| blob | 7253d890bf2595631b9597ccaf6528df1290651b |
1 /***************************************************************************
2 * __________ __ ___.
3 * Open \______ \ ____ ____ | | _\_ |__ _______ ___
4 * Source | _// _ \_/ ___\| |/ /| __ \ / _ \ \/ /
5 * Jukebox | | ( <_> ) \___| < | \_\ ( <_> > < <
6 * Firmware |____|_ /\____/ \___ >__|_ \|___ /\____/__/\_ \
7 * \/ \/ \/ \/ \/
8 * $Id$
9 *
10 * Copyright (C) 2007 by Michael Sevakis
11 *
12 * This program is free software; you can redistribute it and/or
13 * modify it under the terms of the GNU General Public License
14 * as published by the Free Software Foundation; either version 2
15 * of the License, or (at your option) any later version.
16 *
17 * This software is distributed on an "AS IS" basis, WITHOUT WARRANTY OF ANY
18 * KIND, either express or implied.
19 *
20 ****************************************************************************/
22 .global mpeg2_idct_copy
23 .type mpeg2_idct_copy, %function
24 .global mpeg2_idct_add
25 .type mpeg2_idct_add, %function
28 /* Custom calling convention:
29 * r0 contains block pointer and is non-volatile
30 * all non-volatile c context saved and restored on its behalf
31 */
32 .idct:
33 add r12, r0, #128
34 1:
35 ldrsh r1, [r0, #0] /* d0 */
36 ldrsh r2, [r0, #2] /* d1 */
37 ldrsh r3, [r0, #4] /* d2 */
38 ldrsh r4, [r0, #6] /* d3 */
39 ldrsh r5, [r0, #8] /* d0 */
40 ldrsh r6, [r0, #10] /* d1 */
41 ldrsh r7, [r0, #12] /* d2 */
42 ldrsh r8, [r0, #14] /* d3 */
43 orrs r9, r2, r3
44 orreqs r9, r4, r5
45 orreqs r9, r6, r7
46 cmpeq r8, #0
47 bne 2f
48 mov r1, r1, asl #15
49 bic r1, r1, #0x8000
50 orr r1, r1, r1, lsr #16
51 str r1, [r0], #4
52 str r1, [r0], #4
53 str r1, [r0], #4
54 str r1, [r0], #4
55 cmp r0, r12
56 blo 1b
57 b 3f
58 2:
59 mov r1, r1, asl #11 /* r1 = d0 = (block[0] << 11) + 2048 */
60 add r1, r1, #2048
61 add r1, r1, r3, asl #11 /* r1 = t0 = d0 + (block[2] << 11) */
62 sub r3, r1, r3, asl #12 /* r3 = t1 = d0 - (block[2] << 11) */
64 add r9, r2, r4 /* r9 = tmp = (d1+d3)*(1108/4) */
65 add r10, r9, r9, asl #2
66 add r10, r10, r9, asl #4
67 add r9, r10, r9, asl #8
69 add r10, r2, r2, asl #4 /* r2 = t2 = tmp + (d1*(1568/32)*8) */
70 add r2, r10, r2, asl #5
71 add r2, r9, r2, asl #3
73 add r10, r4, r4, asl #2 /* r4 = t3 = tmp - (d3*(3784/8)*2) */
74 rsb r10, r10, r4, asl #6
75 add r4, r4, r10, asl #3
76 sub r4, r9, r4, asl #1
77 /* t2 & t3 are 1/4 final value here */
78 add r1, r1, r2, asl #2 /* r1 = a0 = t0 + t2 */
79 sub r2, r1, r2, asl #3 /* r2 = a3 = t0 - t2 */
80 add r3, r3, r4, asl #2 /* r3 = a1 = t1 + t3 */
81 sub r4, r3, r4, asl #3 /* r4 = a2 = t1 - t3 */
83 add r9, r8, r5 /* r9 = tmp = 565*(d3 + d0) */
84 add r10, r9, r9, asl #4
85 add r10, r10, r10, asl #5
86 add r9, r10, r9, asl #2
88 add r10, r5, r5, asl #4 /* r5 = t0 = tmp + (((2276/4)*d0)*4) */
89 add r10, r10, r10, asl #5
90 add r5, r10, r5, asl #3
91 add r5, r9, r5, asl #2
93 add r10, r8, r8, asl #2 /* r8 = t1 = tmp - (((3406/2)*d3)*2) */
94 add r10, r10, r10, asl #4
95 add r10, r10, r8, asl #7
96 rsb r8, r8, r10, asl #3
97 sub r8, r9, r8, asl #1
99 add r9, r6, r7 /* r9 = tmp = (2408/8)*(d1 + d2) */
100 add r10, r9, r9, asl #3
101 add r10, r10, r10, asl #5
102 add r9, r10, r9, asl #2
104 add r10, r7, r7, asl #3 /* r7 = t2 = (tmp*8) - 799*d2 */
105 add r10, r10, r7, asl #4
106 rsb r7, r7, r10, asl #5
107 rsb r7, r7, r9, asl #3
109 sub r10, r6, r6, asl #4 /* r6 = t3 = (tmp*8) - 4017*d1 */
110 sub r10, r10, r6, asl #6
111 add r10, r10, r6, asl #12
112 add r6, r10, r6
113 rsb r6, r6, r9, asl #3
114 /* t0 = r5, t1 = r8, t2 = r7, t3 = r6*/
115 add r9, r5, r7 /* r9 = b0 = t0 + t2 */
116 add r10, r8, r6 /* r10 = b3 = t1 + t3 */
117 sub r5, r5, r7 /* t0 -= t2 */
118 sub r8, r8, r6 /* t1 -= t3 */
119 add r6, r5, r8 /* r6 = t0 + t1 */
120 sub r7, r5, r8 /* r7 = t0 - t1 */
122 add r11, r6, r6, asr #2 /* r6 = b1 = r6*(181/128) */
123 add r11, r11, r11, asr #5
124 add r6, r11, r6, asr #3
125 add r11, r7, r7, asr #2 /* r7 = b2 = r7*(181/128) */
126 add r11, r11, r11, asr #5
127 add r7, r11, r7, asr #3
128 /* r1 = a0, r3 = a1, r4 = a2, r2 = a3 */
129 /* r9 = b0, r6 = b1*2, r7 = b2*2, r10 = b3 */
130 add r5, r1, r9 /* block[0] = (a0 + b0) >> 12 */
131 mov r5, r5, asr #12
132 strh r5, [r0], #2
133 add r8, r3, r6, asr #1 /* block[1] = (a1 + b1) >> 12 */
134 mov r8, r8, asr #12
135 strh r8, [r0], #2
136 add r5, r4, r7, asr #1 /* block[2] = (a2 + b2) >> 12 */
137 mov r5, r5, asr #12
138 strh r5, [r0], #2
139 add r8, r2, r10 /* block[3] = (a3 + b3) >> 12 */
140 mov r8, r8, asr #12
141 strh r8, [r0], #2
142 sub r5, r2, r10 /* block[4] = (a3 - b3) >> 12 */
143 mov r5, r5, asr #12
144 strh r5, [r0], #2
145 sub r8, r4, r7, asr #1 /* block[5] = (a2 - b2) >> 12 */
146 mov r8, r8, asr #12
147 strh r8, [r0], #2
148 sub r5, r3, r6, asr #1 /* block[6] = (a1 - b1) >> 12 */
149 mov r5, r5, asr #12
150 strh r5, [r0], #2
151 sub r8, r1, r9 /* block[7] = (a0 - b0) >> 12 */
152 mov r8, r8, asr #12
153 strh r8, [r0], #2
154 cmp r0, r12
155 blo 1b
156 3:
157 sub r0, r0, #128
158 add r12, r0, #16
159 4:
160 ldrsh r1, [r0, #0*8] /* d0 */
161 ldrsh r2, [r0, #2*8] /* d1 */
162 ldrsh r3, [r0, #4*8] /* d2 */
163 ldrsh r4, [r0, #6*8] /* d3 */
164 ldrsh r5, [r0, #8*8] /* d0 */
165 ldrsh r6, [r0, #10*8] /* d1 */
166 ldrsh r7, [r0, #12*8] /* d2 */
167 ldrsh r8, [r0, #14*8] /* d3 */
169 mov r1, r1, asl #11 /* r1 = d0 = (block[0] << 11) + 2048 */
170 add r1, r1, #65536
171 add r1, r1, r3, asl #11 /* r1 = t0 = d0 + d2:(block[2] << 11) */
172 sub r3, r1, r3, asl #12 /* r3 = t1 = d0 - d2:(block[2] << 11) */
174 add r9, r2, r4 /* r9 = tmp = (d1+d3)*(1108/4) */
175 add r10, r9, r9, asl #2
176 add r10, r10, r9, asl #4
177 add r9, r10, r9, asl #8
179 add r11, r2, r2, asl #4 /* r2 = t2 = tmp + (d1*(1568/32)*8) */
180 add r2, r11, r2, asl #5
181 add r2, r9, r2, asl #3
183 add r10, r4, r4, asl #2 /* r4 = t3 = tmp - (d3*(3784/8)*2) */
184 rsb r10, r10, r4, asl #6
185 add r4, r4, r10, asl #3
186 sub r4, r9, r4, asl #1
187 /* t2 & t3 are 1/4 final value here */
188 add r1, r1, r2, asl #2 /* r1 = a0 = t0 + t2 */
189 sub r2, r1, r2, asl #3 /* r2 = a3 = t0 - t2 */
190 add r3, r3, r4, asl #2 /* r3 = a1 = t1 + t3 */
191 sub r4, r3, r4, asl #3 /* r4 = a2 = t1 - t3 */
193 add r9, r8, r5 /* r9 = tmp = 565*(d3 + d0) */
194 add r10, r9, r9, asl #4
195 add r10, r10, r10, asl #5
196 add r9, r10, r9, asl #2
198 add r10, r5, r5, asl #4 /* r5 = t0 = tmp + (((2276/4)*d0)*4) */
199 add r10, r10, r10, asl #5
200 add r5, r10, r5, asl #3
201 add r5, r9, r5, asl #2
203 add r10, r8, r8, asl #2 /* r8 = t1 = tmp - (((3406/2)*d3)*2) */
204 add r10, r10, r10, asl #4
205 add r10, r10, r8, asl #7
206 rsb r8, r8, r10, asl #3
207 sub r8, r9, r8, asl #1
209 add r9, r6, r7 /* r9 = tmp = (2408/8)*(d1 + d2) */
210 add r10, r9, r9, asl #3
211 add r10, r10, r10, asl #5
212 add r9, r10, r9, asl #2
214 add r10, r7, r7, asl #3 /* r7 = t2 = (tmp*8) - 799*d2 */
215 add r10, r10, r7, asl #4
216 rsb r7, r7, r10, asl #5
217 rsb r7, r7, r9, asl #3
219 sub r10, r6, r6, asl #4 /* r6 = t3 = (tmp*8) - 4017*d1 */
220 sub r10, r10, r6, asl #6
221 add r10, r10, r6, asl #12
222 add r6, r10, r6
223 rsb r6, r6, r9, asl #3
224 /* t0=r5, t1=r8, t2=r7, t3=r6*/
225 add r9, r5, r7 /* r9 = b0 = t0 + t2 */
226 add r10, r8, r6 /* r10 = b3 = t1 + t3 */
227 sub r5, r5, r7 /* t0 -= t2 */
228 sub r8, r8, r6 /* t1 -= t3 */
229 add r6, r5, r8 /* r6 = t0 + t1 */
230 sub r7, r5, r8 /* r7 = t0 - t1 */
232 add r11, r6, r6, asr #2 /* r6 = b1 = r5*(181/128) */
233 add r11, r11, r11, asr #5
234 add r6, r11, r6, asr #3
235 add r11, r7, r7, asr #2 /* r7 = b2 = r6*(181/128) */
236 add r11, r11, r11, asr #5
237 add r7, r11, r7, asr #3
238 /* r1 = a0, r3 = a1, r4 = a2, r2 = a3 */
239 /* r9 = b0, r6 = b1*2, r7 = b2*2, r10 = b3 */
240 add r5, r1, r9 /* block[0] = (a0 + b0) >> 17 */
241 mov r5, r5, asr #17
242 strh r5, [r0, #0*8]
243 add r8, r3, r6, asr #1 /* block[1] = (a1 + b1) >> 17 */
244 mov r8, r8, asr #17
245 strh r8, [r0, #2*8]
246 add r5, r4, r7, asr #1 /* block[2] = (a2 + b2) >> 17 */
247 mov r5, r5, asr #17
248 strh r5, [r0, #4*8]
249 add r8, r2, r10 /* block[3] = (a3 + b3) >> 17 */
250 mov r8, r8, asr #17
251 strh r8, [r0, #6*8]
252 sub r5, r2, r10 /* block[4] = (a3 - b3) >> 17 */
253 mov r5, r5, asr #17
254 strh r5, [r0, #8*8]
255 sub r8, r4, r7, asr #1 /* block[5] = (a2 - b2) >> 17 */
256 mov r8, r8, asr #17
257 strh r8, [r0, #10*8]
258 sub r5, r3, r6, asr #1 /* block[6] = (a1 - b1) >> 17 */
259 mov r5, r5, asr #17
260 strh r5, [r0, #12*8]
261 sub r8, r1, r9 /* block[7] = (a0 - b0) >> 17 */
262 mov r8, r8, asr #17
263 strh r8, [r0, #14*8]
264 add r0, r0, #2
265 cmp r0, r12
266 blo 4b
267 sub r0, r0, #16
268 bx lr
270 mpeg2_idct_copy:
271 stmfd sp!, { r1-r2, r4-r11, lr }
272 bl .idct
273 ldmfd sp!, { r1-r2 }
274 mov r11, #0
275 add r12, r0, #128
276 1:
277 ldrsh r3, [r0, #0]
278 ldrsh r4, [r0, #2]
279 ldrsh r5, [r0, #4]
280 ldrsh r6, [r0, #6]
281 ldrsh r7, [r0, #8]
282 ldrsh r8, [r0, #10]
283 ldrsh r9, [r0, #12]
284 ldrsh r10, [r0, #14]
285 cmp r3, #255
286 mvnhi r3, r3, asr #31
287 strb r3, [r1, #0]
288 str r11, [r0], #4
289 cmp r4, #255
290 mvnhi r4, r4, asr #31
291 strb r4, [r1, #1]
292 cmp r5, #255
293 mvnhi r5, r5, asr #31
294 strb r5, [r1, #2]
295 str r11, [r0], #4
296 cmp r6, #255
297 mvnhi r6, r6, asr #31
298 strb r6, [r1, #3]
299 cmp r7, #255
300 mvnhi r7, r7, asr #31
301 strb r7, [r1, #4]
302 str r11, [r0], #4
303 cmp r8, #255
304 mvnhi r8, r8, asr #31
305 strb r8, [r1, #5]
306 cmp r9, #255
307 mvnhi r9, r9, asr #31
308 strb r9, [r1, #6]
309 str r11, [r0], #4
310 cmp r10, #255
311 mvnhi r10, r10, asr #31
312 strb r10, [r1, #7]
313 add r1, r1, r2
314 cmp r0, r12
315 blo 1b
316 ldmfd sp!, { r4-r11, pc }
318 mpeg2_idct_add:
319 cmp r0, #129
320 mov r0, r1
321 ldreqsh r1, [r0, #0]
322 bne 1f
323 and r1, r1, #0x70
324 cmp r1, #0x40
325 bne 3f
326 1:
327 stmfd sp!, { r2-r11, lr }
328 bl .idct
329 ldmfd sp!, { r1-r2 }
330 mov r11, #0
331 add r12, r0, #128
332 2:
333 ldrb r3, [r1, #0]
334 ldrb r4, [r1, #1]
335 ldrb r5, [r1, #2]
336 ldrb r6, [r1, #3]
337 ldrsh r7, [r0, #0]
338 ldrsh r8, [r0, #2]
339 ldrsh r9, [r0, #4]
340 ldrsh r10, [r0, #6]
341 add r7, r7, r3
342 ldrb r3, [r1, #4]
343 cmp r7, #255
344 mvnhi r7, r7, asr #31
345 strb r7, [r1, #0]
346 ldrsh r7, [r0, #8]
347 add r8, r8, r4
348 ldrb r4, [r1, #5]
349 cmp r8, #255
350 mvnhi r8, r8, asr #31
351 strb r8, [r1, #1]
352 ldrsh r8, [r0, #10]
353 add r9, r9, r5
354 ldrb r5, [r1, #6]
355 cmp r9, #255
356 mvnhi r9, r9, asr #31
357 strb r9, [r1, #2]
358 ldrsh r9, [r0, #12]
359 add r10, r10, r6
360 ldrb r6, [r1, #7]
361 cmp r10, #255
362 mvnhi r10, r10, asr #31
363 strb r10, [r1, #3]
364 ldrsh r10, [r0, #14]
365 str r11, [r0], #4
366 add r7, r7, r3
367 cmp r7, #255
368 mvnhi r7, r7, asr #31
369 strb r7, [r1, #4]
370 str r11, [r0], #4
371 add r8, r8, r4
372 cmp r8, #255
373 mvnhi r8, r8, asr #31
374 strb r8, [r1, #5]
375 str r11, [r0], #4
376 add r9, r9, r5
377 cmp r9, #255
378 mvnhi r9, r9, asr #31
379 strb r9, [r1, #6]
380 add r10, r10, r6
381 cmp r10, #255
382 mvnhi r10, r10, asr #31
383 strb r10, [r1, #7]
384 str r11, [r0], #4
385 add r1, r1, r2
386 cmp r0, r12
387 blo 2b
388 ldmfd sp!, { r4-r11, pc }
389 3:
390 stmfd sp!, { r4-r5, lr }
391 ldrsh r1, [r0, #0] /* r1 = block[0] */
392 mov r4, #0
393 strh r4, [r0, #0] /* block[0] = 0 */
394 strh r4, [r0, #126] /* block[63] = 0 */
395 add r1, r1, #64 /* r1 = DC << 7 */
396 add r0, r2, r3, asl #3
397 4:
398 ldrb r4, [r2, #0]
399 ldrb r5, [r2, #1]
400 ldrb r12, [r2, #2]
401 ldrb lr, [r2, #3]
402 add r4, r4, r1, asr #7
403 cmp r4, #255
404 mvnhi r4, r4, asr #31
405 strb r4, [r2, #0]
406 add r5, r5, r1, asr #7
407 cmp r5, #255
408 mvnhi r5, r5, asr #31
409 strb r5, [r2, #1]
410 add r12, r12, r1, asr #7
411 cmp r12, #255
412 mvnhi r12, r12, asr #31
413 strb r12, [r2, #2]
414 add lr, lr, r1, asr #7
415 cmp lr, #255
416 mvnhi lr, lr, asr #31
417 strb lr, [r2, #3]
418 ldrb r4, [r2, #4]
419 ldrb r5, [r2, #5]
420 ldrb r12, [r2, #6]
421 ldrb lr, [r2, #7]
422 add r4, r4, r1, asr #7
423 cmp r4, #255
424 mvnhi r4, r4, asr #31
425 strb r4, [r2, #4]
426 add r5, r5, r1, asr #7
427 cmp r5, #255
428 mvnhi r5, r5, asr #31
429 strb r5, [r2, #5]
430 add r12, r12, r1, asr #7
431 cmp r12, #255
432 mvnhi r12, r12, asr #31
433 strb r12, [r2, #6]
434 add lr, lr, r1, asr #7
435 cmp lr, #255
436 mvnhi lr, lr, asr #31
437 strb lr, [r2, #7]
438 add r2, r2, r3
439 cmp r2, r0
440 blo 4b
441 ldmfd sp!, { r4-r5, pc }