FFmpeg
 All Data Structures Namespaces Files Functions Variables Typedefs Enumerations Enumerator Macros Groups Pages
simple_idct.c
Go to the documentation of this file.
1 /*
2  * Simple IDCT MMX
3  *
4  * Copyright (c) 2001, 2002 Michael Niedermayer <michaelni@gmx.at>
5  *
6  * This file is part of FFmpeg.
7  *
8  * FFmpeg is free software; you can redistribute it and/or
9  * modify it under the terms of the GNU Lesser General Public
10  * License as published by the Free Software Foundation; either
11  * version 2.1 of the License, or (at your option) any later version.
12  *
13  * FFmpeg is distributed in the hope that it will be useful,
14  * but WITHOUT ANY WARRANTY; without even the implied warranty of
15  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
16  * Lesser General Public License for more details.
17  *
18  * You should have received a copy of the GNU Lesser General Public
19  * License along with FFmpeg; if not, write to the Free Software
20  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
21  */
22 
23 #include "libavutil/mem.h"
24 #include "libavutil/x86/asm.h"
25 
26 #include "libavcodec/idctdsp.h"
27 
28 #include "idctdsp.h"
29 #include "simple_idct.h"
30 
31 #if HAVE_INLINE_ASM
32 
33 /*
34 23170.475006
35 22725.260826
36 21406.727617
37 19265.545870
38 16384.000000
39 12872.826198
40 8866.956905
41 4520.335430
42 */
43 #define C0 23170 //cos(i*M_PI/16)*sqrt(2)*(1<<14) + 0.5
44 #define C1 22725 //cos(i*M_PI/16)*sqrt(2)*(1<<14) + 0.5
45 #define C2 21407 //cos(i*M_PI/16)*sqrt(2)*(1<<14) + 0.5
46 #define C3 19266 //cos(i*M_PI/16)*sqrt(2)*(1<<14) + 0.5
47 #define C4 16383 //cos(i*M_PI/16)*sqrt(2)*(1<<14) - 0.5
48 #define C5 12873 //cos(i*M_PI/16)*sqrt(2)*(1<<14) + 0.5
49 #define C6 8867 //cos(i*M_PI/16)*sqrt(2)*(1<<14) + 0.5
50 #define C7 4520 //cos(i*M_PI/16)*sqrt(2)*(1<<14) + 0.5
51 
52 #define ROW_SHIFT 11
53 #define COL_SHIFT 20 // 6
54 
55 DECLARE_ASM_CONST(8, uint64_t, wm1010)= 0xFFFF0000FFFF0000ULL;
56 DECLARE_ASM_CONST(8, uint64_t, d40000)= 0x0000000000040000ULL;
57 
58 DECLARE_ALIGNED(8, static const int16_t, coeffs)[]= {
59  1<<(ROW_SHIFT-1), 0, 1<<(ROW_SHIFT-1), 0,
60 // 1<<(COL_SHIFT-1), 0, 1<<(COL_SHIFT-1), 0,
61 // 0, 1<<(COL_SHIFT-1-16), 0, 1<<(COL_SHIFT-1-16),
62  1<<(ROW_SHIFT-1), 1, 1<<(ROW_SHIFT-1), 0,
63  // the 1 = ((1<<(COL_SHIFT-1))/C4)<<ROW_SHIFT :)
64 // 0, 0, 0, 0,
65 // 0, 0, 0, 0,
66 
67  C4, C4, C4, C4,
68  C4, -C4, C4, -C4,
69 
70  C2, C6, C2, C6,
71  C6, -C2, C6, -C2,
72 
73  C1, C3, C1, C3,
74  C5, C7, C5, C7,
75 
76  C3, -C7, C3, -C7,
77 -C1, -C5, -C1, -C5,
78 
79  C5, -C1, C5, -C1,
80  C7, C3, C7, C3,
81 
82  C7, -C5, C7, -C5,
83  C3, -C1, C3, -C1
84 };
85 
86 static inline void idct(int16_t *block)
87 {
88  LOCAL_ALIGNED_8(int64_t, align_tmp, [16]);
89  int16_t * const temp= (int16_t*)align_tmp;
90 
91  __asm__ volatile(
92 #define DC_COND_IDCT(src0, src4, src1, src5, dst, rounder, shift) \
93  "movq " #src0 ", %%mm0 \n\t" /* R4 R0 r4 r0 */\
94  "movq " #src4 ", %%mm1 \n\t" /* R6 R2 r6 r2 */\
95  "movq " #src1 ", %%mm2 \n\t" /* R3 R1 r3 r1 */\
96  "movq " #src5 ", %%mm3 \n\t" /* R7 R5 r7 r5 */\
97  "movq "MANGLE(wm1010)", %%mm4 \n\t"\
98  "pand %%mm0, %%mm4 \n\t"\
99  "por %%mm1, %%mm4 \n\t"\
100  "por %%mm2, %%mm4 \n\t"\
101  "por %%mm3, %%mm4 \n\t"\
102  "packssdw %%mm4,%%mm4 \n\t"\
103  "movd %%mm4, %%eax \n\t"\
104  "orl %%eax, %%eax \n\t"\
105  "jz 1f \n\t"\
106  "movq 16(%2), %%mm4 \n\t" /* C4 C4 C4 C4 */\
107  "pmaddwd %%mm0, %%mm4 \n\t" /* C4R4+C4R0 C4r4+C4r0 */\
108  "movq 24(%2), %%mm5 \n\t" /* -C4 C4 -C4 C4 */\
109  "pmaddwd %%mm5, %%mm0 \n\t" /* -C4R4+C4R0 -C4r4+C4r0 */\
110  "movq 32(%2), %%mm5 \n\t" /* C6 C2 C6 C2 */\
111  "pmaddwd %%mm1, %%mm5 \n\t" /* C6R6+C2R2 C6r6+C2r2 */\
112  "movq 40(%2), %%mm6 \n\t" /* -C2 C6 -C2 C6 */\
113  "pmaddwd %%mm6, %%mm1 \n\t" /* -C2R6+C6R2 -C2r6+C6r2 */\
114  "movq 48(%2), %%mm7 \n\t" /* C3 C1 C3 C1 */\
115  "pmaddwd %%mm2, %%mm7 \n\t" /* C3R3+C1R1 C3r3+C1r1 */\
116  #rounder ", %%mm4 \n\t"\
117  "movq %%mm4, %%mm6 \n\t" /* C4R4+C4R0 C4r4+C4r0 */\
118  "paddd %%mm5, %%mm4 \n\t" /* A0 a0 */\
119  "psubd %%mm5, %%mm6 \n\t" /* A3 a3 */\
120  "movq 56(%2), %%mm5 \n\t" /* C7 C5 C7 C5 */\
121  "pmaddwd %%mm3, %%mm5 \n\t" /* C7R7+C5R5 C7r7+C5r5 */\
122  #rounder ", %%mm0 \n\t"\
123  "paddd %%mm0, %%mm1 \n\t" /* A1 a1 */\
124  "paddd %%mm0, %%mm0 \n\t" \
125  "psubd %%mm1, %%mm0 \n\t" /* A2 a2 */\
126  "pmaddwd 64(%2), %%mm2 \n\t" /* -C7R3+C3R1 -C7r3+C3r1 */\
127  "paddd %%mm5, %%mm7 \n\t" /* B0 b0 */\
128  "movq 72(%2), %%mm5 \n\t" /* -C5 -C1 -C5 -C1 */\
129  "pmaddwd %%mm3, %%mm5 \n\t" /* -C5R7-C1R5 -C5r7-C1r5 */\
130  "paddd %%mm4, %%mm7 \n\t" /* A0+B0 a0+b0 */\
131  "paddd %%mm4, %%mm4 \n\t" /* 2A0 2a0 */\
132  "psubd %%mm7, %%mm4 \n\t" /* A0-B0 a0-b0 */\
133  "paddd %%mm2, %%mm5 \n\t" /* B1 b1 */\
134  "psrad $" #shift ", %%mm7 \n\t"\
135  "psrad $" #shift ", %%mm4 \n\t"\
136  "movq %%mm1, %%mm2 \n\t" /* A1 a1 */\
137  "paddd %%mm5, %%mm1 \n\t" /* A1+B1 a1+b1 */\
138  "psubd %%mm5, %%mm2 \n\t" /* A1-B1 a1-b1 */\
139  "psrad $" #shift ", %%mm1 \n\t"\
140  "psrad $" #shift ", %%mm2 \n\t"\
141  "packssdw %%mm1, %%mm7 \n\t" /* A1+B1 a1+b1 A0+B0 a0+b0 */\
142  "packssdw %%mm4, %%mm2 \n\t" /* A0-B0 a0-b0 A1-B1 a1-b1 */\
143  "movq %%mm7, " #dst " \n\t"\
144  "movq " #src1 ", %%mm1 \n\t" /* R3 R1 r3 r1 */\
145  "movq 80(%2), %%mm4 \n\t" /* -C1 C5 -C1 C5 */\
146  "movq %%mm2, 24+" #dst " \n\t"\
147  "pmaddwd %%mm1, %%mm4 \n\t" /* -C1R3+C5R1 -C1r3+C5r1 */\
148  "movq 88(%2), %%mm7 \n\t" /* C3 C7 C3 C7 */\
149  "pmaddwd 96(%2), %%mm1 \n\t" /* -C5R3+C7R1 -C5r3+C7r1 */\
150  "pmaddwd %%mm3, %%mm7 \n\t" /* C3R7+C7R5 C3r7+C7r5 */\
151  "movq %%mm0, %%mm2 \n\t" /* A2 a2 */\
152  "pmaddwd 104(%2), %%mm3 \n\t" /* -C1R7+C3R5 -C1r7+C3r5 */\
153  "paddd %%mm7, %%mm4 \n\t" /* B2 b2 */\
154  "paddd %%mm4, %%mm2 \n\t" /* A2+B2 a2+b2 */\
155  "psubd %%mm4, %%mm0 \n\t" /* a2-B2 a2-b2 */\
156  "psrad $" #shift ", %%mm2 \n\t"\
157  "psrad $" #shift ", %%mm0 \n\t"\
158  "movq %%mm6, %%mm4 \n\t" /* A3 a3 */\
159  "paddd %%mm1, %%mm3 \n\t" /* B3 b3 */\
160  "paddd %%mm3, %%mm6 \n\t" /* A3+B3 a3+b3 */\
161  "psubd %%mm3, %%mm4 \n\t" /* a3-B3 a3-b3 */\
162  "psrad $" #shift ", %%mm6 \n\t"\
163  "packssdw %%mm6, %%mm2 \n\t" /* A3+B3 a3+b3 A2+B2 a2+b2 */\
164  "movq %%mm2, 8+" #dst " \n\t"\
165  "psrad $" #shift ", %%mm4 \n\t"\
166  "packssdw %%mm0, %%mm4 \n\t" /* A2-B2 a2-b2 A3-B3 a3-b3 */\
167  "movq %%mm4, 16+" #dst " \n\t"\
168  "jmp 2f \n\t"\
169  "1: \n\t"\
170  "pslld $16, %%mm0 \n\t"\
171  "paddd "MANGLE(d40000)", %%mm0 \n\t"\
172  "psrad $13, %%mm0 \n\t"\
173  "packssdw %%mm0, %%mm0 \n\t"\
174  "movq %%mm0, " #dst " \n\t"\
175  "movq %%mm0, 8+" #dst " \n\t"\
176  "movq %%mm0, 16+" #dst " \n\t"\
177  "movq %%mm0, 24+" #dst " \n\t"\
178  "2: \n\t"
179 
180 #define Z_COND_IDCT(src0, src4, src1, src5, dst, rounder, shift, bt) \
181  "movq " #src0 ", %%mm0 \n\t" /* R4 R0 r4 r0 */\
182  "movq " #src4 ", %%mm1 \n\t" /* R6 R2 r6 r2 */\
183  "movq " #src1 ", %%mm2 \n\t" /* R3 R1 r3 r1 */\
184  "movq " #src5 ", %%mm3 \n\t" /* R7 R5 r7 r5 */\
185  "movq %%mm0, %%mm4 \n\t"\
186  "por %%mm1, %%mm4 \n\t"\
187  "por %%mm2, %%mm4 \n\t"\
188  "por %%mm3, %%mm4 \n\t"\
189  "packssdw %%mm4,%%mm4 \n\t"\
190  "movd %%mm4, %%eax \n\t"\
191  "orl %%eax, %%eax \n\t"\
192  "jz " #bt " \n\t"\
193  "movq 16(%2), %%mm4 \n\t" /* C4 C4 C4 C4 */\
194  "pmaddwd %%mm0, %%mm4 \n\t" /* C4R4+C4R0 C4r4+C4r0 */\
195  "movq 24(%2), %%mm5 \n\t" /* -C4 C4 -C4 C4 */\
196  "pmaddwd %%mm5, %%mm0 \n\t" /* -C4R4+C4R0 -C4r4+C4r0 */\
197  "movq 32(%2), %%mm5 \n\t" /* C6 C2 C6 C2 */\
198  "pmaddwd %%mm1, %%mm5 \n\t" /* C6R6+C2R2 C6r6+C2r2 */\
199  "movq 40(%2), %%mm6 \n\t" /* -C2 C6 -C2 C6 */\
200  "pmaddwd %%mm6, %%mm1 \n\t" /* -C2R6+C6R2 -C2r6+C6r2 */\
201  "movq 48(%2), %%mm7 \n\t" /* C3 C1 C3 C1 */\
202  "pmaddwd %%mm2, %%mm7 \n\t" /* C3R3+C1R1 C3r3+C1r1 */\
203  #rounder ", %%mm4 \n\t"\
204  "movq %%mm4, %%mm6 \n\t" /* C4R4+C4R0 C4r4+C4r0 */\
205  "paddd %%mm5, %%mm4 \n\t" /* A0 a0 */\
206  "psubd %%mm5, %%mm6 \n\t" /* A3 a3 */\
207  "movq 56(%2), %%mm5 \n\t" /* C7 C5 C7 C5 */\
208  "pmaddwd %%mm3, %%mm5 \n\t" /* C7R7+C5R5 C7r7+C5r5 */\
209  #rounder ", %%mm0 \n\t"\
210  "paddd %%mm0, %%mm1 \n\t" /* A1 a1 */\
211  "paddd %%mm0, %%mm0 \n\t" \
212  "psubd %%mm1, %%mm0 \n\t" /* A2 a2 */\
213  "pmaddwd 64(%2), %%mm2 \n\t" /* -C7R3+C3R1 -C7r3+C3r1 */\
214  "paddd %%mm5, %%mm7 \n\t" /* B0 b0 */\
215  "movq 72(%2), %%mm5 \n\t" /* -C5 -C1 -C5 -C1 */\
216  "pmaddwd %%mm3, %%mm5 \n\t" /* -C5R7-C1R5 -C5r7-C1r5 */\
217  "paddd %%mm4, %%mm7 \n\t" /* A0+B0 a0+b0 */\
218  "paddd %%mm4, %%mm4 \n\t" /* 2A0 2a0 */\
219  "psubd %%mm7, %%mm4 \n\t" /* A0-B0 a0-b0 */\
220  "paddd %%mm2, %%mm5 \n\t" /* B1 b1 */\
221  "psrad $" #shift ", %%mm7 \n\t"\
222  "psrad $" #shift ", %%mm4 \n\t"\
223  "movq %%mm1, %%mm2 \n\t" /* A1 a1 */\
224  "paddd %%mm5, %%mm1 \n\t" /* A1+B1 a1+b1 */\
225  "psubd %%mm5, %%mm2 \n\t" /* A1-B1 a1-b1 */\
226  "psrad $" #shift ", %%mm1 \n\t"\
227  "psrad $" #shift ", %%mm2 \n\t"\
228  "packssdw %%mm1, %%mm7 \n\t" /* A1+B1 a1+b1 A0+B0 a0+b0 */\
229  "packssdw %%mm4, %%mm2 \n\t" /* A0-B0 a0-b0 A1-B1 a1-b1 */\
230  "movq %%mm7, " #dst " \n\t"\
231  "movq " #src1 ", %%mm1 \n\t" /* R3 R1 r3 r1 */\
232  "movq 80(%2), %%mm4 \n\t" /* -C1 C5 -C1 C5 */\
233  "movq %%mm2, 24+" #dst " \n\t"\
234  "pmaddwd %%mm1, %%mm4 \n\t" /* -C1R3+C5R1 -C1r3+C5r1 */\
235  "movq 88(%2), %%mm7 \n\t" /* C3 C7 C3 C7 */\
236  "pmaddwd 96(%2), %%mm1 \n\t" /* -C5R3+C7R1 -C5r3+C7r1 */\
237  "pmaddwd %%mm3, %%mm7 \n\t" /* C3R7+C7R5 C3r7+C7r5 */\
238  "movq %%mm0, %%mm2 \n\t" /* A2 a2 */\
239  "pmaddwd 104(%2), %%mm3 \n\t" /* -C1R7+C3R5 -C1r7+C3r5 */\
240  "paddd %%mm7, %%mm4 \n\t" /* B2 b2 */\
241  "paddd %%mm4, %%mm2 \n\t" /* A2+B2 a2+b2 */\
242  "psubd %%mm4, %%mm0 \n\t" /* a2-B2 a2-b2 */\
243  "psrad $" #shift ", %%mm2 \n\t"\
244  "psrad $" #shift ", %%mm0 \n\t"\
245  "movq %%mm6, %%mm4 \n\t" /* A3 a3 */\
246  "paddd %%mm1, %%mm3 \n\t" /* B3 b3 */\
247  "paddd %%mm3, %%mm6 \n\t" /* A3+B3 a3+b3 */\
248  "psubd %%mm3, %%mm4 \n\t" /* a3-B3 a3-b3 */\
249  "psrad $" #shift ", %%mm6 \n\t"\
250  "packssdw %%mm6, %%mm2 \n\t" /* A3+B3 a3+b3 A2+B2 a2+b2 */\
251  "movq %%mm2, 8+" #dst " \n\t"\
252  "psrad $" #shift ", %%mm4 \n\t"\
253  "packssdw %%mm0, %%mm4 \n\t" /* A2-B2 a2-b2 A3-B3 a3-b3 */\
254  "movq %%mm4, 16+" #dst " \n\t"\
255 
256 #define ROW_IDCT(src0, src4, src1, src5, dst, rounder, shift) \
257  "movq " #src0 ", %%mm0 \n\t" /* R4 R0 r4 r0 */\
258  "movq " #src4 ", %%mm1 \n\t" /* R6 R2 r6 r2 */\
259  "movq " #src1 ", %%mm2 \n\t" /* R3 R1 r3 r1 */\
260  "movq " #src5 ", %%mm3 \n\t" /* R7 R5 r7 r5 */\
261  "movq 16(%2), %%mm4 \n\t" /* C4 C4 C4 C4 */\
262  "pmaddwd %%mm0, %%mm4 \n\t" /* C4R4+C4R0 C4r4+C4r0 */\
263  "movq 24(%2), %%mm5 \n\t" /* -C4 C4 -C4 C4 */\
264  "pmaddwd %%mm5, %%mm0 \n\t" /* -C4R4+C4R0 -C4r4+C4r0 */\
265  "movq 32(%2), %%mm5 \n\t" /* C6 C2 C6 C2 */\
266  "pmaddwd %%mm1, %%mm5 \n\t" /* C6R6+C2R2 C6r6+C2r2 */\
267  "movq 40(%2), %%mm6 \n\t" /* -C2 C6 -C2 C6 */\
268  "pmaddwd %%mm6, %%mm1 \n\t" /* -C2R6+C6R2 -C2r6+C6r2 */\
269  "movq 48(%2), %%mm7 \n\t" /* C3 C1 C3 C1 */\
270  "pmaddwd %%mm2, %%mm7 \n\t" /* C3R3+C1R1 C3r3+C1r1 */\
271  #rounder ", %%mm4 \n\t"\
272  "movq %%mm4, %%mm6 \n\t" /* C4R4+C4R0 C4r4+C4r0 */\
273  "paddd %%mm5, %%mm4 \n\t" /* A0 a0 */\
274  "psubd %%mm5, %%mm6 \n\t" /* A3 a3 */\
275  "movq 56(%2), %%mm5 \n\t" /* C7 C5 C7 C5 */\
276  "pmaddwd %%mm3, %%mm5 \n\t" /* C7R7+C5R5 C7r7+C5r5 */\
277  #rounder ", %%mm0 \n\t"\
278  "paddd %%mm0, %%mm1 \n\t" /* A1 a1 */\
279  "paddd %%mm0, %%mm0 \n\t" \
280  "psubd %%mm1, %%mm0 \n\t" /* A2 a2 */\
281  "pmaddwd 64(%2), %%mm2 \n\t" /* -C7R3+C3R1 -C7r3+C3r1 */\
282  "paddd %%mm5, %%mm7 \n\t" /* B0 b0 */\
283  "movq 72(%2), %%mm5 \n\t" /* -C5 -C1 -C5 -C1 */\
284  "pmaddwd %%mm3, %%mm5 \n\t" /* -C5R7-C1R5 -C5r7-C1r5 */\
285  "paddd %%mm4, %%mm7 \n\t" /* A0+B0 a0+b0 */\
286  "paddd %%mm4, %%mm4 \n\t" /* 2A0 2a0 */\
287  "psubd %%mm7, %%mm4 \n\t" /* A0-B0 a0-b0 */\
288  "paddd %%mm2, %%mm5 \n\t" /* B1 b1 */\
289  "psrad $" #shift ", %%mm7 \n\t"\
290  "psrad $" #shift ", %%mm4 \n\t"\
291  "movq %%mm1, %%mm2 \n\t" /* A1 a1 */\
292  "paddd %%mm5, %%mm1 \n\t" /* A1+B1 a1+b1 */\
293  "psubd %%mm5, %%mm2 \n\t" /* A1-B1 a1-b1 */\
294  "psrad $" #shift ", %%mm1 \n\t"\
295  "psrad $" #shift ", %%mm2 \n\t"\
296  "packssdw %%mm1, %%mm7 \n\t" /* A1+B1 a1+b1 A0+B0 a0+b0 */\
297  "packssdw %%mm4, %%mm2 \n\t" /* A0-B0 a0-b0 A1-B1 a1-b1 */\
298  "movq %%mm7, " #dst " \n\t"\
299  "movq " #src1 ", %%mm1 \n\t" /* R3 R1 r3 r1 */\
300  "movq 80(%2), %%mm4 \n\t" /* -C1 C5 -C1 C5 */\
301  "movq %%mm2, 24+" #dst " \n\t"\
302  "pmaddwd %%mm1, %%mm4 \n\t" /* -C1R3+C5R1 -C1r3+C5r1 */\
303  "movq 88(%2), %%mm7 \n\t" /* C3 C7 C3 C7 */\
304  "pmaddwd 96(%2), %%mm1 \n\t" /* -C5R3+C7R1 -C5r3+C7r1 */\
305  "pmaddwd %%mm3, %%mm7 \n\t" /* C3R7+C7R5 C3r7+C7r5 */\
306  "movq %%mm0, %%mm2 \n\t" /* A2 a2 */\
307  "pmaddwd 104(%2), %%mm3 \n\t" /* -C1R7+C3R5 -C1r7+C3r5 */\
308  "paddd %%mm7, %%mm4 \n\t" /* B2 b2 */\
309  "paddd %%mm4, %%mm2 \n\t" /* A2+B2 a2+b2 */\
310  "psubd %%mm4, %%mm0 \n\t" /* a2-B2 a2-b2 */\
311  "psrad $" #shift ", %%mm2 \n\t"\
312  "psrad $" #shift ", %%mm0 \n\t"\
313  "movq %%mm6, %%mm4 \n\t" /* A3 a3 */\
314  "paddd %%mm1, %%mm3 \n\t" /* B3 b3 */\
315  "paddd %%mm3, %%mm6 \n\t" /* A3+B3 a3+b3 */\
316  "psubd %%mm3, %%mm4 \n\t" /* a3-B3 a3-b3 */\
317  "psrad $" #shift ", %%mm6 \n\t"\
318  "packssdw %%mm6, %%mm2 \n\t" /* A3+B3 a3+b3 A2+B2 a2+b2 */\
319  "movq %%mm2, 8+" #dst " \n\t"\
320  "psrad $" #shift ", %%mm4 \n\t"\
321  "packssdw %%mm0, %%mm4 \n\t" /* A2-B2 a2-b2 A3-B3 a3-b3 */\
322  "movq %%mm4, 16+" #dst " \n\t"\
323 
324 //IDCT( src0, src4, src1, src5, dst, rounder, shift)
325 DC_COND_IDCT( 0(%0), 8(%0), 16(%0), 24(%0), 0(%1),paddd 8(%2), 11)
326 Z_COND_IDCT( 32(%0), 40(%0), 48(%0), 56(%0), 32(%1),paddd (%2), 11, 4f)
327 Z_COND_IDCT( 64(%0), 72(%0), 80(%0), 88(%0), 64(%1),paddd (%2), 11, 2f)
328 Z_COND_IDCT( 96(%0),104(%0),112(%0),120(%0), 96(%1),paddd (%2), 11, 1f)
329 
330 #undef IDCT
331 #define IDCT(src0, src4, src1, src5, dst, shift) \
332  "movq " #src0 ", %%mm0 \n\t" /* R4 R0 r4 r0 */\
333  "movq " #src4 ", %%mm1 \n\t" /* R6 R2 r6 r2 */\
334  "movq " #src1 ", %%mm2 \n\t" /* R3 R1 r3 r1 */\
335  "movq " #src5 ", %%mm3 \n\t" /* R7 R5 r7 r5 */\
336  "movq 16(%2), %%mm4 \n\t" /* C4 C4 C4 C4 */\
337  "pmaddwd %%mm0, %%mm4 \n\t" /* C4R4+C4R0 C4r4+C4r0 */\
338  "movq 24(%2), %%mm5 \n\t" /* -C4 C4 -C4 C4 */\
339  "pmaddwd %%mm5, %%mm0 \n\t" /* -C4R4+C4R0 -C4r4+C4r0 */\
340  "movq 32(%2), %%mm5 \n\t" /* C6 C2 C6 C2 */\
341  "pmaddwd %%mm1, %%mm5 \n\t" /* C6R6+C2R2 C6r6+C2r2 */\
342  "movq 40(%2), %%mm6 \n\t" /* -C2 C6 -C2 C6 */\
343  "pmaddwd %%mm6, %%mm1 \n\t" /* -C2R6+C6R2 -C2r6+C6r2 */\
344  "movq %%mm4, %%mm6 \n\t" /* C4R4+C4R0 C4r4+C4r0 */\
345  "movq 48(%2), %%mm7 \n\t" /* C3 C1 C3 C1 */\
346  "pmaddwd %%mm2, %%mm7 \n\t" /* C3R3+C1R1 C3r3+C1r1 */\
347  "paddd %%mm5, %%mm4 \n\t" /* A0 a0 */\
348  "psubd %%mm5, %%mm6 \n\t" /* A3 a3 */\
349  "movq %%mm0, %%mm5 \n\t" /* -C4R4+C4R0 -C4r4+C4r0 */\
350  "paddd %%mm1, %%mm0 \n\t" /* A1 a1 */\
351  "psubd %%mm1, %%mm5 \n\t" /* A2 a2 */\
352  "movq 56(%2), %%mm1 \n\t" /* C7 C5 C7 C5 */\
353  "pmaddwd %%mm3, %%mm1 \n\t" /* C7R7+C5R5 C7r7+C5r5 */\
354  "pmaddwd 64(%2), %%mm2 \n\t" /* -C7R3+C3R1 -C7r3+C3r1 */\
355  "paddd %%mm1, %%mm7 \n\t" /* B0 b0 */\
356  "movq 72(%2), %%mm1 \n\t" /* -C5 -C1 -C5 -C1 */\
357  "pmaddwd %%mm3, %%mm1 \n\t" /* -C5R7-C1R5 -C5r7-C1r5 */\
358  "paddd %%mm4, %%mm7 \n\t" /* A0+B0 a0+b0 */\
359  "paddd %%mm4, %%mm4 \n\t" /* 2A0 2a0 */\
360  "psubd %%mm7, %%mm4 \n\t" /* A0-B0 a0-b0 */\
361  "paddd %%mm2, %%mm1 \n\t" /* B1 b1 */\
362  "psrad $" #shift ", %%mm7 \n\t"\
363  "psrad $" #shift ", %%mm4 \n\t"\
364  "movq %%mm0, %%mm2 \n\t" /* A1 a1 */\
365  "paddd %%mm1, %%mm0 \n\t" /* A1+B1 a1+b1 */\
366  "psubd %%mm1, %%mm2 \n\t" /* A1-B1 a1-b1 */\
367  "psrad $" #shift ", %%mm0 \n\t"\
368  "psrad $" #shift ", %%mm2 \n\t"\
369  "packssdw %%mm7, %%mm7 \n\t" /* A0+B0 a0+b0 */\
370  "movd %%mm7, " #dst " \n\t"\
371  "packssdw %%mm0, %%mm0 \n\t" /* A1+B1 a1+b1 */\
372  "movd %%mm0, 16+" #dst " \n\t"\
373  "packssdw %%mm2, %%mm2 \n\t" /* A1-B1 a1-b1 */\
374  "movd %%mm2, 96+" #dst " \n\t"\
375  "packssdw %%mm4, %%mm4 \n\t" /* A0-B0 a0-b0 */\
376  "movd %%mm4, 112+" #dst " \n\t"\
377  "movq " #src1 ", %%mm0 \n\t" /* R3 R1 r3 r1 */\
378  "movq 80(%2), %%mm4 \n\t" /* -C1 C5 -C1 C5 */\
379  "pmaddwd %%mm0, %%mm4 \n\t" /* -C1R3+C5R1 -C1r3+C5r1 */\
380  "movq 88(%2), %%mm7 \n\t" /* C3 C7 C3 C7 */\
381  "pmaddwd 96(%2), %%mm0 \n\t" /* -C5R3+C7R1 -C5r3+C7r1 */\
382  "pmaddwd %%mm3, %%mm7 \n\t" /* C3R7+C7R5 C3r7+C7r5 */\
383  "movq %%mm5, %%mm2 \n\t" /* A2 a2 */\
384  "pmaddwd 104(%2), %%mm3 \n\t" /* -C1R7+C3R5 -C1r7+C3r5 */\
385  "paddd %%mm7, %%mm4 \n\t" /* B2 b2 */\
386  "paddd %%mm4, %%mm2 \n\t" /* A2+B2 a2+b2 */\
387  "psubd %%mm4, %%mm5 \n\t" /* a2-B2 a2-b2 */\
388  "psrad $" #shift ", %%mm2 \n\t"\
389  "psrad $" #shift ", %%mm5 \n\t"\
390  "movq %%mm6, %%mm4 \n\t" /* A3 a3 */\
391  "paddd %%mm0, %%mm3 \n\t" /* B3 b3 */\
392  "paddd %%mm3, %%mm6 \n\t" /* A3+B3 a3+b3 */\
393  "psubd %%mm3, %%mm4 \n\t" /* a3-B3 a3-b3 */\
394  "psrad $" #shift ", %%mm6 \n\t"\
395  "psrad $" #shift ", %%mm4 \n\t"\
396  "packssdw %%mm2, %%mm2 \n\t" /* A2+B2 a2+b2 */\
397  "packssdw %%mm6, %%mm6 \n\t" /* A3+B3 a3+b3 */\
398  "movd %%mm2, 32+" #dst " \n\t"\
399  "packssdw %%mm4, %%mm4 \n\t" /* A3-B3 a3-b3 */\
400  "packssdw %%mm5, %%mm5 \n\t" /* A2-B2 a2-b2 */\
401  "movd %%mm6, 48+" #dst " \n\t"\
402  "movd %%mm4, 64+" #dst " \n\t"\
403  "movd %%mm5, 80+" #dst " \n\t"
404 
405 
406 //IDCT( src0, src4, src1, src5, dst, shift)
407 IDCT( (%1), 64(%1), 32(%1), 96(%1), 0(%0), 20)
408 IDCT( 8(%1), 72(%1), 40(%1), 104(%1), 4(%0), 20)
409 IDCT( 16(%1), 80(%1), 48(%1), 112(%1), 8(%0), 20)
410 IDCT( 24(%1), 88(%1), 56(%1), 120(%1), 12(%0), 20)
411  "jmp 9f \n\t"
412 
413  "# .p2align 4 \n\t"\
414  "4: \n\t"
415 Z_COND_IDCT( 64(%0), 72(%0), 80(%0), 88(%0), 64(%1),paddd (%2), 11, 6f)
416 Z_COND_IDCT( 96(%0),104(%0),112(%0),120(%0), 96(%1),paddd (%2), 11, 5f)
417 
418 #undef IDCT
419 #define IDCT(src0, src4, src1, src5, dst, shift) \
420  "movq " #src0 ", %%mm0 \n\t" /* R4 R0 r4 r0 */\
421  "movq " #src4 ", %%mm1 \n\t" /* R6 R2 r6 r2 */\
422  "movq " #src5 ", %%mm3 \n\t" /* R7 R5 r7 r5 */\
423  "movq 16(%2), %%mm4 \n\t" /* C4 C4 C4 C4 */\
424  "pmaddwd %%mm0, %%mm4 \n\t" /* C4R4+C4R0 C4r4+C4r0 */\
425  "movq 24(%2), %%mm5 \n\t" /* -C4 C4 -C4 C4 */\
426  "pmaddwd %%mm5, %%mm0 \n\t" /* -C4R4+C4R0 -C4r4+C4r0 */\
427  "movq 32(%2), %%mm5 \n\t" /* C6 C2 C6 C2 */\
428  "pmaddwd %%mm1, %%mm5 \n\t" /* C6R6+C2R2 C6r6+C2r2 */\
429  "movq 40(%2), %%mm6 \n\t" /* -C2 C6 -C2 C6 */\
430  "pmaddwd %%mm6, %%mm1 \n\t" /* -C2R6+C6R2 -C2r6+C6r2 */\
431  "movq %%mm4, %%mm6 \n\t" /* C4R4+C4R0 C4r4+C4r0 */\
432  "paddd %%mm5, %%mm4 \n\t" /* A0 a0 */\
433  "psubd %%mm5, %%mm6 \n\t" /* A3 a3 */\
434  "movq %%mm0, %%mm5 \n\t" /* -C4R4+C4R0 -C4r4+C4r0 */\
435  "paddd %%mm1, %%mm0 \n\t" /* A1 a1 */\
436  "psubd %%mm1, %%mm5 \n\t" /* A2 a2 */\
437  "movq 56(%2), %%mm1 \n\t" /* C7 C5 C7 C5 */\
438  "pmaddwd %%mm3, %%mm1 \n\t" /* C7R7+C5R5 C7r7+C5r5 */\
439  "movq 72(%2), %%mm7 \n\t" /* -C5 -C1 -C5 -C1 */\
440  "pmaddwd %%mm3, %%mm7 \n\t" /* -C5R7-C1R5 -C5r7-C1r5 */\
441  "paddd %%mm4, %%mm1 \n\t" /* A0+B0 a0+b0 */\
442  "paddd %%mm4, %%mm4 \n\t" /* 2A0 2a0 */\
443  "psubd %%mm1, %%mm4 \n\t" /* A0-B0 a0-b0 */\
444  "psrad $" #shift ", %%mm1 \n\t"\
445  "psrad $" #shift ", %%mm4 \n\t"\
446  "movq %%mm0, %%mm2 \n\t" /* A1 a1 */\
447  "paddd %%mm7, %%mm0 \n\t" /* A1+B1 a1+b1 */\
448  "psubd %%mm7, %%mm2 \n\t" /* A1-B1 a1-b1 */\
449  "psrad $" #shift ", %%mm0 \n\t"\
450  "psrad $" #shift ", %%mm2 \n\t"\
451  "packssdw %%mm1, %%mm1 \n\t" /* A0+B0 a0+b0 */\
452  "movd %%mm1, " #dst " \n\t"\
453  "packssdw %%mm0, %%mm0 \n\t" /* A1+B1 a1+b1 */\
454  "movd %%mm0, 16+" #dst " \n\t"\
455  "packssdw %%mm2, %%mm2 \n\t" /* A1-B1 a1-b1 */\
456  "movd %%mm2, 96+" #dst " \n\t"\
457  "packssdw %%mm4, %%mm4 \n\t" /* A0-B0 a0-b0 */\
458  "movd %%mm4, 112+" #dst " \n\t"\
459  "movq 88(%2), %%mm1 \n\t" /* C3 C7 C3 C7 */\
460  "pmaddwd %%mm3, %%mm1 \n\t" /* C3R7+C7R5 C3r7+C7r5 */\
461  "movq %%mm5, %%mm2 \n\t" /* A2 a2 */\
462  "pmaddwd 104(%2), %%mm3 \n\t" /* -C1R7+C3R5 -C1r7+C3r5 */\
463  "paddd %%mm1, %%mm2 \n\t" /* A2+B2 a2+b2 */\
464  "psubd %%mm1, %%mm5 \n\t" /* a2-B2 a2-b2 */\
465  "psrad $" #shift ", %%mm2 \n\t"\
466  "psrad $" #shift ", %%mm5 \n\t"\
467  "movq %%mm6, %%mm1 \n\t" /* A3 a3 */\
468  "paddd %%mm3, %%mm6 \n\t" /* A3+B3 a3+b3 */\
469  "psubd %%mm3, %%mm1 \n\t" /* a3-B3 a3-b3 */\
470  "psrad $" #shift ", %%mm6 \n\t"\
471  "psrad $" #shift ", %%mm1 \n\t"\
472  "packssdw %%mm2, %%mm2 \n\t" /* A2+B2 a2+b2 */\
473  "packssdw %%mm6, %%mm6 \n\t" /* A3+B3 a3+b3 */\
474  "movd %%mm2, 32+" #dst " \n\t"\
475  "packssdw %%mm1, %%mm1 \n\t" /* A3-B3 a3-b3 */\
476  "packssdw %%mm5, %%mm5 \n\t" /* A2-B2 a2-b2 */\
477  "movd %%mm6, 48+" #dst " \n\t"\
478  "movd %%mm1, 64+" #dst " \n\t"\
479  "movd %%mm5, 80+" #dst " \n\t"
480 
481 //IDCT( src0, src4, src1, src5, dst, shift)
482 IDCT( (%1), 64(%1), 32(%1), 96(%1), 0(%0), 20)
483 IDCT( 8(%1), 72(%1), 40(%1), 104(%1), 4(%0), 20)
484 IDCT( 16(%1), 80(%1), 48(%1), 112(%1), 8(%0), 20)
485 IDCT( 24(%1), 88(%1), 56(%1), 120(%1), 12(%0), 20)
486  "jmp 9f \n\t"
487 
488  "# .p2align 4 \n\t"\
489  "6: \n\t"
490 Z_COND_IDCT( 96(%0),104(%0),112(%0),120(%0), 96(%1),paddd (%2), 11, 7f)
491 
492 #undef IDCT
493 #define IDCT(src0, src4, src1, src5, dst, shift) \
494  "movq " #src0 ", %%mm0 \n\t" /* R4 R0 r4 r0 */\
495  "movq " #src5 ", %%mm3 \n\t" /* R7 R5 r7 r5 */\
496  "movq 16(%2), %%mm4 \n\t" /* C4 C4 C4 C4 */\
497  "pmaddwd %%mm0, %%mm4 \n\t" /* C4R4+C4R0 C4r4+C4r0 */\
498  "movq 24(%2), %%mm5 \n\t" /* -C4 C4 -C4 C4 */\
499  "pmaddwd %%mm5, %%mm0 \n\t" /* -C4R4+C4R0 -C4r4+C4r0 */\
500  "movq %%mm4, %%mm6 \n\t" /* C4R4+C4R0 C4r4+C4r0 */\
501  "movq %%mm0, %%mm5 \n\t" /* -C4R4+C4R0 -C4r4+C4r0 */\
502  "movq 56(%2), %%mm1 \n\t" /* C7 C5 C7 C5 */\
503  "pmaddwd %%mm3, %%mm1 \n\t" /* C7R7+C5R5 C7r7+C5r5 */\
504  "movq 72(%2), %%mm7 \n\t" /* -C5 -C1 -C5 -C1 */\
505  "pmaddwd %%mm3, %%mm7 \n\t" /* -C5R7-C1R5 -C5r7-C1r5 */\
506  "paddd %%mm4, %%mm1 \n\t" /* A0+B0 a0+b0 */\
507  "paddd %%mm4, %%mm4 \n\t" /* 2A0 2a0 */\
508  "psubd %%mm1, %%mm4 \n\t" /* A0-B0 a0-b0 */\
509  "psrad $" #shift ", %%mm1 \n\t"\
510  "psrad $" #shift ", %%mm4 \n\t"\
511  "movq %%mm0, %%mm2 \n\t" /* A1 a1 */\
512  "paddd %%mm7, %%mm0 \n\t" /* A1+B1 a1+b1 */\
513  "psubd %%mm7, %%mm2 \n\t" /* A1-B1 a1-b1 */\
514  "psrad $" #shift ", %%mm0 \n\t"\
515  "psrad $" #shift ", %%mm2 \n\t"\
516  "packssdw %%mm1, %%mm1 \n\t" /* A0+B0 a0+b0 */\
517  "movd %%mm1, " #dst " \n\t"\
518  "packssdw %%mm0, %%mm0 \n\t" /* A1+B1 a1+b1 */\
519  "movd %%mm0, 16+" #dst " \n\t"\
520  "packssdw %%mm2, %%mm2 \n\t" /* A1-B1 a1-b1 */\
521  "movd %%mm2, 96+" #dst " \n\t"\
522  "packssdw %%mm4, %%mm4 \n\t" /* A0-B0 a0-b0 */\
523  "movd %%mm4, 112+" #dst " \n\t"\
524  "movq 88(%2), %%mm1 \n\t" /* C3 C7 C3 C7 */\
525  "pmaddwd %%mm3, %%mm1 \n\t" /* C3R7+C7R5 C3r7+C7r5 */\
526  "movq %%mm5, %%mm2 \n\t" /* A2 a2 */\
527  "pmaddwd 104(%2), %%mm3 \n\t" /* -C1R7+C3R5 -C1r7+C3r5 */\
528  "paddd %%mm1, %%mm2 \n\t" /* A2+B2 a2+b2 */\
529  "psubd %%mm1, %%mm5 \n\t" /* a2-B2 a2-b2 */\
530  "psrad $" #shift ", %%mm2 \n\t"\
531  "psrad $" #shift ", %%mm5 \n\t"\
532  "movq %%mm6, %%mm1 \n\t" /* A3 a3 */\
533  "paddd %%mm3, %%mm6 \n\t" /* A3+B3 a3+b3 */\
534  "psubd %%mm3, %%mm1 \n\t" /* a3-B3 a3-b3 */\
535  "psrad $" #shift ", %%mm6 \n\t"\
536  "psrad $" #shift ", %%mm1 \n\t"\
537  "packssdw %%mm2, %%mm2 \n\t" /* A2+B2 a2+b2 */\
538  "packssdw %%mm6, %%mm6 \n\t" /* A3+B3 a3+b3 */\
539  "movd %%mm2, 32+" #dst " \n\t"\
540  "packssdw %%mm1, %%mm1 \n\t" /* A3-B3 a3-b3 */\
541  "packssdw %%mm5, %%mm5 \n\t" /* A2-B2 a2-b2 */\
542  "movd %%mm6, 48+" #dst " \n\t"\
543  "movd %%mm1, 64+" #dst " \n\t"\
544  "movd %%mm5, 80+" #dst " \n\t"
545 
546 
547 //IDCT( src0, src4, src1, src5, dst, shift)
548 IDCT( (%1), 64(%1), 32(%1), 96(%1), 0(%0), 20)
549 IDCT( 8(%1), 72(%1), 40(%1), 104(%1), 4(%0), 20)
550 IDCT( 16(%1), 80(%1), 48(%1), 112(%1), 8(%0), 20)
551 IDCT( 24(%1), 88(%1), 56(%1), 120(%1), 12(%0), 20)
552  "jmp 9f \n\t"
553 
554  "# .p2align 4 \n\t"\
555  "2: \n\t"
556 Z_COND_IDCT( 96(%0),104(%0),112(%0),120(%0), 96(%1),paddd (%2), 11, 3f)
557 
558 #undef IDCT
559 #define IDCT(src0, src4, src1, src5, dst, shift) \
560  "movq " #src0 ", %%mm0 \n\t" /* R4 R0 r4 r0 */\
561  "movq " #src1 ", %%mm2 \n\t" /* R3 R1 r3 r1 */\
562  "movq " #src5 ", %%mm3 \n\t" /* R7 R5 r7 r5 */\
563  "movq 16(%2), %%mm4 \n\t" /* C4 C4 C4 C4 */\
564  "pmaddwd %%mm0, %%mm4 \n\t" /* C4R4+C4R0 C4r4+C4r0 */\
565  "movq 24(%2), %%mm5 \n\t" /* -C4 C4 -C4 C4 */\
566  "pmaddwd %%mm5, %%mm0 \n\t" /* -C4R4+C4R0 -C4r4+C4r0 */\
567  "movq %%mm4, %%mm6 \n\t" /* C4R4+C4R0 C4r4+C4r0 */\
568  "movq 48(%2), %%mm7 \n\t" /* C3 C1 C3 C1 */\
569  "pmaddwd %%mm2, %%mm7 \n\t" /* C3R3+C1R1 C3r3+C1r1 */\
570  "movq %%mm0, %%mm5 \n\t" /* -C4R4+C4R0 -C4r4+C4r0 */\
571  "movq 56(%2), %%mm1 \n\t" /* C7 C5 C7 C5 */\
572  "pmaddwd %%mm3, %%mm1 \n\t" /* C7R7+C5R5 C7r7+C5r5 */\
573  "pmaddwd 64(%2), %%mm2 \n\t" /* -C7R3+C3R1 -C7r3+C3r1 */\
574  "paddd %%mm1, %%mm7 \n\t" /* B0 b0 */\
575  "movq 72(%2), %%mm1 \n\t" /* -C5 -C1 -C5 -C1 */\
576  "pmaddwd %%mm3, %%mm1 \n\t" /* -C5R7-C1R5 -C5r7-C1r5 */\
577  "paddd %%mm4, %%mm7 \n\t" /* A0+B0 a0+b0 */\
578  "paddd %%mm4, %%mm4 \n\t" /* 2A0 2a0 */\
579  "psubd %%mm7, %%mm4 \n\t" /* A0-B0 a0-b0 */\
580  "paddd %%mm2, %%mm1 \n\t" /* B1 b1 */\
581  "psrad $" #shift ", %%mm7 \n\t"\
582  "psrad $" #shift ", %%mm4 \n\t"\
583  "movq %%mm0, %%mm2 \n\t" /* A1 a1 */\
584  "paddd %%mm1, %%mm0 \n\t" /* A1+B1 a1+b1 */\
585  "psubd %%mm1, %%mm2 \n\t" /* A1-B1 a1-b1 */\
586  "psrad $" #shift ", %%mm0 \n\t"\
587  "psrad $" #shift ", %%mm2 \n\t"\
588  "packssdw %%mm7, %%mm7 \n\t" /* A0+B0 a0+b0 */\
589  "movd %%mm7, " #dst " \n\t"\
590  "packssdw %%mm0, %%mm0 \n\t" /* A1+B1 a1+b1 */\
591  "movd %%mm0, 16+" #dst " \n\t"\
592  "packssdw %%mm2, %%mm2 \n\t" /* A1-B1 a1-b1 */\
593  "movd %%mm2, 96+" #dst " \n\t"\
594  "packssdw %%mm4, %%mm4 \n\t" /* A0-B0 a0-b0 */\
595  "movd %%mm4, 112+" #dst " \n\t"\
596  "movq " #src1 ", %%mm0 \n\t" /* R3 R1 r3 r1 */\
597  "movq 80(%2), %%mm4 \n\t" /* -C1 C5 -C1 C5 */\
598  "pmaddwd %%mm0, %%mm4 \n\t" /* -C1R3+C5R1 -C1r3+C5r1 */\
599  "movq 88(%2), %%mm7 \n\t" /* C3 C7 C3 C7 */\
600  "pmaddwd 96(%2), %%mm0 \n\t" /* -C5R3+C7R1 -C5r3+C7r1 */\
601  "pmaddwd %%mm3, %%mm7 \n\t" /* C3R7+C7R5 C3r7+C7r5 */\
602  "movq %%mm5, %%mm2 \n\t" /* A2 a2 */\
603  "pmaddwd 104(%2), %%mm3 \n\t" /* -C1R7+C3R5 -C1r7+C3r5 */\
604  "paddd %%mm7, %%mm4 \n\t" /* B2 b2 */\
605  "paddd %%mm4, %%mm2 \n\t" /* A2+B2 a2+b2 */\
606  "psubd %%mm4, %%mm5 \n\t" /* a2-B2 a2-b2 */\
607  "psrad $" #shift ", %%mm2 \n\t"\
608  "psrad $" #shift ", %%mm5 \n\t"\
609  "movq %%mm6, %%mm4 \n\t" /* A3 a3 */\
610  "paddd %%mm0, %%mm3 \n\t" /* B3 b3 */\
611  "paddd %%mm3, %%mm6 \n\t" /* A3+B3 a3+b3 */\
612  "psubd %%mm3, %%mm4 \n\t" /* a3-B3 a3-b3 */\
613  "psrad $" #shift ", %%mm6 \n\t"\
614  "psrad $" #shift ", %%mm4 \n\t"\
615  "packssdw %%mm2, %%mm2 \n\t" /* A2+B2 a2+b2 */\
616  "packssdw %%mm6, %%mm6 \n\t" /* A3+B3 a3+b3 */\
617  "movd %%mm2, 32+" #dst " \n\t"\
618  "packssdw %%mm4, %%mm4 \n\t" /* A3-B3 a3-b3 */\
619  "packssdw %%mm5, %%mm5 \n\t" /* A2-B2 a2-b2 */\
620  "movd %%mm6, 48+" #dst " \n\t"\
621  "movd %%mm4, 64+" #dst " \n\t"\
622  "movd %%mm5, 80+" #dst " \n\t"
623 
624 //IDCT( src0, src4, src1, src5, dst, shift)
625 IDCT( (%1), 64(%1), 32(%1), 96(%1), 0(%0), 20)
626 IDCT( 8(%1), 72(%1), 40(%1), 104(%1), 4(%0), 20)
627 IDCT( 16(%1), 80(%1), 48(%1), 112(%1), 8(%0), 20)
628 IDCT( 24(%1), 88(%1), 56(%1), 120(%1), 12(%0), 20)
629  "jmp 9f \n\t"
630 
631  "# .p2align 4 \n\t"\
632  "3: \n\t"
633 #undef IDCT
634 #define IDCT(src0, src4, src1, src5, dst, shift) \
635  "movq " #src0 ", %%mm0 \n\t" /* R4 R0 r4 r0 */\
636  "movq " #src1 ", %%mm2 \n\t" /* R3 R1 r3 r1 */\
637  "movq 16(%2), %%mm4 \n\t" /* C4 C4 C4 C4 */\
638  "pmaddwd %%mm0, %%mm4 \n\t" /* C4R4+C4R0 C4r4+C4r0 */\
639  "movq 24(%2), %%mm5 \n\t" /* -C4 C4 -C4 C4 */\
640  "pmaddwd %%mm5, %%mm0 \n\t" /* -C4R4+C4R0 -C4r4+C4r0 */\
641  "movq %%mm4, %%mm6 \n\t" /* C4R4+C4R0 C4r4+C4r0 */\
642  "movq 48(%2), %%mm7 \n\t" /* C3 C1 C3 C1 */\
643  "pmaddwd %%mm2, %%mm7 \n\t" /* C3R3+C1R1 C3r3+C1r1 */\
644  "movq %%mm0, %%mm5 \n\t" /* -C4R4+C4R0 -C4r4+C4r0 */\
645  "movq 64(%2), %%mm3 \n\t"\
646  "pmaddwd %%mm2, %%mm3 \n\t" /* -C7R3+C3R1 -C7r3+C3r1 */\
647  "paddd %%mm4, %%mm7 \n\t" /* A0+B0 a0+b0 */\
648  "paddd %%mm4, %%mm4 \n\t" /* 2A0 2a0 */\
649  "psubd %%mm7, %%mm4 \n\t" /* A0-B0 a0-b0 */\
650  "psrad $" #shift ", %%mm7 \n\t"\
651  "psrad $" #shift ", %%mm4 \n\t"\
652  "movq %%mm0, %%mm1 \n\t" /* A1 a1 */\
653  "paddd %%mm3, %%mm0 \n\t" /* A1+B1 a1+b1 */\
654  "psubd %%mm3, %%mm1 \n\t" /* A1-B1 a1-b1 */\
655  "psrad $" #shift ", %%mm0 \n\t"\
656  "psrad $" #shift ", %%mm1 \n\t"\
657  "packssdw %%mm7, %%mm7 \n\t" /* A0+B0 a0+b0 */\
658  "movd %%mm7, " #dst " \n\t"\
659  "packssdw %%mm0, %%mm0 \n\t" /* A1+B1 a1+b1 */\
660  "movd %%mm0, 16+" #dst " \n\t"\
661  "packssdw %%mm1, %%mm1 \n\t" /* A1-B1 a1-b1 */\
662  "movd %%mm1, 96+" #dst " \n\t"\
663  "packssdw %%mm4, %%mm4 \n\t" /* A0-B0 a0-b0 */\
664  "movd %%mm4, 112+" #dst " \n\t"\
665  "movq 80(%2), %%mm4 \n\t" /* -C1 C5 -C1 C5 */\
666  "pmaddwd %%mm2, %%mm4 \n\t" /* -C1R3+C5R1 -C1r3+C5r1 */\
667  "pmaddwd 96(%2), %%mm2 \n\t" /* -C5R3+C7R1 -C5r3+C7r1 */\
668  "movq %%mm5, %%mm1 \n\t" /* A2 a2 */\
669  "paddd %%mm4, %%mm1 \n\t" /* A2+B2 a2+b2 */\
670  "psubd %%mm4, %%mm5 \n\t" /* a2-B2 a2-b2 */\
671  "psrad $" #shift ", %%mm1 \n\t"\
672  "psrad $" #shift ", %%mm5 \n\t"\
673  "movq %%mm6, %%mm4 \n\t" /* A3 a3 */\
674  "paddd %%mm2, %%mm6 \n\t" /* A3+B3 a3+b3 */\
675  "psubd %%mm2, %%mm4 \n\t" /* a3-B3 a3-b3 */\
676  "psrad $" #shift ", %%mm6 \n\t"\
677  "psrad $" #shift ", %%mm4 \n\t"\
678  "packssdw %%mm1, %%mm1 \n\t" /* A2+B2 a2+b2 */\
679  "packssdw %%mm6, %%mm6 \n\t" /* A3+B3 a3+b3 */\
680  "movd %%mm1, 32+" #dst " \n\t"\
681  "packssdw %%mm4, %%mm4 \n\t" /* A3-B3 a3-b3 */\
682  "packssdw %%mm5, %%mm5 \n\t" /* A2-B2 a2-b2 */\
683  "movd %%mm6, 48+" #dst " \n\t"\
684  "movd %%mm4, 64+" #dst " \n\t"\
685  "movd %%mm5, 80+" #dst " \n\t"
686 
687 
688 //IDCT( src0, src4, src1, src5, dst, shift)
689 IDCT( (%1), 64(%1), 32(%1), 96(%1), 0(%0), 20)
690 IDCT( 8(%1), 72(%1), 40(%1), 104(%1), 4(%0), 20)
691 IDCT( 16(%1), 80(%1), 48(%1), 112(%1), 8(%0), 20)
692 IDCT( 24(%1), 88(%1), 56(%1), 120(%1), 12(%0), 20)
693  "jmp 9f \n\t"
694 
695  "# .p2align 4 \n\t"\
696  "5: \n\t"
697 #undef IDCT
698 #define IDCT(src0, src4, src1, src5, dst, shift) \
699  "movq " #src0 ", %%mm0 \n\t" /* R4 R0 r4 r0 */\
700  "movq " #src4 ", %%mm1 \n\t" /* R6 R2 r6 r2 */\
701  "movq 16(%2), %%mm4 \n\t" /* C4 C4 C4 C4 */\
702  "pmaddwd %%mm0, %%mm4 \n\t" /* C4R4+C4R0 C4r4+C4r0 */\
703  "movq 24(%2), %%mm5 \n\t" /* -C4 C4 -C4 C4 */\
704  "pmaddwd %%mm5, %%mm0 \n\t" /* -C4R4+C4R0 -C4r4+C4r0 */\
705  "movq 32(%2), %%mm5 \n\t" /* C6 C2 C6 C2 */\
706  "pmaddwd %%mm1, %%mm5 \n\t" /* C6R6+C2R2 C6r6+C2r2 */\
707  "movq 40(%2), %%mm6 \n\t" /* -C2 C6 -C2 C6 */\
708  "pmaddwd %%mm6, %%mm1 \n\t" /* -C2R6+C6R2 -C2r6+C6r2 */\
709  "movq %%mm4, %%mm6 \n\t" /* C4R4+C4R0 C4r4+C4r0 */\
710  "paddd %%mm5, %%mm4 \n\t" /* A0 a0 */\
711  "psubd %%mm5, %%mm6 \n\t" /* A3 a3 */\
712  "movq %%mm0, %%mm5 \n\t" /* -C4R4+C4R0 -C4r4+C4r0 */\
713  "paddd %%mm1, %%mm0 \n\t" /* A1 a1 */\
714  "psubd %%mm1, %%mm5 \n\t" /* A2 a2 */\
715  "movq 8+" #src0 ", %%mm2 \n\t" /* R4 R0 r4 r0 */\
716  "movq 8+" #src4 ", %%mm3 \n\t" /* R6 R2 r6 r2 */\
717  "movq 16(%2), %%mm1 \n\t" /* C4 C4 C4 C4 */\
718  "pmaddwd %%mm2, %%mm1 \n\t" /* C4R4+C4R0 C4r4+C4r0 */\
719  "movq 24(%2), %%mm7 \n\t" /* -C4 C4 -C4 C4 */\
720  "pmaddwd %%mm7, %%mm2 \n\t" /* -C4R4+C4R0 -C4r4+C4r0 */\
721  "movq 32(%2), %%mm7 \n\t" /* C6 C2 C6 C2 */\
722  "pmaddwd %%mm3, %%mm7 \n\t" /* C6R6+C2R2 C6r6+C2r2 */\
723  "pmaddwd 40(%2), %%mm3 \n\t" /* -C2R6+C6R2 -C2r6+C6r2 */\
724  "paddd %%mm1, %%mm7 \n\t" /* A0 a0 */\
725  "paddd %%mm1, %%mm1 \n\t" /* 2C0 2c0 */\
726  "psubd %%mm7, %%mm1 \n\t" /* A3 a3 */\
727  "paddd %%mm2, %%mm3 \n\t" /* A1 a1 */\
728  "paddd %%mm2, %%mm2 \n\t" /* 2C1 2c1 */\
729  "psubd %%mm3, %%mm2 \n\t" /* A2 a2 */\
730  "psrad $" #shift ", %%mm4 \n\t"\
731  "psrad $" #shift ", %%mm7 \n\t"\
732  "psrad $" #shift ", %%mm3 \n\t"\
733  "packssdw %%mm7, %%mm4 \n\t" /* A0 a0 */\
734  "movq %%mm4, " #dst " \n\t"\
735  "psrad $" #shift ", %%mm0 \n\t"\
736  "packssdw %%mm3, %%mm0 \n\t" /* A1 a1 */\
737  "movq %%mm0, 16+" #dst " \n\t"\
738  "movq %%mm0, 96+" #dst " \n\t"\
739  "movq %%mm4, 112+" #dst " \n\t"\
740  "psrad $" #shift ", %%mm5 \n\t"\
741  "psrad $" #shift ", %%mm6 \n\t"\
742  "psrad $" #shift ", %%mm2 \n\t"\
743  "packssdw %%mm2, %%mm5 \n\t" /* A2-B2 a2-b2 */\
744  "movq %%mm5, 32+" #dst " \n\t"\
745  "psrad $" #shift ", %%mm1 \n\t"\
746  "packssdw %%mm1, %%mm6 \n\t" /* A3+B3 a3+b3 */\
747  "movq %%mm6, 48+" #dst " \n\t"\
748  "movq %%mm6, 64+" #dst " \n\t"\
749  "movq %%mm5, 80+" #dst " \n\t"
750 
751 
752 //IDCT( src0, src4, src1, src5, dst, shift)
753 IDCT( 0(%1), 64(%1), 32(%1), 96(%1), 0(%0), 20)
754 //IDCT( 8(%1), 72(%1), 40(%1), 104(%1), 4(%0), 20)
755 IDCT( 16(%1), 80(%1), 48(%1), 112(%1), 8(%0), 20)
756 //IDCT( 24(%1), 88(%1), 56(%1), 120(%1), 12(%0), 20)
757  "jmp 9f \n\t"
758 
759 
760  "# .p2align 4 \n\t"\
761  "1: \n\t"
762 #undef IDCT
763 #define IDCT(src0, src4, src1, src5, dst, shift) \
764  "movq " #src0 ", %%mm0 \n\t" /* R4 R0 r4 r0 */\
765  "movq " #src4 ", %%mm1 \n\t" /* R6 R2 r6 r2 */\
766  "movq " #src1 ", %%mm2 \n\t" /* R3 R1 r3 r1 */\
767  "movq 16(%2), %%mm4 \n\t" /* C4 C4 C4 C4 */\
768  "pmaddwd %%mm0, %%mm4 \n\t" /* C4R4+C4R0 C4r4+C4r0 */\
769  "movq 24(%2), %%mm5 \n\t" /* -C4 C4 -C4 C4 */\
770  "pmaddwd %%mm5, %%mm0 \n\t" /* -C4R4+C4R0 -C4r4+C4r0 */\
771  "movq 32(%2), %%mm5 \n\t" /* C6 C2 C6 C2 */\
772  "pmaddwd %%mm1, %%mm5 \n\t" /* C6R6+C2R2 C6r6+C2r2 */\
773  "movq 40(%2), %%mm6 \n\t" /* -C2 C6 -C2 C6 */\
774  "pmaddwd %%mm6, %%mm1 \n\t" /* -C2R6+C6R2 -C2r6+C6r2 */\
775  "movq %%mm4, %%mm6 \n\t" /* C4R4+C4R0 C4r4+C4r0 */\
776  "movq 48(%2), %%mm7 \n\t" /* C3 C1 C3 C1 */\
777  "pmaddwd %%mm2, %%mm7 \n\t" /* C3R3+C1R1 C3r3+C1r1 */\
778  "paddd %%mm5, %%mm4 \n\t" /* A0 a0 */\
779  "psubd %%mm5, %%mm6 \n\t" /* A3 a3 */\
780  "movq %%mm0, %%mm5 \n\t" /* -C4R4+C4R0 -C4r4+C4r0 */\
781  "paddd %%mm1, %%mm0 \n\t" /* A1 a1 */\
782  "psubd %%mm1, %%mm5 \n\t" /* A2 a2 */\
783  "movq 64(%2), %%mm1 \n\t"\
784  "pmaddwd %%mm2, %%mm1 \n\t" /* -C7R3+C3R1 -C7r3+C3r1 */\
785  "paddd %%mm4, %%mm7 \n\t" /* A0+B0 a0+b0 */\
786  "paddd %%mm4, %%mm4 \n\t" /* 2A0 2a0 */\
787  "psubd %%mm7, %%mm4 \n\t" /* A0-B0 a0-b0 */\
788  "psrad $" #shift ", %%mm7 \n\t"\
789  "psrad $" #shift ", %%mm4 \n\t"\
790  "movq %%mm0, %%mm3 \n\t" /* A1 a1 */\
791  "paddd %%mm1, %%mm0 \n\t" /* A1+B1 a1+b1 */\
792  "psubd %%mm1, %%mm3 \n\t" /* A1-B1 a1-b1 */\
793  "psrad $" #shift ", %%mm0 \n\t"\
794  "psrad $" #shift ", %%mm3 \n\t"\
795  "packssdw %%mm7, %%mm7 \n\t" /* A0+B0 a0+b0 */\
796  "movd %%mm7, " #dst " \n\t"\
797  "packssdw %%mm0, %%mm0 \n\t" /* A1+B1 a1+b1 */\
798  "movd %%mm0, 16+" #dst " \n\t"\
799  "packssdw %%mm3, %%mm3 \n\t" /* A1-B1 a1-b1 */\
800  "movd %%mm3, 96+" #dst " \n\t"\
801  "packssdw %%mm4, %%mm4 \n\t" /* A0-B0 a0-b0 */\
802  "movd %%mm4, 112+" #dst " \n\t"\
803  "movq 80(%2), %%mm4 \n\t" /* -C1 C5 -C1 C5 */\
804  "pmaddwd %%mm2, %%mm4 \n\t" /* -C1R3+C5R1 -C1r3+C5r1 */\
805  "pmaddwd 96(%2), %%mm2 \n\t" /* -C5R3+C7R1 -C5r3+C7r1 */\
806  "movq %%mm5, %%mm3 \n\t" /* A2 a2 */\
807  "paddd %%mm4, %%mm3 \n\t" /* A2+B2 a2+b2 */\
808  "psubd %%mm4, %%mm5 \n\t" /* a2-B2 a2-b2 */\
809  "psrad $" #shift ", %%mm3 \n\t"\
810  "psrad $" #shift ", %%mm5 \n\t"\
811  "movq %%mm6, %%mm4 \n\t" /* A3 a3 */\
812  "paddd %%mm2, %%mm6 \n\t" /* A3+B3 a3+b3 */\
813  "psubd %%mm2, %%mm4 \n\t" /* a3-B3 a3-b3 */\
814  "psrad $" #shift ", %%mm6 \n\t"\
815  "packssdw %%mm3, %%mm3 \n\t" /* A2+B2 a2+b2 */\
816  "movd %%mm3, 32+" #dst " \n\t"\
817  "psrad $" #shift ", %%mm4 \n\t"\
818  "packssdw %%mm6, %%mm6 \n\t" /* A3+B3 a3+b3 */\
819  "movd %%mm6, 48+" #dst " \n\t"\
820  "packssdw %%mm4, %%mm4 \n\t" /* A3-B3 a3-b3 */\
821  "packssdw %%mm5, %%mm5 \n\t" /* A2-B2 a2-b2 */\
822  "movd %%mm4, 64+" #dst " \n\t"\
823  "movd %%mm5, 80+" #dst " \n\t"
824 
825 
826 //IDCT( src0, src4, src1, src5, dst, shift)
827 IDCT( (%1), 64(%1), 32(%1), 96(%1), 0(%0), 20)
828 IDCT( 8(%1), 72(%1), 40(%1), 104(%1), 4(%0), 20)
829 IDCT( 16(%1), 80(%1), 48(%1), 112(%1), 8(%0), 20)
830 IDCT( 24(%1), 88(%1), 56(%1), 120(%1), 12(%0), 20)
831  "jmp 9f \n\t"
832 
833 
834  "# .p2align 4 \n\t"
835  "7: \n\t"
836 #undef IDCT
837 #define IDCT(src0, src4, src1, src5, dst, shift) \
838  "movq " #src0 ", %%mm0 \n\t" /* R4 R0 r4 r0 */\
839  "movq 16(%2), %%mm4 \n\t" /* C4 C4 C4 C4 */\
840  "pmaddwd %%mm0, %%mm4 \n\t" /* C4R4+C4R0 C4r4+C4r0 */\
841  "movq 24(%2), %%mm5 \n\t" /* -C4 C4 -C4 C4 */\
842  "pmaddwd %%mm5, %%mm0 \n\t" /* -C4R4+C4R0 -C4r4+C4r0 */\
843  "psrad $" #shift ", %%mm4 \n\t"\
844  "psrad $" #shift ", %%mm0 \n\t"\
845  "movq 8+" #src0 ", %%mm2 \n\t" /* R4 R0 r4 r0 */\
846  "movq 16(%2), %%mm1 \n\t" /* C4 C4 C4 C4 */\
847  "pmaddwd %%mm2, %%mm1 \n\t" /* C4R4+C4R0 C4r4+C4r0 */\
848  "movq 24(%2), %%mm7 \n\t" /* -C4 C4 -C4 C4 */\
849  "pmaddwd %%mm7, %%mm2 \n\t" /* -C4R4+C4R0 -C4r4+C4r0 */\
850  "movq 32(%2), %%mm7 \n\t" /* C6 C2 C6 C2 */\
851  "psrad $" #shift ", %%mm1 \n\t"\
852  "packssdw %%mm1, %%mm4 \n\t" /* A0 a0 */\
853  "movq %%mm4, " #dst " \n\t"\
854  "psrad $" #shift ", %%mm2 \n\t"\
855  "packssdw %%mm2, %%mm0 \n\t" /* A1 a1 */\
856  "movq %%mm0, 16+" #dst " \n\t"\
857  "movq %%mm0, 96+" #dst " \n\t"\
858  "movq %%mm4, 112+" #dst " \n\t"\
859  "movq %%mm0, 32+" #dst " \n\t"\
860  "movq %%mm4, 48+" #dst " \n\t"\
861  "movq %%mm4, 64+" #dst " \n\t"\
862  "movq %%mm0, 80+" #dst " \n\t"
863 
864 //IDCT( src0, src4, src1, src5, dst, shift)
865 IDCT( 0(%1), 64(%1), 32(%1), 96(%1), 0(%0), 20)
866 //IDCT( 8(%1), 72(%1), 40(%1), 104(%1), 4(%0), 20)
867 IDCT( 16(%1), 80(%1), 48(%1), 112(%1), 8(%0), 20)
868 //IDCT( 24(%1), 88(%1), 56(%1), 120(%1), 12(%0), 20)
869 
870 
871 /*
872 Input
873  00 40 04 44 20 60 24 64
874  10 30 14 34 50 70 54 74
875  01 41 03 43 21 61 23 63
876  11 31 13 33 51 71 53 73
877  02 42 06 46 22 62 26 66
878  12 32 16 36 52 72 56 76
879  05 45 07 47 25 65 27 67
880  15 35 17 37 55 75 57 77
881 
882 Temp
883  00 04 10 14 20 24 30 34
884  40 44 50 54 60 64 70 74
885  01 03 11 13 21 23 31 33
886  41 43 51 53 61 63 71 73
887  02 06 12 16 22 26 32 36
888  42 46 52 56 62 66 72 76
889  05 07 15 17 25 27 35 37
890  45 47 55 57 65 67 75 77
891 */
892 
893 "9: \n\t"
894  :: "r" (block), "r" (temp), "r" (coeffs)
895  NAMED_CONSTRAINTS_ADD(wm1010,d40000)
896  : "%eax"
897  );
898 }
899 
900 void ff_simple_idct_mmx(int16_t *block)
901 {
902  idct(block);
903 }
904 
905 //FIXME merge add/put into the idct
906 
907 void ff_simple_idct_put_mmx(uint8_t *dest, ptrdiff_t line_size, int16_t *block)
908 {
909  idct(block);
910  ff_put_pixels_clamped(block, dest, line_size);
911 }
912 void ff_simple_idct_add_mmx(uint8_t *dest, ptrdiff_t line_size, int16_t *block)
913 {
914  idct(block);
915  ff_add_pixels_clamped(block, dest, line_size);
916 }
917 
918 #endif /* HAVE_INLINE_ASM */
ff_put_pixels_clamped
void(* ff_put_pixels_clamped)(const int16_t *block, uint8_t *pixels, ptrdiff_t line_size)
Definition: idctdsp.c:83
DC_COND_IDCT
#define DC_COND_IDCT(src0, src4, src1, src5, dst, rounder, rarg, shift)
simple_idct.h
idct
static void idct(int16_t block[64])
Definition: 4xm.c:163
mem.h
Memory handling functions.
temp
else temp
Definition: vf_mcdeint.c:259
C2
#define C2
Definition: simple_idct.c:137
block
static int16_t block[64]
Definition: dct.c:115
ROW_SHIFT
#define ROW_SHIFT
Definition: simple_idct_alpha.c:42
uint8_t
uint8_t
Definition: audio_convert.c:194
ff_add_pixels_clamped
void(* ff_add_pixels_clamped)(const int16_t *block, uint8_t *pixels, ptrdiff_t line_size)
Definition: idctdsp.c:84
C5
#define C5
Definition: simple_idct_mmi.c:33
DECLARE_ALIGNED
#define DECLARE_ALIGNED(n, t, v)
Declare a variable that is aligned in memory.
Definition: mem.h:104
C6
#define C6
Definition: simple_idct_mmi.c:34
C1
#define C1
Definition: simple_idct.c:136
LOCAL_ALIGNED_8
#define LOCAL_ALIGNED_8(t, v,...)
Definition: internal.h:118
Z_COND_IDCT
#define Z_COND_IDCT(src0, src4, src1, src5, dst, rounder, rarg, shift, bt)
r
const char * r
Definition: vf_curves.c:111
DECLARE_ASM_CONST
#define DECLARE_ASM_CONST(n, t, v)
Declare a static constant aligned variable appropriate for use in inline assembly code...
Definition: mem.h:105
ff_simple_idct_add_mmx
void ff_simple_idct_add_mmx(uint8_t *dest, ptrdiff_t line_size, int16_t *block)
IDCT
#define IDCT(H)
Definition: hevcdsp_template.c:240
n
int n
Definition: avisynth_c.h:684
C4
#define C4
Definition: simple_idct_mmi.c:32
ff_simple_idct_mmx
void ff_simple_idct_mmx(int16_t *block)
C3
#define C3
Definition: simple_idct.c:138
ff_simple_idct_put_mmx
void ff_simple_idct_put_mmx(uint8_t *dest, ptrdiff_t line_size, int16_t *block)
coeffs
static const int16_t coeffs[]
Definition: simple_idct_mmi.c:40
C7
#define C7
Definition: simple_idct_mmi.c:35
NAMED_CONSTRAINTS_ADD
#define NAMED_CONSTRAINTS_ADD(...)
Definition: asm.h:145
Generated on Fri Jan 12 2018 01:45:46 for FFmpeg by   doxygen 1.8.6