vp8/common/arm/armv6/recon_v6.asm

   1 ;
   2 ;  Copyright (c) 2010 The VP8 project authors. All Rights Reserved.
   3 ;
   4 ;  Use of this source code is governed by a BSD-style license and patent
   5 ;  grant that can be found in the LICENSE file in the root of the source
   6 ;  tree. All contributing project authors may be found in the AUTHORS
   7 ;  file in the root of the source tree.
   8 ;
   9
  10
  11     EXPORT  |vp8_recon_b_armv6|
  12     EXPORT  |vp8_recon2b_armv6|
  13     EXPORT  |vp8_recon4b_armv6|
  14
  15     AREA    |.text|, CODE, READONLY  ; name this block of code
  16 prd     RN  r0
  17 dif     RN  r1
  18 dst     RN  r2
  19 stride      RN  r3
  20
  21 ;void recon_b(unsigned char *pred_ptr, short *diff_ptr, unsigned char *dst_ptr, int stride)
  22 ; R0 char* pred_ptr
  23 ; R1 short * dif_ptr
  24 ; R2 char * dst_ptr
  25 ; R3 int stride
  26
  27 ; Description:
  28 ; Loop through the block adding the Pred and Diff together.  Clamp and then
  29 ; store back into the Dst.
  30
  31 ; Restrictions :
  32 ; all buffers are expected to be 4 byte aligned coming in and
  33 ; going out.
  34 ;-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
  35 ;
  36 ;
  37 ;
  38 ;-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
  39 |vp8_recon_b_armv6| PROC
  40     stmdb   sp!, {r4 - r9, lr}
  41
  42     ;0, 1, 2, 3
  43     ldr     r4, [prd], #16          ; 3 | 2 | 1 | 0
  44     ldr     r6, [dif, #0]           ;     1 |     0
  45     ldr     r7, [dif, #4]           ;     3 |     2
  46
  47     pkhbt   r8, r6, r7, lsl #16     ;     2 |     0
  48     pkhtb   r9, r7, r6, asr #16     ;     3 |     1
  49
  50     uxtab16 r8, r8, r4              ;     2 |     0  +  3 | 2 | 2 | 0
  51     uxtab16 r9, r9, r4, ror #8      ;     3 |     1  +  0 | 3 | 2 | 1
  52
  53     usat16  r8, #8, r8
  54     usat16  r9, #8, r9
  55     add     dif, dif, #32
  56     orr     r8, r8, r9, lsl #8
  57
  58     str     r8, [dst], stride
  59
  60     ;0, 1, 2, 3
  61     ldr     r4, [prd], #16          ; 3 | 2 | 1 | 0
  62 ;;  ldr     r6, [dif, #8]           ;     1 |     0
  63 ;;  ldr     r7, [dif, #12]          ;     3 |     2
  64     ldr     r6, [dif, #0]           ;     1 |     0
  65     ldr     r7, [dif, #4]           ;     3 |     2
  66
  67     pkhbt   r8, r6, r7, lsl #16     ;     2 |     0
  68     pkhtb   r9, r7, r6, asr #16     ;     3 |     1
  69
  70     uxtab16 r8, r8, r4              ;     2 |     0  +  3 | 2 | 2 | 0
  71     uxtab16 r9, r9, r4, ror #8      ;     3 |     1  +  0 | 3 | 2 | 1
  72
  73     usat16  r8, #8, r8
  74     usat16  r9, #8, r9
  75     add     dif, dif, #32
  76     orr     r8, r8, r9, lsl #8
  77
  78     str     r8, [dst], stride
  79
  80     ;0, 1, 2, 3
  81     ldr     r4, [prd], #16          ; 3 | 2 | 1 | 0
  82 ;;  ldr     r6, [dif, #16]          ;     1 |     0
  83 ;;  ldr     r7, [dif, #20]          ;     3 |     2
  84     ldr     r6, [dif, #0]           ;     1 |     0
  85     ldr     r7, [dif, #4]           ;     3 |     2
  86
  87     pkhbt   r8, r6, r7, lsl #16     ;     2 |     0
  88     pkhtb   r9, r7, r6, asr #16     ;     3 |     1
  89
  90     uxtab16 r8, r8, r4              ;     2 |     0  +  3 | 2 | 2 | 0
  91     uxtab16 r9, r9, r4, ror #8      ;     3 |     1  +  0 | 3 | 2 | 1
  92
  93     usat16  r8, #8, r8
  94     usat16  r9, #8, r9
  95     add     dif, dif, #32
  96     orr     r8, r8, r9, lsl #8
  97
  98     str     r8, [dst], stride
  99
 100     ;0, 1, 2, 3
 101     ldr     r4, [prd], #16          ; 3 | 2 | 1 | 0
 102 ;;  ldr     r6, [dif, #24]          ;     1 |     0
 103 ;;  ldr     r7, [dif, #28]          ;     3 |     2
 104     ldr     r6, [dif, #0]           ;     1 |     0
 105     ldr     r7, [dif, #4]           ;     3 |     2
 106
 107     pkhbt   r8, r6, r7, lsl #16     ;     2 |     0
 108     pkhtb   r9, r7, r6, asr #16     ;     3 |     1
 109
 110     uxtab16 r8, r8, r4              ;     2 |     0  +  3 | 2 | 2 | 0
 111     uxtab16 r9, r9, r4, ror #8      ;     3 |     1  +  0 | 3 | 2 | 1
 112
 113     usat16  r8, #8, r8
 114     usat16  r9, #8, r9
 115     orr     r8, r8, r9, lsl #8
 116
 117     str     r8, [dst], stride
 118
 119     ldmia   sp!, {r4 - r9, pc}
 120
 121     ENDP    ; |recon_b|
 122
 123 ;-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
 124 ;
 125 ;
 126 ;
 127 ;-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
 128 ; R0 char  *pred_ptr
 129 ; R1 short *dif_ptr
 130 ; R2 char  *dst_ptr
 131 ; R3 int stride
 132 |vp8_recon4b_armv6| PROC
 133     stmdb   sp!, {r4 - r9, lr}
 134
 135     mov     lr, #4
 136
 137 recon4b_loop
 138     ;0, 1, 2, 3
 139     ldr     r4, [prd], #4           ; 3 | 2 | 1 | 0
 140     ldr     r6, [dif, #0]           ;     1 |     0
 141     ldr     r7, [dif, #4]           ;     3 |     2
 142
 143     pkhbt   r8, r6, r7, lsl #16     ;     2 |     0
 144     pkhtb   r9, r7, r6, asr #16     ;     3 |     1
 145
 146     uxtab16 r8, r8, r4              ;     2 |     0  +  3 | 2 | 2 | 0
 147     uxtab16 r9, r9, r4, ror #8      ;     3 |     1  +  0 | 3 | 2 | 1
 148
 149     usat16  r8, #8, r8
 150     usat16  r9, #8, r9
 151     orr     r8, r8, r9, lsl #8
 152
 153     str     r8, [dst]
 154
 155     ;4, 5, 6, 7
 156     ldr     r4, [prd], #4
 157 ;;  ldr     r6, [dif, #32]
 158 ;;  ldr     r7, [dif, #36]
 159     ldr     r6, [dif, #8]
 160     ldr     r7, [dif, #12]
 161
 162     pkhbt   r8, r6, r7, lsl #16
 163     pkhtb   r9, r7, r6, asr #16
 164
 165     uxtab16 r8, r8, r4
 166     uxtab16 r9, r9, r4, ror #8
 167     usat16  r8, #8, r8
 168     usat16  r9, #8, r9
 169     orr     r8, r8, r9, lsl #8
 170
 171     str     r8, [dst, #4]
 172
 173     ;8, 9, 10, 11
 174     ldr     r4, [prd], #4
 175 ;;  ldr     r6, [dif, #64]
 176 ;;  ldr     r7, [dif, #68]
 177     ldr     r6, [dif, #16]
 178     ldr     r7, [dif, #20]
 179
 180     pkhbt   r8, r6, r7, lsl #16
 181     pkhtb   r9, r7, r6, asr #16
 182
 183     uxtab16 r8, r8, r4
 184     uxtab16 r9, r9, r4, ror #8
 185     usat16  r8, #8, r8
 186     usat16  r9, #8, r9
 187     orr     r8, r8, r9, lsl #8
 188
 189     str     r8, [dst, #8]
 190
 191     ;12, 13, 14, 15
 192     ldr     r4, [prd], #4
 193 ;;  ldr     r6, [dif, #96]
 194 ;;  ldr     r7, [dif, #100]
 195     ldr     r6, [dif, #24]
 196     ldr     r7, [dif, #28]
 197
 198     pkhbt   r8, r6, r7, lsl #16
 199     pkhtb   r9, r7, r6, asr #16
 200
 201     uxtab16 r8, r8, r4
 202     uxtab16 r9, r9, r4, ror #8
 203     usat16  r8, #8, r8
 204     usat16  r9, #8, r9
 205     orr     r8, r8, r9, lsl #8
 206
 207     str     r8, [dst, #12]
 208
 209     add     dst, dst, stride
 210 ;;  add     dif, dif, #8
 211     add     dif, dif, #32
 212
 213     subs    lr, lr, #1
 214     bne     recon4b_loop
 215
 216     ldmia   sp!, {r4 - r9, pc}
 217
 218     ENDP    ; |Recon4B|
 219
 220 ;-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
 221 ;
 222 ;
 223 ;
 224 ;-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
 225 ; R0 char  *pred_ptr
 226 ; R1 short *dif_ptr
 227 ; R2 char  *dst_ptr
 228 ; R3 int stride
 229 |vp8_recon2b_armv6| PROC
 230     stmdb   sp!, {r4 - r9, lr}
 231
 232     mov     lr, #4
 233
 234 recon2b_loop
 235     ;0, 1, 2, 3
 236     ldr     r4, [prd], #4
 237     ldr     r6, [dif, #0]
 238     ldr     r7, [dif, #4]
 239
 240     pkhbt   r8, r6, r7, lsl #16
 241     pkhtb   r9, r7, r6, asr #16
 242
 243     uxtab16 r8, r8, r4
 244     uxtab16 r9, r9, r4, ror #8
 245     usat16  r8, #8, r8
 246     usat16  r9, #8, r9
 247     orr     r8, r8, r9, lsl #8
 248
 249     str     r8, [dst]
 250
 251     ;4, 5, 6, 7
 252     ldr     r4, [prd], #4
 253 ;;  ldr     r6, [dif, #32]
 254 ;;  ldr     r7, [dif, #36]
 255     ldr     r6, [dif, #8]
 256     ldr     r7, [dif, #12]
 257
 258     pkhbt   r8, r6, r7, lsl #16
 259     pkhtb   r9, r7, r6, asr #16
 260
 261     uxtab16 r8, r8, r4
 262     uxtab16 r9, r9, r4, ror #8
 263     usat16  r8, #8, r8
 264     usat16  r9, #8, r9
 265     orr     r8, r8, r9, lsl #8
 266
 267     str     r8, [dst, #4]
 268
 269     add     dst, dst, stride
 270 ;;  add     dif, dif, #8
 271     add     dif, dif, #16
 272
 273     subs    lr, lr, #1
 274     bne     recon2b_loop
 275
 276     ldmia   sp!, {r4 - r9, pc}
 277
 278     ENDP    ; |Recon2B|
 279
 280     END