asm_arm.h 4.82 KB
Newer Older
Monty's avatar
 
Monty committed
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
/********************************************************************
 *                                                                  *
 * THIS FILE IS PART OF THE OggVorbis 'TREMOR' CODEC SOURCE CODE.   *
 *                                                                  *
 * THE OggVorbis 'TREMOR' SOURCE CODE IS (C) COPYRIGHT 1994-2002    *
 * BY THE Xiph.Org FOUNDATION http://www.xiph.org/                  *
 * ALL REDISTRIBUTION RIGHTS RESERVED.                              *
 *                                                                  *
 ********************************************************************

 function: arm7 and later wide math functions

 ********************************************************************/

#ifdef _ARM_ASSEM_
#ifndef _V_WIDE_MATH
#define _V_WIDE_MATH

static inline ogg_int32_t MULT32(ogg_int32_t x, ogg_int32_t y) {
  int lo,hi;
  asm volatile("smull  %0,%1,%2,%3;\n"
               : "=&r"(lo),"=&r"(hi)
               : "%r"(x),"r"(y));
  return(hi);
}

static inline ogg_int32_t MULT31(ogg_int32_t x, ogg_int32_t y) {
  return MULT32(x,y)<<1;
}

static inline ogg_int32_t MULT30(ogg_int32_t x, ogg_int32_t y) {
  return MULT32(x,y)<<2;
}

static inline ogg_int32_t MULT31_SHIFT15(ogg_int32_t x, ogg_int32_t y) {
  int lo,hi;
  asm volatile("smull  %0,%1,%2,%3;\n"
	       "mov    %0,%0, lsr #15;\n"
	       "orr    %1,%0,%1, lsl #17;\n"
               : "=&r"(lo),"=&r"(hi)
               : "%r"(x),"r"(y));
  return(hi);
}

static inline ogg_int32_t CLIP_TO_15(ogg_int32_t x) {
  asm volatile("subs   r0,%0,#32768;\n"
	       "movpl  %0,#0x7f00;\n"
	       "orrpl  %0,%0,#0xff;\n"
	       "adds   r0,%0,#32768;\n"
	       "movmi  %0,#0x8000;\n"
	       : "+r"(x)
	       :
	       : "r0","cc");
  return(x);
}

#endif

#ifndef _V_LSP_MATH_ASM
#define _V_LSP_MATH_ASM

static inline void lsp_loop_asm(ogg_uint32_t *qip,ogg_uint32_t *pip,
				ogg_int32_t *qexpp,
				ogg_int32_t *ilsp,ogg_int32_t wi,
				ogg_int32_t m){
  
  ogg_uint32_t qi=*qip,pi=*pip;
  ogg_int32_t qexp=*qexpp;

  asm("mov     r0,%3;"
      "mov     r1,%5,asr#1;"
      "add     r0,r0,r1,lsl#3;"
      "1:"
      
      "ldmdb   r0!,{r1,r3};"
      "subs    r1,r1,%4;"          //ilsp[j]-wi
      "rsbmi   r1,r1,#0;"          //labs(ilsp[j]-wi)
      "umull   %0,r2,r1,%0;"       //qi*=labs(ilsp[j]-wi)
      
      "subs    r1,r3,%4;"          //ilsp[j+1]-wi
      "rsbmi   r1,r1,#0;"          //labs(ilsp[j+1]-wi)
      "umull   %1,r3,r1,%1;"       //pi*=labs(ilsp[j+1]-wi)
      
      "cmn     r2,r3;"             // shift down 16?
      "beq     0f;"
      "add     %2,%2,#16;"
      "mov     %0,%0,lsr #16;"
      "orr     %0,%0,r2,lsl #16;"
      "mov     %1,%1,lsr #16;"
      "orr     %1,%1,r3,lsl #16;"
      "0:"
      "cmp     r0,%3;\n"
      "bhi     1b;\n"
      
      // odd filter assymetry
      "ands    r0,%5,#1;\n"
      "beq     2f;\n"
      "add     r0,%3,%5,lsl#2;\n"
      
      "ldr     r1,[r0,#-4];\n"
      "mov     r0,#0x4000;\n"
      
      "subs    r1,r1,%4;\n"          //ilsp[j]-wi
      "rsbmi   r1,r1,#0;\n"          //labs(ilsp[j]-wi)
      "umull   %0,r2,r1,%0;\n"       //qi*=labs(ilsp[j]-wi)
      "umull   %1,r3,r0,%1;\n"       //pi*=labs(ilsp[j+1]-wi)
      
      "cmn     r2,r3;\n"             // shift down 16?
      "beq     2f;\n"
      "add     %2,%2,#16;\n"
      "mov     %0,%0,lsr #16;\n"
      "orr     %0,%0,r2,lsl #16;\n"
      "mov     %1,%1,lsr #16;\n"
      "orr     %1,%1,r3,lsl #16;\n"
      
      //qi=(pi>>shift)*labs(ilsp[j]-wi);
      //pi=(qi>>shift)*labs(ilsp[j+1]-wi);
      //qexp+=shift;
      
      //}
	 
      /* normalize to max 16 sig figs */
      "2:"
      "mov     r2,#0;"
      "orr     r1,%0,%1;"
      "tst     r1,#0xff000000;"
      "addne   r2,r2,#8;"
      "movne   r1,r1,lsr #8;"
      "tst     r1,#0x00f00000;"
      "addne   r2,r2,#4;"
      "movne   r1,r1,lsr #4;"
      "tst     r1,#0x000c0000;"
      "addne   r2,r2,#2;"
      "movne   r1,r1,lsr #2;"
      "tst     r1,#0x00020000;"
      "addne   r2,r2,#1;"
      "movne   r1,r1,lsr #1;"
      "tst     r1,#0x00010000;"
      "addne   r2,r2,#1;"
      "mov     %0,%0,lsr r2;"
      "mov     %1,%1,lsr r2;"
      "add     %2,%2,r2;"
      
      : "+r"(qi),"+r"(pi),"+r"(qexp)
      : "r"(ilsp),"r"(wi),"r"(m)
      : "r0","r1","r2","r3","cc");
  
  *qip=qi;
  *pip=pi;
  *qexpp=qexp;
}

static inline void lsp_norm_asm(ogg_uint32_t *qip,ogg_int32_t *qexpp){

  ogg_uint32_t qi=*qip;
  ogg_int32_t qexp=*qexpp;

  asm("tst     %0,#0x0000ff00;"
      "moveq   %0,%0,lsl #8;"
      "subeq   %1,%1,#8;"
      "tst     %0,#0x0000f000;"
      "moveq   %0,%0,lsl #4;"
      "subeq   %1,%1,#4;"
      "tst     %0,#0x0000c000;"
      "moveq   %0,%0,lsl #2;"
      "subeq   %1,%1,#2;"
      "tst     %0,#0x00008000;"
      "moveq   %0,%0,lsl #1;"
      "subeq   %1,%1,#1;"
      : "+r"(qi),"+r"(qexp)
      :
      : "cc");
  *qip=qi;
  *qexpp=qexp;
}

#endif
#endif