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
249
250
251
252
|
/****************************************************************************
* __________ __ ___.
* Open \______ \ ____ ____ | | _\_ |__ _______ ___
* Source | _// _ \_/ ___\| |/ /| __ \ / _ \ \/ /
* Jukebox | | ( <_> ) \___| < | \_\ ( <_> > < <
* Firmware |____|_ /\____/ \___ >__|_ \|___ /\____/__/\_ \
* \/ \/ \/ \/ \/
*
* Copyright (C) 2007 Michael Giacomelli
*
* 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 software is distributed on an "AS IS" basis, WITHOUT WARRANTY OF ANY
* KIND, either express or implied.
*
****************************************************************************/
#include "wmadec.h"
#include "wmafixed.h"
#include <codecs.h>
fixed64 IntTo64(int x){
fixed64 res = 0;
unsigned char *p = (unsigned char *)&res;
#ifdef ROCKBOX_BIG_ENDIAN
p[5] = x & 0xff;
p[4] = (x & 0xff00)>>8;
p[3] = (x & 0xff0000)>>16;
p[2] = (x & 0xff000000)>>24;
#else
p[2] = x & 0xff;
p[3] = (x & 0xff00)>>8;
p[4] = (x & 0xff0000)>>16;
p[5] = (x & 0xff000000)>>24;
#endif
return res;
}
int IntFrom64(fixed64 x)
{
int res = 0;
unsigned char *p = (unsigned char *)&x;
#ifdef ROCKBOX_BIG_ENDIAN
res = p[5] | (p[4]<<8) | (p[3]<<16) | (p[2]<<24);
#else
res = p[2] | (p[3]<<8) | (p[4]<<16) | (p[5]<<24);
#endif
return res;
}
fixed32 Fixed32From64(fixed64 x)
{
return x & 0xFFFFFFFF;
}
fixed64 Fixed32To64(fixed32 x)
{
return (fixed64)x;
}
/*
* Helper functions for wma_window.
*
*
*/
#ifdef CPU_ARM
inline void vector_fmul_add_add(fixed32 *dst, const fixed32 *data,
const fixed32 *window, int n)
{
/* Block sizes are always power of two */
asm volatile (
"0:"
"ldmia %[d]!, {r0, r1};"
"ldmia %[w]!, {r4, r5};"
/* consume the first data and window value so we can use those
* registers again */
"smull r8, r9, r0, r4;"
"ldmia %[dst], {r0, r4};"
"add r0, r0, r9, lsl #1;" /* *dst=*dst+(r9<<1)*/
"smull r8, r9, r1, r5;"
"add r1, r4, r9, lsl #1;"
"stmia %[dst]!, {r0, r1};"
"subs %[n], %[n], #2;"
"bne 0b;"
: [d] "+r" (data), [w] "+r" (window), [dst] "+r" (dst), [n] "+r" (n)
: : "r0", "r1", "r4", "r5", "r8", "r9", "memory", "cc");
}
inline void vector_fmul_reverse(fixed32 *dst, const fixed32 *src0, const fixed32 *src1,
int len)
{
/* Block sizes are always power of two */
asm volatile (
"add %[s1], %[s1], %[n], lsl #2;"
"0:"
"ldmia %[s0]!, {r0, r1};"
"ldmdb %[s1]!, {r4, r5};"
"smull r8, r9, r0, r5;"
"mov r0, r9, lsl #1;"
"smull r8, r9, r1, r4;"
"mov r1, r9, lsl #1;"
"stmia %[dst]!, {r0, r1};"
"subs %[n], %[n], #2;"
"bne 0b;"
: [s0] "+r" (src0), [s1] "+r" (src1), [dst] "+r" (dst), [n] "+r" (len)
: : "r0", "r1", "r4", "r5", "r8", "r9", "memory", "cc");
}
#elif defined(CPU_COLDFIRE)
inline void vector_fmul_add_add(fixed32 *dst, const fixed32 *data,
const fixed32 *window, int n)
{
/* Block sizes are always power of two. Smallest block is always way bigger
* than four too.*/
asm volatile (
"0:"
"movem.l (%[d]), %%d0-%%d3;"
"movem.l (%[w]), %%d4-%%d5/%%a0-%%a1;"
"mac.l %%d0, %%d4, %%acc0;"
"mac.l %%d1, %%d5, %%acc1;"
"mac.l %%d2, %%a0, %%acc2;"
"mac.l %%d3, %%a1, %%acc3;"
"lea.l (16, %[d]), %[d];"
"lea.l (16, %[w]), %[w];"
"movclr.l %%acc0, %%d0;"
"movclr.l %%acc1, %%d1;"
"movclr.l %%acc2, %%d2;"
"movclr.l %%acc3, %%d3;"
"movem.l (%[dst]), %%d4-%%d5/%%a0-%%a1;"
"add.l %%d4, %%d0;"
"add.l %%d5, %%d1;"
"add.l %%a0, %%d2;"
"add.l %%a1, %%d3;"
"movem.l %%d0-%%d3, (%[dst]);"
"lea.l (16, %[dst]), %[dst];"
"subq.l #4, %[n];"
"jne 0b;"
: [d] "+a" (data), [w] "+a" (window), [dst] "+a" (dst), [n] "+d" (n)
: : "d0", "d1", "d2", "d3", "d4", "d5", "a0", "a1", "memory", "cc");
}
inline void vector_fmul_reverse(fixed32 *dst, const fixed32 *src0, const fixed32 *src1,
int len)
{
/* Block sizes are always power of two. Smallest block is always way bigger
* than four too.*/
asm volatile (
"lea.l (-16, %[s1], %[n]*4), %[s1];"
"0:"
"movem.l (%[s0]), %%d0-%%d3;"
"movem.l (%[s1]), %%d4-%%d5/%%a0-%%a1;"
"mac.l %%d0, %%a1, %%acc0;"
"mac.l %%d1, %%a0, %%acc1;"
"mac.l %%d2, %%d5, %%acc2;"
"mac.l %%d3, %%d4, %%acc3;"
"lea.l (16, %[s0]), %[s0];"
"lea.l (-16, %[s1]), %[s1];"
"movclr.l %%acc0, %%d0;"
"movclr.l %%acc1, %%d1;"
"movclr.l %%acc2, %%d2;"
"movclr.l %%acc3, %%d3;"
"movem.l %%d0-%%d3, (%[dst]);"
"lea.l (16, %[dst]), %[dst];"
"subq.l #4, %[n];"
"jne 0b;"
: [s0] "+a" (src0), [s1] "+a" (src1), [dst] "+a" (dst), [n] "+d" (len)
: : "d0", "d1", "d2", "d3", "d4", "d5", "a0", "a1", "memory", "cc");
}
#else
inline void vector_fmul_add_add(fixed32 *dst, const fixed32 *src0, const fixed32 *src1, int len){
int i;
for(i=0; i<len; i++)
dst[i] = fixmul32b(src0[i], src1[i]) + dst[i];
}
inline void vector_fmul_reverse(fixed32 *dst, const fixed32 *src0, const fixed32 *src1, int len){
int i;
src1 += len-1;
for(i=0; i<len; i++)
dst[i] = fixmul32b(src0[i], src1[-i]);
}
#endif
/*
Not performance senstitive code here
*/
fixed32 fixdiv32(fixed32 x, fixed32 y)
{
fixed64 temp;
if(x == 0)
return 0;
if(y == 0)
return 0x7fffffff;
temp = x;
temp <<= PRECISION;
return (fixed32)(temp / y);
}
fixed64 fixdiv64(fixed64 x, fixed64 y)
{
fixed64 temp;
if(x == 0)
return 0;
if(y == 0)
return 0x07ffffffffffffffLL;
temp = x;
temp <<= PRECISION64;
return (fixed64)(temp / y);
}
fixed32 fixsqrt32(fixed32 x)
{
unsigned long r = 0, s, v = (unsigned long)x;
#define STEP(k) s = r + (1 << k * 2); r >>= 1; \
if (s <= v) { v -= s; r |= (1 << k * 2); }
STEP(15);
STEP(14);
STEP(13);
STEP(12);
STEP(11);
STEP(10);
STEP(9);
STEP(8);
STEP(7);
STEP(6);
STEP(5);
STEP(4);
STEP(3);
STEP(2);
STEP(1);
STEP(0);
return (fixed32)(r << (PRECISION / 2));
}
|
