libavcodec/arm/mpegvideo_neon.S

   1 /*
   2  * Copyright (c) 2010 Mans Rullgard
   3  *
   4  * This file is part of FFmpeg.
   5  *
   6  * FFmpeg is free software; you can redistribute it and/or
   7  * modify it under the terms of the GNU Lesser General Public
   8  * License as published by the Free Software Foundation; either
   9  * version 2.1 of the License, or (at your option) any later version.
  10  *
  11  * FFmpeg is distributed in the hope that it will be useful,
  12  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  13  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  14  * Lesser General Public License for more details.
  15  *
  16  * You should have received a copy of the GNU Lesser General Public
  17  * License along with FFmpeg; if not, write to the Free Software
  18  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
  19  */
  20
  21 #include "asm.S"
  22
  23 #define Y_DC_SCALE               0xa54
  24 #define C_DC_SCALE               0xa58
  25 #define AC_PRED                  0xa80
  26 #define BLOCK_LAST_INDEX         0x2278
  27 #define INTER_SCANTAB_RASTER_END 0x2478
  28 #define H263_AIC                 0x2728
  29
  30 function ff_dct_unquantize_h263_inter_neon, export=1
  31         add             r0,  r0,  #0x2200
  32         add             r12, r0,  #BLOCK_LAST_INDEX-0x2200
  33         ldr             r12, [r12, r2, lsl #2]
  34         add             r0,  r0,  #INTER_SCANTAB_RASTER_END-0x2200
  35         ldrb            r12, [r0, r12]
  36         sub             r2,  r3,  #1
  37         lsl             r0,  r3,  #1
  38         orr             r2,  r2,  #1
  39         add             r3,  r12, #1
  40 endfunc
  41
  42 function ff_dct_unquantize_h263_neon, export=1
  43         vdup.16         q15, r0                 @ qmul
  44         vdup.16         q14, r2                 @ qadd
  45         vneg.s16        q13, q14
  46         cmp             r3,  #4
  47         mov             r0,  r1
  48         ble             2f
  49 1:
  50         vld1.16         {q0},     [r0,:128]!
  51         vclt.s16        q3,  q0,  #0
  52         vld1.16         {q8},     [r0,:128]!
  53         vceq.s16        q1,  q0,  #0
  54         vmul.s16        q2,  q0,  q15
  55         vclt.s16        q11, q8,  #0
  56         vmul.s16        q10, q8,  q15
  57         vbsl            q3,  q13, q14
  58         vbsl            q11, q13, q14
  59         vadd.s16        q2,  q2,  q3
  60         vceq.s16        q9,  q8,  #0
  61         vadd.s16        q10, q10, q11
  62         vbif            q0,  q2,  q1
  63         vbif            q8,  q10, q9
  64         subs            r3,  r3,  #16
  65         vst1.16         {q0},     [r1,:128]!
  66         vst1.16         {q8},     [r1,:128]!
  67         bxle            lr
  68         cmp             r3,  #8
  69         bgt             1b
  70 2:
  71         vld1.16         {d0},     [r0,:64]
  72         vclt.s16        d3,  d0,  #0
  73         vceq.s16        d1,  d0,  #0
  74         vmul.s16        d2,  d0,  d30
  75         vbsl            d3,  d26, d28
  76         vadd.s16        d2,  d2,  d3
  77         vbif            d0,  d2,  d1
  78         vst1.16         {d0},     [r1,:64]
  79         bx              lr
  80 endfunc
  81
  82 function ff_dct_unquantize_h263_intra_neon, export=1
  83         push            {r4-r6,lr}
  84         ldr             r6,  [r0, #AC_PRED]
  85         add             r5,  r0,  #0x2700
  86         cmp             r6,  #0
  87         movne           r12, #63
  88         bne             1f
  89         add             lr,  r0,  #0x2200
  90         add             r12, lr,  #BLOCK_LAST_INDEX-0x2200
  91         add             lr,  lr,  #INTER_SCANTAB_RASTER_END-0x2200
  92         ldr             r12, [r12, r2, lsl #2]
  93         ldrb            r12, [lr, r12]
  94 1:      ldr             r5,  [r5, #H263_AIC-0x2700]
  95         ldrsh           r4,  [r1]
  96         cmp             r5,  #0
  97         mov             r5,  r1
  98         movne           r2,  #0
  99         bne             2f
 100         cmp             r2,  #4
 101         addge           r0,  r0,  #4
 102         sub             r2,  r3,  #1
 103         ldr             r6,  [r0, #Y_DC_SCALE]
 104         orr             r2,  r2,  #1
 105         smulbb          r4,  r4,  r6
 106 2:      lsl             r0,  r3,  #1
 107         add             r3,  r12, #1
 108         bl              ff_dct_unquantize_h263_neon
 109         vmov.16         d0[0], r4
 110         vst1.16         {d0[0]},  [r5]
 111         pop             {r4-r6,pc}
 112 endfunc