtd_m2r(var,reg) mmx_m2r (pcmpgtd, var, reg) | |
118 #define pcmpgtd_r2r(regs,regd) mmx_r2r (pcmpgtd, regs, regd) | |
119 #define pcmpgtw_m2r(var,reg) mmx_m2r (pcmpgtw, var, reg) | |
120 #define pcmpgtw_r2r(regs,regd) mmx_r2r (pcmpgtw, regs, regd) | |
121 | |
122 #define pmaddwd_m2r(var,reg) mmx_m2r (pmaddwd, var, reg) | |
123 #define pmaddwd_r2r(regs,regd) mmx_r2r (pmaddwd, regs, regd) | |
124 | |
125 #define pmulhw_m2r(var,reg) mmx_m2r (pmulhw, var, reg) | |
126 #define pmulhw_r2r(regs,regd) mmx_r2r (pmulhw, regs, regd) | |
127 | |
128 #define pmullw_m2r(var,reg) mmx_m2r (pmullw, var, reg) | |
129 #define pmullw_r2r(regs,regd) mmx_r2r (pmullw, regs, regd) | |
130 | |
131 #define por_m2r(var,reg) mmx_m2r (por, var, reg) | |
132 #define por_r2r(regs,regd) mmx_r2r (por, regs, regd) | |
133 | |
134 #define pslld_i2r(imm,reg) mmx_i2r (pslld, imm, reg) | |
135 #define pslld_m2r(var,reg) mmx_m2r (pslld, var, reg) | |
136 #define pslld_r2r(regs,regd) mmx_r2r (pslld, regs, regd) | |
137 #define psllq_i2r(imm,reg) mmx_i2r (psllq, imm, reg) | |
138 #define psllq_m2r(var,reg) mmx_m2r (psllq, var, reg) | |
139 #define psllq_r2r(regs,regd) mmx_r2r (psllq, regs, regd) | |
140 #define psllw_i2r(imm,reg) mmx_i2r (psllw, imm, reg) | |
141 #define psllw_m2r(var,reg) mmx_m2r (psllw, var, reg) | |
142 #define psllw_r2r(regs,regd) mmx_r2r (psllw, regs, regd) | |
143 | |
144 #define psrad_i2r(imm,reg) mmx_i2r (psrad, imm, reg) | |
145 #define psrad_m2r(var,reg) mmx_m2r (psrad, var, reg) | |
146 #define psrad_r2r(regs,regd) mmx_r2r (psrad, regs, regd) | |
147 #define psraw_i2r(imm,reg) mmx_i2r (psraw, imm, reg) | |
148 #define psraw_m2r(var,reg) mmx_m2r (psraw, var, reg) | |
149 #define psraw_r2r(regs,regd) mmx_r2r (psraw, regs, regd) | |
150 | |
151 #define psrld_i2r(imm,reg) mmx_i2r (psrld, imm, reg) | |
152 #define psrld_m2r(var,reg) mmx_m2r (psrld, var, reg) | |
153 #define psrld_r2r(regs,regd) mmx_r2r (psrld, regs, regd) | |
154 #define psrlq_i2r(imm,reg) mmx_i2r (psrlq, imm, reg) | |
155 #define psrlq_m2r(var,reg) mmx_m2r (psrlq, var, reg) | |
156 #define psrlq_r2r(regs,regd) mmx_r2r (psrlq, regs, regd) | |
157 #define psrlw_i2r(imm,reg) mmx_i2r (psrlw, imm, reg) | |
158 #define psrlw_m2r(var,reg) mmx_m2r (psrlw, var, reg) | |
159 #define psrlw_r2r(regs,regd) mmx_r2r (psrlw, regs, regd) | |
160 | |
161 #define psubb_m2r(var,reg) mmx_m2r (psubb, var, reg) | |
162 #define psubb_r2r(regs,regd) mmx_r2r (psubb, regs, regd) | |
163 #define psubd_m2r(var,reg) mmx_m2r (psubd, var, reg) | |
164 #define psubd_r2r(regs,regd) mmx_r2r (psubd, regs, regd) | |
165 #define psubw_m2r(var,reg) mmx_m2r (psubw, var, reg) | |
166 #define psubw_r2r(regs,regd) mmx_r2r (psubw, regs, regd) | |
167 | |
168 #define psubsb_m2r(var,reg) mmx_m2r (psubsb, var, reg) | |
169 #define psubsb_r2r(regs,regd) mmx_r2r (psubsb, regs, regd) | |
170 #define psubsw_m2r(var,reg) mmx_m2r (psubsw, var, reg) | |
171 #define psubsw_r2r(regs,regd) mmx_r2r (psubsw, regs, regd) | |
172 | |
173 #define psubusb_m2r(var,reg) mmx_m2r (psubusb, var, reg) | |
174 #define psubusb_r2r(regs,regd) mmx_r2r (psubusb, regs, regd) | |
175 #define psubusw_m2r(var,reg) mmx_m2r (psubusw, var, reg) | |
176 #define psubusw_r2r(regs,regd) mmx_r2r (psubusw, regs, regd) | |
177 | |
178 #define punpckhbw_m2r(var,reg) mmx_m2r (punpckhbw, var, reg) | |
179 #define punpckhbw_r2r(regs,regd) mmx_r2r (punpckhbw, regs, regd) | |
180 #define punpckhdq_m2r(var,reg) mmx_m2r (punpckhdq, var, reg) | |
181 #define punpckhdq_r2r(regs,regd) mmx_r2r (punpckhdq, regs, regd) | |
182 #define punpckhwd_m2r(var,reg) mmx_m2r (punpckhwd, var, reg) | |
183 #define punpckhwd_r2r(regs,regd) mmx_r2r (punpckhwd, regs, regd) | |
184 | |
185 #define punpcklbw_m2r(var,reg) mmx_m2r (punpcklbw, var, reg) | |
186 #define punpcklbw_r2r(regs,regd) mmx_r2r (punpcklbw, regs, regd) | |
187 #define punpckldq_m2r(var,reg) mmx_m2r (punpckldq, var, reg) | |
188 #define punpckldq_r2r(regs,regd) mmx_r2r (punpckldq, regs, regd) | |
189 #define punpcklwd_m2r(var,reg) mmx_m2r (punpcklwd, var, reg) | |
190 #define punpcklwd_r2r(regs,regd) mmx_r2r (punpcklwd, regs, regd) | |
191 | |
192 #define pxor_m2r(var,reg) mmx_m2r (pxor, var, reg) | |
193 #define pxor_r2r(regs,regd) mmx_r2r (pxor, regs, regd) | |
194 | |
195 | |
196 /* 3DNOW extensions */ | |
197 | |
198 #define pavgusb_m2r(var,reg) mmx_m2r (pavgusb, var, reg) | |
199 #define pavgusb_r2r(regs,regd) mmx_r2r (pavgusb, regs, regd) | |
200 | |
201 | |
202 /* AMD MMX extensions - also available in intel SSE */ | |
203 | |
204 | |
205 #define mmx_m2ri(op,mem,reg,imm) \ | |
9852 | 206 __asm__ __volatile__ (#op " %1, %0, %%" #reg \ |
207 : /* nothing */ \ | |
208 : "m" (mem), "i" (imm)) | |
209 | |
1 | 210 #define mmx_r2ri(op,regs,regd,imm) \ |
9852 | 211 __asm__ __volatile__ (#op " %0, %%" #regs ", %%" #regd \ |
212 : /* nothing */ \ | |
213 : "i" (imm) ) | |
1 | 214 |
215 #define mmx_fetch(mem,hint) \ | |
216 __asm__ __volatile__ ("prefetch" #hint " %0" \ | |
217 : /* nothing */ \ | |
9852 | 218 : "m" (mem)) |
1 | 219 |
220 | |
221 #define maskmovq(regs,maskreg) mmx_r2ri (maskmovq, regs, maskreg) | |
222 | |
223 #define movntq_r2m(mmreg,var) mmx_r2m (movntq, mmreg, var) | |
224 | |
225 #define pavgb_m2r(var,reg) mmx_m2r (pavgb, var, reg) | |
226 #define pavgb_r2r(regs,regd) mmx_r2r (pavgb, regs, regd) | |
227 #define pavgw_m2r(var,reg) mmx_m2r (pavgw, var, reg) | |
228 #define pavgw_r2r(regs,regd) mmx_r2r (pavgw, regs, regd) | |
229 | |
230 #define pextrw_r2r(mmreg,reg,imm) mmx_r2ri (pextrw, mmreg, reg, imm) | |
231 | |
232 #define pinsrw_r2r(reg,mmreg,imm) mmx_r2ri (pinsrw, reg, mmreg, imm) | |
233 | |
234 #define pmaxsw_m2r(var,reg) mmx_m2r (pmaxsw, var, reg) | |
235 #define pmaxsw_r2r(regs,regd) mmx_r2r (pmaxsw, regs, regd) | |
236 | |
237 #define pmaxub_m2r(var,reg) mmx_m2r (pmaxub, var, reg) | |
238 #define pmaxub_r2r(regs,regd) mmx_r2r (pmaxub, regs, regd) | |
239 | |
240 #define pminsw_m2r(var,reg) mmx_m2r (pminsw, var, reg) | |
241 #define pminsw_r2r(regs,regd) mmx_r2r (pminsw, regs, regd) | |
242 | |
243 #define pminub_m2r(var,reg) mmx_m2r (pminub, var, reg) | |
244 #define pminub_r2r(regs,regd) mmx_r2r (pminub, regs, regd) | |
245 | |
246 #define pmovmskb(mmreg,reg) \ | |
247 __asm__ __volatile__ ("movmskps %" #mmreg ", %" #reg) | |
248 | |
249 #define pmulhuw_m2r(var,reg) mmx_m2r (pmulhuw, var, reg) | |
250 #define pmulhuw_r2r(regs,regd) mmx_r2r (pmulhuw, regs, regd) | |
251 | |
252 #define prefetcht0(mem) mmx_fetch (mem, t0) | |
253 #define prefetcht1(mem) mmx_fetch (mem, t1) | |
254 #define prefetcht2(mem) mmx_fetch (mem, t2) | |
255 #define prefetchnta(mem) mmx_fetch (mem, nta) | |
256 | |
257 #define psadbw_m2r(var,reg) mmx_m2r (psadbw, var, reg) | |
258 #define psadbw_r2r(regs,regd) mmx_r2r (psadbw, regs, regd) | |
259 | |
26393
2506f1b0bdbe
Backport SSE2-optimized IDCT routines from upstream libmpeg2.
diego
parents:
12932
diff
changeset
|
260 |
2506f1b0bdbe
Backport SSE2-optimized IDCT routines from upstream libmpeg2.
diego
parents:
12932
diff
changeset
|
261 /* SSE2 */ |
2506f1b0bdbe
Backport SSE2-optimized IDCT routines from upstream libmpeg2.
diego
parents:
12932
diff
changeset
|
262 |
2506f1b0bdbe
Backport SSE2-optimized IDCT routines from upstream libmpeg2.
diego
parents:
12932
diff
changeset
|
263 typedef union { |
2506f1b0bdbe
Backport SSE2-optimized IDCT routines from upstream libmpeg2.
diego
parents:
12932
diff
changeset
|
264 long long q[2]; /* Quadword (64-bit) value */ |
2506f1b0bdbe
Backport SSE2-optimized IDCT routines from upstream libmpeg2.
diego
parents:
12932
diff
changeset
|
265 unsigned long long uq[2]; /* Unsigned Quadword */ |
2506f1b0bdbe
Backport SSE2-optimized IDCT routines from upstream libmpeg2.
diego
parents:
12932
diff
changeset
|
266 int d[4]; /* 2 Doubleword (32-bit) values */ |
2506f1b0bdbe
Backport SSE2-optimized IDCT routines from upstream libmpeg2.
diego
parents:
12932
diff
changeset
|
267 unsigned int ud[4]; /* 2 Unsigned Doubleword */ |
2506f1b0bdbe
Backport SSE2-optimized IDCT routines from upstream libmpeg2.
diego
parents:
12932
diff
changeset
|
268 short w[8]; /* 4 Word (16-bit) values */ |
2506f1b0bdbe
Backport SSE2-optimized IDCT routines from upstream libmpeg2.
diego
parents:
12932
diff
changeset
|
269 unsigned short uw[8]; /* 4 Unsigned Word */ |
2506f1b0bdbe
Backport SSE2-optimized IDCT routines from upstream libmpeg2.
diego
parents:
12932
diff
changeset
|
270 char b[16]; /* 8 Byte (8-bit) values */ |
2506f1b0bdbe
Backport SSE2-optimized IDCT routines from upstream libmpeg2.
diego
parents:
12932
diff
changeset
|
271 unsigned char ub[16]; /* 8 Unsigned Byte */ |
2506f1b0bdbe
Backport SSE2-optimized IDCT routines from upstream libmpeg2.
diego
parents:
12932
diff
changeset
|
272 float s[4]; /* Single-precision (32-bit) value */ |
2506f1b0bdbe
Backport SSE2-optimized IDCT routines from upstream libmpeg2.
diego
parents:
12932
diff
changeset
|
273 } ATTR_ALIGN(16) sse_t; /* On an 16-byte (128-bit) boundary */ |
2506f1b0bdbe
Backport SSE2-optimized IDCT routines from upstream libmpeg2.
diego
parents:
12932
diff
changeset
|
274 |
2506f1b0bdbe
Backport SSE2-optimized IDCT routines from upstream libmpeg2.
diego
parents:
12932
diff
changeset
|
275 #define movdqu_m2r(var,reg) mmx_m2r (movdqu, var, reg) |
2506f1b0bdbe
Backport SSE2-optimized IDCT routines from upstream libmpeg2.
diego
parents:
12932
diff
changeset
|
276 #define movdqu_r2m(reg,var) mmx_r2m (movdqu, reg, var) |
2506f1b0bdbe
Backport SSE2-optimized IDCT routines from upstream libmpeg2.
diego
parents:
12932
diff
changeset
|
277 #define movdqu_r2r(regs,regd) mmx_r2r (movdqu, regs, regd) |
2506f1b0bdbe
Backport SSE2-optimized IDCT routines from upstream libmpeg2.
diego
parents:
12932
diff
changeset
|
278 #define movdqa_m2r(var,reg) mmx_m2r (movdqa, var, reg) |
2506f1b0bdbe
Backport SSE2-optimized IDCT routines from upstream libmpeg2.
diego
parents:
12932
diff
changeset
|
279 #define movdqa_r2m(reg,var) mmx_r2m (movdqa, reg, var) |
2506f1b0bdbe
Backport SSE2-optimized IDCT routines from upstream libmpeg2.
diego
parents:
12932
diff
changeset
|
280 #define movdqa_r2r(regs,regd) mmx_r2r (movdqa, regs, regd) |
2506f1b0bdbe
Backport SSE2-optimized IDCT routines from upstream libmpeg2.
diego
parents:
12932
diff
changeset
|
281 |
2506f1b0bdbe
Backport SSE2-optimized IDCT routines from upstream libmpeg2.
diego
parents:
12932
diff
changeset
|
282 #define pshufd_r2r(regs,regd,imm) mmx_r2ri(pshufd, regs, regd, imm) |
2506f1b0bdbe
Backport SSE2-optimized IDCT routines from upstream libmpeg2.
diego
parents:
12932
diff
changeset
|
283 |
1 | 284 #define pshufw_m2r(var,reg,imm) mmx_m2ri(pshufw, var, reg, imm) |
285 #define pshufw_r2r(regs,regd,imm) mmx_r2ri(pshufw, regs, regd, imm) | |
286 | |
287 #define sfence() __asm__ __volatile__ ("sfence\n\t") |