git.sesse.net Git - ffmpeg/blob - libavcodec/ppc/h264_altivec.c

   1 /*
   2  * Copyright (c) 2004 Romain Dolbeau <romain@dolbeau.org>
   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 "libavcodec/dsputil.h"
  22 #include "libavcodec/h264data.h"
  23 #include "libavcodec/h264dsp.h"
  24
  25 #include "dsputil_altivec.h"
  26 #include "util_altivec.h"
  27 #include "types_altivec.h"
  28
  29 #define PUT_OP_U8_ALTIVEC(d, s, dst) d = s
  30 #define AVG_OP_U8_ALTIVEC(d, s, dst) d = vec_avg(dst, s)
  31
  32 #define OP_U8_ALTIVEC                          PUT_OP_U8_ALTIVEC
  33 #define PREFIX_h264_chroma_mc8_altivec         put_h264_chroma_mc8_altivec
  34 #define PREFIX_no_rnd_vc1_chroma_mc8_altivec   put_no_rnd_vc1_chroma_mc8_altivec
  35 #define PREFIX_h264_chroma_mc8_num             altivec_put_h264_chroma_mc8_num
  36 #define PREFIX_h264_qpel16_h_lowpass_altivec   put_h264_qpel16_h_lowpass_altivec
  37 #define PREFIX_h264_qpel16_h_lowpass_num       altivec_put_h264_qpel16_h_lowpass_num
  38 #define PREFIX_h264_qpel16_v_lowpass_altivec   put_h264_qpel16_v_lowpass_altivec
  39 #define PREFIX_h264_qpel16_v_lowpass_num       altivec_put_h264_qpel16_v_lowpass_num
  40 #define PREFIX_h264_qpel16_hv_lowpass_altivec  put_h264_qpel16_hv_lowpass_altivec
  41 #define PREFIX_h264_qpel16_hv_lowpass_num      altivec_put_h264_qpel16_hv_lowpass_num
  42 #include "h264_template_altivec.c"
  43 #undef OP_U8_ALTIVEC
  44 #undef PREFIX_h264_chroma_mc8_altivec
  45 #undef PREFIX_no_rnd_vc1_chroma_mc8_altivec
  46 #undef PREFIX_h264_chroma_mc8_num
  47 #undef PREFIX_h264_qpel16_h_lowpass_altivec
  48 #undef PREFIX_h264_qpel16_h_lowpass_num
  49 #undef PREFIX_h264_qpel16_v_lowpass_altivec
  50 #undef PREFIX_h264_qpel16_v_lowpass_num
  51 #undef PREFIX_h264_qpel16_hv_lowpass_altivec
  52 #undef PREFIX_h264_qpel16_hv_lowpass_num
  53
  54 #define OP_U8_ALTIVEC                          AVG_OP_U8_ALTIVEC
  55 #define PREFIX_h264_chroma_mc8_altivec         avg_h264_chroma_mc8_altivec
  56 #define PREFIX_no_rnd_vc1_chroma_mc8_altivec   avg_no_rnd_vc1_chroma_mc8_altivec
  57 #define PREFIX_h264_chroma_mc8_num             altivec_avg_h264_chroma_mc8_num
  58 #define PREFIX_h264_qpel16_h_lowpass_altivec   avg_h264_qpel16_h_lowpass_altivec
  59 #define PREFIX_h264_qpel16_h_lowpass_num       altivec_avg_h264_qpel16_h_lowpass_num
  60 #define PREFIX_h264_qpel16_v_lowpass_altivec   avg_h264_qpel16_v_lowpass_altivec
  61 #define PREFIX_h264_qpel16_v_lowpass_num       altivec_avg_h264_qpel16_v_lowpass_num
  62 #define PREFIX_h264_qpel16_hv_lowpass_altivec  avg_h264_qpel16_hv_lowpass_altivec
  63 #define PREFIX_h264_qpel16_hv_lowpass_num      altivec_avg_h264_qpel16_hv_lowpass_num
  64 #include "h264_template_altivec.c"
  65 #undef OP_U8_ALTIVEC
  66 #undef PREFIX_h264_chroma_mc8_altivec
  67 #undef PREFIX_no_rnd_vc1_chroma_mc8_altivec
  68 #undef PREFIX_h264_chroma_mc8_num
  69 #undef PREFIX_h264_qpel16_h_lowpass_altivec
  70 #undef PREFIX_h264_qpel16_h_lowpass_num
  71 #undef PREFIX_h264_qpel16_v_lowpass_altivec
  72 #undef PREFIX_h264_qpel16_v_lowpass_num
  73 #undef PREFIX_h264_qpel16_hv_lowpass_altivec
  74 #undef PREFIX_h264_qpel16_hv_lowpass_num
  75
  76 #define H264_MC(OPNAME, SIZE, CODETYPE) \
  77 static void OPNAME ## h264_qpel ## SIZE ## _mc00_ ## CODETYPE (uint8_t *dst, uint8_t *src, int stride){\
  78     OPNAME ## pixels ## SIZE ## _ ## CODETYPE(dst, src, stride, SIZE);\
  79 }\
  80 \
  81 static void OPNAME ## h264_qpel ## SIZE ## _mc10_ ## CODETYPE(uint8_t *dst, uint8_t *src, int stride){ \
  82     DECLARE_ALIGNED(16, uint8_t, half)[SIZE*SIZE];\
  83     put_h264_qpel ## SIZE ## _h_lowpass_ ## CODETYPE(half, src, SIZE, stride);\
  84     OPNAME ## pixels ## SIZE ## _l2_ ## CODETYPE(dst, src, half, stride, stride, SIZE);\
  85 }\
  86 \
  87 static void OPNAME ## h264_qpel ## SIZE ## _mc20_ ## CODETYPE(uint8_t *dst, uint8_t *src, int stride){\
  88     OPNAME ## h264_qpel ## SIZE ## _h_lowpass_ ## CODETYPE(dst, src, stride, stride);\
  89 }\
  90 \
  91 static void OPNAME ## h264_qpel ## SIZE ## _mc30_ ## CODETYPE(uint8_t *dst, uint8_t *src, int stride){\
  92     DECLARE_ALIGNED(16, uint8_t, half)[SIZE*SIZE];\
  93     put_h264_qpel ## SIZE ## _h_lowpass_ ## CODETYPE(half, src, SIZE, stride);\
  94     OPNAME ## pixels ## SIZE ## _l2_ ## CODETYPE(dst, src+1, half, stride, stride, SIZE);\
  95 }\
  96 \
  97 static void OPNAME ## h264_qpel ## SIZE ## _mc01_ ## CODETYPE(uint8_t *dst, uint8_t *src, int stride){\
  98     DECLARE_ALIGNED(16, uint8_t, half)[SIZE*SIZE];\
  99     put_h264_qpel ## SIZE ## _v_lowpass_ ## CODETYPE(half, src, SIZE, stride);\
 100     OPNAME ## pixels ## SIZE ## _l2_ ## CODETYPE(dst, src, half, stride, stride, SIZE);\
 101 }\
 102 \
 103 static void OPNAME ## h264_qpel ## SIZE ## _mc02_ ## CODETYPE(uint8_t *dst, uint8_t *src, int stride){\
 104     OPNAME ## h264_qpel ## SIZE ## _v_lowpass_ ## CODETYPE(dst, src, stride, stride);\
 105 }\
 106 \
 107 static void OPNAME ## h264_qpel ## SIZE ## _mc03_ ## CODETYPE(uint8_t *dst, uint8_t *src, int stride){\
 108     DECLARE_ALIGNED(16, uint8_t, half)[SIZE*SIZE];\
 109     put_h264_qpel ## SIZE ## _v_lowpass_ ## CODETYPE(half, src, SIZE, stride);\
 110     OPNAME ## pixels ## SIZE ## _l2_ ## CODETYPE(dst, src+stride, half, stride, stride, SIZE);\
 111 }\
 112 \
 113 static void OPNAME ## h264_qpel ## SIZE ## _mc11_ ## CODETYPE(uint8_t *dst, uint8_t *src, int stride){\
 114     DECLARE_ALIGNED(16, uint8_t, halfH)[SIZE*SIZE];\
 115     DECLARE_ALIGNED(16, uint8_t, halfV)[SIZE*SIZE];\
 116     put_h264_qpel ## SIZE ## _h_lowpass_ ## CODETYPE(halfH, src, SIZE, stride);\
 117     put_h264_qpel ## SIZE ## _v_lowpass_ ## CODETYPE(halfV, src, SIZE, stride);\
 118     OPNAME ## pixels ## SIZE ## _l2_ ## CODETYPE(dst, halfH, halfV, stride, SIZE, SIZE);\
 119 }\
 120 \
 121 static void OPNAME ## h264_qpel ## SIZE ## _mc31_ ## CODETYPE(uint8_t *dst, uint8_t *src, int stride){\
 122     DECLARE_ALIGNED(16, uint8_t, halfH)[SIZE*SIZE];\
 123     DECLARE_ALIGNED(16, uint8_t, halfV)[SIZE*SIZE];\
 124     put_h264_qpel ## SIZE ## _h_lowpass_ ## CODETYPE(halfH, src, SIZE, stride);\
 125     put_h264_qpel ## SIZE ## _v_lowpass_ ## CODETYPE(halfV, src+1, SIZE, stride);\
 126     OPNAME ## pixels ## SIZE ## _l2_ ## CODETYPE(dst, halfH, halfV, stride, SIZE, SIZE);\
 127 }\
 128 \
 129 static void OPNAME ## h264_qpel ## SIZE ## _mc13_ ## CODETYPE(uint8_t *dst, uint8_t *src, int stride){\
 130     DECLARE_ALIGNED(16, uint8_t, halfH)[SIZE*SIZE];\
 131     DECLARE_ALIGNED(16, uint8_t, halfV)[SIZE*SIZE];\
 132     put_h264_qpel ## SIZE ## _h_lowpass_ ## CODETYPE(halfH, src + stride, SIZE, stride);\
 133     put_h264_qpel ## SIZE ## _v_lowpass_ ## CODETYPE(halfV, src, SIZE, stride);\
 134     OPNAME ## pixels ## SIZE ## _l2_ ## CODETYPE(dst, halfH, halfV, stride, SIZE, SIZE);\
 135 }\
 136 \
 137 static void OPNAME ## h264_qpel ## SIZE ## _mc33_ ## CODETYPE(uint8_t *dst, uint8_t *src, int stride){\
 138     DECLARE_ALIGNED(16, uint8_t, halfH)[SIZE*SIZE];\
 139     DECLARE_ALIGNED(16, uint8_t, halfV)[SIZE*SIZE];\
 140     put_h264_qpel ## SIZE ## _h_lowpass_ ## CODETYPE(halfH, src + stride, SIZE, stride);\
 141     put_h264_qpel ## SIZE ## _v_lowpass_ ## CODETYPE(halfV, src+1, SIZE, stride);\
 142     OPNAME ## pixels ## SIZE ## _l2_ ## CODETYPE(dst, halfH, halfV, stride, SIZE, SIZE);\
 143 }\
 144 \
 145 static void OPNAME ## h264_qpel ## SIZE ## _mc22_ ## CODETYPE(uint8_t *dst, uint8_t *src, int stride){\
 146     DECLARE_ALIGNED(16, int16_t, tmp)[SIZE*(SIZE+8)];\
 147     OPNAME ## h264_qpel ## SIZE ## _hv_lowpass_ ## CODETYPE(dst, tmp, src, stride, SIZE, stride);\
 148 }\
 149 \
 150 static void OPNAME ## h264_qpel ## SIZE ## _mc21_ ## CODETYPE(uint8_t *dst, uint8_t *src, int stride){\
 151     DECLARE_ALIGNED(16, uint8_t, halfH)[SIZE*SIZE];\
 152     DECLARE_ALIGNED(16, uint8_t, halfHV)[SIZE*SIZE];\
 153     DECLARE_ALIGNED(16, int16_t, tmp)[SIZE*(SIZE+8)];\
 154     put_h264_qpel ## SIZE ## _h_lowpass_ ## CODETYPE(halfH, src, SIZE, stride);\
 155     put_h264_qpel ## SIZE ## _hv_lowpass_ ## CODETYPE(halfHV, tmp, src, SIZE, SIZE, stride);\
 156     OPNAME ## pixels ## SIZE ## _l2_ ## CODETYPE(dst, halfH, halfHV, stride, SIZE, SIZE);\
 157 }\
 158 \
 159 static void OPNAME ## h264_qpel ## SIZE ## _mc23_ ## CODETYPE(uint8_t *dst, uint8_t *src, int stride){\
 160     DECLARE_ALIGNED(16, uint8_t, halfH)[SIZE*SIZE];\
 161     DECLARE_ALIGNED(16, uint8_t, halfHV)[SIZE*SIZE];\
 162     DECLARE_ALIGNED(16, int16_t, tmp)[SIZE*(SIZE+8)];\
 163     put_h264_qpel ## SIZE ## _h_lowpass_ ## CODETYPE(halfH, src + stride, SIZE, stride);\
 164     put_h264_qpel ## SIZE ## _hv_lowpass_ ## CODETYPE(halfHV, tmp, src, SIZE, SIZE, stride);\
 165     OPNAME ## pixels ## SIZE ## _l2_ ## CODETYPE(dst, halfH, halfHV, stride, SIZE, SIZE);\
 166 }\
 167 \
 168 static void OPNAME ## h264_qpel ## SIZE ## _mc12_ ## CODETYPE(uint8_t *dst, uint8_t *src, int stride){\
 169     DECLARE_ALIGNED(16, uint8_t, halfV)[SIZE*SIZE];\
 170     DECLARE_ALIGNED(16, uint8_t, halfHV)[SIZE*SIZE];\
 171     DECLARE_ALIGNED(16, int16_t, tmp)[SIZE*(SIZE+8)];\
 172     put_h264_qpel ## SIZE ## _v_lowpass_ ## CODETYPE(halfV, src, SIZE, stride);\
 173     put_h264_qpel ## SIZE ## _hv_lowpass_ ## CODETYPE(halfHV, tmp, src, SIZE, SIZE, stride);\
 174     OPNAME ## pixels ## SIZE ## _l2_ ## CODETYPE(dst, halfV, halfHV, stride, SIZE, SIZE);\
 175 }\
 176 \
 177 static void OPNAME ## h264_qpel ## SIZE ## _mc32_ ## CODETYPE(uint8_t *dst, uint8_t *src, int stride){\
 178     DECLARE_ALIGNED(16, uint8_t, halfV)[SIZE*SIZE];\
 179     DECLARE_ALIGNED(16, uint8_t, halfHV)[SIZE*SIZE];\
 180     DECLARE_ALIGNED(16, int16_t, tmp)[SIZE*(SIZE+8)];\
 181     put_h264_qpel ## SIZE ## _v_lowpass_ ## CODETYPE(halfV, src+1, SIZE, stride);\
 182     put_h264_qpel ## SIZE ## _hv_lowpass_ ## CODETYPE(halfHV, tmp, src, SIZE, SIZE, stride);\
 183     OPNAME ## pixels ## SIZE ## _l2_ ## CODETYPE(dst, halfV, halfHV, stride, SIZE, SIZE);\
 184 }\
 185
 186 static inline void put_pixels16_l2_altivec( uint8_t * dst, const uint8_t * src1,
 187                                     const uint8_t * src2, int dst_stride,
 188                                     int src_stride1, int h)
 189 {
 190     int i;
 191     vec_u8 a, b, d, tmp1, tmp2, mask, mask_, edges, align;
 192
 193     mask_ = vec_lvsl(0, src2);
 194
 195     for (i = 0; i < h; i++) {
 196
 197         tmp1 = vec_ld(i * src_stride1, src1);
 198         mask = vec_lvsl(i * src_stride1, src1);
 199         tmp2 = vec_ld(i * src_stride1 + 15, src1);
 200
 201         a = vec_perm(tmp1, tmp2, mask);
 202
 203         tmp1 = vec_ld(i * 16, src2);
 204         tmp2 = vec_ld(i * 16 + 15, src2);
 205
 206         b = vec_perm(tmp1, tmp2, mask_);
 207
 208         tmp1 = vec_ld(0, dst);
 209         mask = vec_lvsl(0, dst);
 210         tmp2 = vec_ld(15, dst);
 211
 212         d = vec_avg(a, b);
 213
 214         edges = vec_perm(tmp2, tmp1, mask);
 215
 216         align = vec_lvsr(0, dst);
 217
 218         tmp2 = vec_perm(d, edges, align);
 219         tmp1 = vec_perm(edges, d, align);
 220
 221         vec_st(tmp2, 15, dst);
 222         vec_st(tmp1, 0 , dst);
 223
 224         dst += dst_stride;
 225     }
 226 }
 227
 228 static inline void avg_pixels16_l2_altivec( uint8_t * dst, const uint8_t * src1,
 229                                     const uint8_t * src2, int dst_stride,
 230                                     int src_stride1, int h)
 231 {
 232     int i;
 233     vec_u8 a, b, d, tmp1, tmp2, mask, mask_, edges, align;
 234
 235     mask_ = vec_lvsl(0, src2);
 236
 237     for (i = 0; i < h; i++) {
 238
 239         tmp1 = vec_ld(i * src_stride1, src1);
 240         mask = vec_lvsl(i * src_stride1, src1);
 241         tmp2 = vec_ld(i * src_stride1 + 15, src1);
 242
 243         a = vec_perm(tmp1, tmp2, mask);
 244
 245         tmp1 = vec_ld(i * 16, src2);
 246         tmp2 = vec_ld(i * 16 + 15, src2);
 247
 248         b = vec_perm(tmp1, tmp2, mask_);
 249
 250         tmp1 = vec_ld(0, dst);
 251         mask = vec_lvsl(0, dst);
 252         tmp2 = vec_ld(15, dst);
 253
 254         d = vec_avg(vec_perm(tmp1, tmp2, mask), vec_avg(a, b));
 255
 256         edges = vec_perm(tmp2, tmp1, mask);
 257
 258         align = vec_lvsr(0, dst);
 259
 260         tmp2 = vec_perm(d, edges, align);
 261         tmp1 = vec_perm(edges, d, align);
 262
 263         vec_st(tmp2, 15, dst);
 264         vec_st(tmp1, 0 , dst);
 265
 266         dst += dst_stride;
 267     }
 268 }
 269
 270 /* Implemented but could be faster
 271 #define put_pixels16_l2_altivec(d,s1,s2,ds,s1s,h) put_pixels16_l2(d,s1,s2,ds,s1s,16,h)
 272 #define avg_pixels16_l2_altivec(d,s1,s2,ds,s1s,h) avg_pixels16_l2(d,s1,s2,ds,s1s,16,h)
 273  */
 274
 275 H264_MC(put_, 16, altivec)
 276 H264_MC(avg_, 16, altivec)
 277
 278
 279 /****************************************************************************
 280  * IDCT transform:
 281  ****************************************************************************/
 282
 283 #define VEC_1D_DCT(vb0,vb1,vb2,vb3,va0,va1,va2,va3)               \
 284     /* 1st stage */                                               \
 285     vz0 = vec_add(vb0,vb2);       /* temp[0] = Y[0] + Y[2] */     \
 286     vz1 = vec_sub(vb0,vb2);       /* temp[1] = Y[0] - Y[2] */     \
 287     vz2 = vec_sra(vb1,vec_splat_u16(1));                          \
 288     vz2 = vec_sub(vz2,vb3);       /* temp[2] = Y[1].1/2 - Y[3] */ \
 289     vz3 = vec_sra(vb3,vec_splat_u16(1));                          \
 290     vz3 = vec_add(vb1,vz3);       /* temp[3] = Y[1] + Y[3].1/2 */ \
 291     /* 2nd stage: output */                                       \
 292     va0 = vec_add(vz0,vz3);       /* x[0] = temp[0] + temp[3] */  \
 293     va1 = vec_add(vz1,vz2);       /* x[1] = temp[1] + temp[2] */  \
 294     va2 = vec_sub(vz1,vz2);       /* x[2] = temp[1] - temp[2] */  \
 295     va3 = vec_sub(vz0,vz3)        /* x[3] = temp[0] - temp[3] */
 296
 297 #define VEC_TRANSPOSE_4(a0,a1,a2,a3,b0,b1,b2,b3) \
 298     b0 = vec_mergeh( a0, a0 ); \
 299     b1 = vec_mergeh( a1, a0 ); \
 300     b2 = vec_mergeh( a2, a0 ); \
 301     b3 = vec_mergeh( a3, a0 ); \
 302     a0 = vec_mergeh( b0, b2 ); \
 303     a1 = vec_mergel( b0, b2 ); \
 304     a2 = vec_mergeh( b1, b3 ); \
 305     a3 = vec_mergel( b1, b3 ); \
 306     b0 = vec_mergeh( a0, a2 ); \
 307     b1 = vec_mergel( a0, a2 ); \
 308     b2 = vec_mergeh( a1, a3 ); \
 309     b3 = vec_mergel( a1, a3 )
 310
 311 #define VEC_LOAD_U8_ADD_S16_STORE_U8(va)                      \
 312     vdst_orig = vec_ld(0, dst);                               \
 313     vdst = vec_perm(vdst_orig, zero_u8v, vdst_mask);          \
 314     vdst_ss = (vec_s16) vec_mergeh(zero_u8v, vdst);         \
 315     va = vec_add(va, vdst_ss);                                \
 316     va_u8 = vec_packsu(va, zero_s16v);                        \
 317     va_u32 = vec_splat((vec_u32)va_u8, 0);                  \
 318     vec_ste(va_u32, element, (uint32_t*)dst);
 319
 320 static void ff_h264_idct_add_altivec(uint8_t *dst, DCTELEM *block, int stride)
 321 {
 322     vec_s16 va0, va1, va2, va3;
 323     vec_s16 vz0, vz1, vz2, vz3;
 324     vec_s16 vtmp0, vtmp1, vtmp2, vtmp3;
 325     vec_u8 va_u8;
 326     vec_u32 va_u32;
 327     vec_s16 vdst_ss;
 328     const vec_u16 v6us = vec_splat_u16(6);
 329     vec_u8 vdst, vdst_orig;
 330     vec_u8 vdst_mask = vec_lvsl(0, dst);
 331     int element = ((unsigned long)dst & 0xf) >> 2;
 332     LOAD_ZERO;
 333
 334     block[0] += 32;  /* add 32 as a DC-level for rounding */
 335
 336     vtmp0 = vec_ld(0,block);
 337     vtmp1 = vec_sld(vtmp0, vtmp0, 8);
 338     vtmp2 = vec_ld(16,block);
 339     vtmp3 = vec_sld(vtmp2, vtmp2, 8);
 340
 341     VEC_1D_DCT(vtmp0,vtmp1,vtmp2,vtmp3,va0,va1,va2,va3);
 342     VEC_TRANSPOSE_4(va0,va1,va2,va3,vtmp0,vtmp1,vtmp2,vtmp3);
 343     VEC_1D_DCT(vtmp0,vtmp1,vtmp2,vtmp3,va0,va1,va2,va3);
 344
 345     va0 = vec_sra(va0,v6us);
 346     va1 = vec_sra(va1,v6us);
 347     va2 = vec_sra(va2,v6us);
 348     va3 = vec_sra(va3,v6us);
 349
 350     VEC_LOAD_U8_ADD_S16_STORE_U8(va0);
 351     dst += stride;
 352     VEC_LOAD_U8_ADD_S16_STORE_U8(va1);
 353     dst += stride;
 354     VEC_LOAD_U8_ADD_S16_STORE_U8(va2);
 355     dst += stride;
 356     VEC_LOAD_U8_ADD_S16_STORE_U8(va3);
 357 }
 358
 359 #define IDCT8_1D_ALTIVEC(s0, s1, s2, s3, s4, s5, s6, s7,  d0, d1, d2, d3, d4, d5, d6, d7) {\
 360     /*        a0  = SRC(0) + SRC(4); */ \
 361     vec_s16 a0v = vec_add(s0, s4);    \
 362     /*        a2  = SRC(0) - SRC(4); */ \
 363     vec_s16 a2v = vec_sub(s0, s4);    \
 364     /*        a4  =           (SRC(2)>>1) - SRC(6); */ \
 365     vec_s16 a4v = vec_sub(vec_sra(s2, onev), s6);    \
 366     /*        a6  =           (SRC(6)>>1) + SRC(2); */ \
 367     vec_s16 a6v = vec_add(vec_sra(s6, onev), s2);    \
 368     /*        b0  =         a0 + a6; */ \
 369     vec_s16 b0v = vec_add(a0v, a6v);  \
 370     /*        b2  =         a2 + a4; */ \
 371     vec_s16 b2v = vec_add(a2v, a4v);  \
 372     /*        b4  =         a2 - a4; */ \
 373     vec_s16 b4v = vec_sub(a2v, a4v);  \
 374     /*        b6  =         a0 - a6; */ \
 375     vec_s16 b6v = vec_sub(a0v, a6v);  \
 376     /* a1 =  SRC(5) - SRC(3) - SRC(7) - (SRC(7)>>1); */ \
 377     /*        a1 =             (SRC(5)-SRC(3)) -  (SRC(7)  +  (SRC(7)>>1)); */ \
 378     vec_s16 a1v = vec_sub( vec_sub(s5, s3), vec_add(s7, vec_sra(s7, onev)) ); \
 379     /* a3 =  SRC(7) + SRC(1) - SRC(3) - (SRC(3)>>1); */ \
 380     /*        a3 =             (SRC(7)+SRC(1)) -  (SRC(3)  +  (SRC(3)>>1)); */ \
 381     vec_s16 a3v = vec_sub( vec_add(s7, s1), vec_add(s3, vec_sra(s3, onev)) );\
 382     /* a5 =  SRC(7) - SRC(1) + SRC(5) + (SRC(5)>>1); */ \
 383     /*        a5 =             (SRC(7)-SRC(1)) +   SRC(5) +   (SRC(5)>>1); */ \
 384     vec_s16 a5v = vec_add( vec_sub(s7, s1), vec_add(s5, vec_sra(s5, onev)) );\
 385     /*        a7 =                SRC(5)+SRC(3) +  SRC(1) +   (SRC(1)>>1); */ \
 386     vec_s16 a7v = vec_add( vec_add(s5, s3), vec_add(s1, vec_sra(s1, onev)) );\
 387     /*        b1 =                  (a7>>2)  +  a1; */ \
 388     vec_s16 b1v = vec_add( vec_sra(a7v, twov), a1v); \
 389     /*        b3 =          a3 +        (a5>>2); */ \
 390     vec_s16 b3v = vec_add(a3v, vec_sra(a5v, twov)); \
 391     /*        b5 =                  (a3>>2)  -   a5; */ \
 392     vec_s16 b5v = vec_sub( vec_sra(a3v, twov), a5v); \
 393     /*        b7 =           a7 -        (a1>>2); */ \
 394     vec_s16 b7v = vec_sub( a7v, vec_sra(a1v, twov)); \
 395     /* DST(0,    b0 + b7); */ \
 396     d0 = vec_add(b0v, b7v); \
 397     /* DST(1,    b2 + b5); */ \
 398     d1 = vec_add(b2v, b5v); \
 399     /* DST(2,    b4 + b3); */ \
 400     d2 = vec_add(b4v, b3v); \
 401     /* DST(3,    b6 + b1); */ \
 402     d3 = vec_add(b6v, b1v); \
 403     /* DST(4,    b6 - b1); */ \
 404     d4 = vec_sub(b6v, b1v); \
 405     /* DST(5,    b4 - b3); */ \
 406     d5 = vec_sub(b4v, b3v); \
 407     /* DST(6,    b2 - b5); */ \
 408     d6 = vec_sub(b2v, b5v); \
 409     /* DST(7,    b0 - b7); */ \
 410     d7 = vec_sub(b0v, b7v); \
 411 }
 412
 413 #define ALTIVEC_STORE_SUM_CLIP(dest, idctv, perm_ldv, perm_stv, sel) { \
 414     /* unaligned load */                                       \
 415     vec_u8 hv = vec_ld( 0, dest );                           \
 416     vec_u8 lv = vec_ld( 7, dest );                           \
 417     vec_u8 dstv   = vec_perm( hv, lv, (vec_u8)perm_ldv );  \
 418     vec_s16 idct_sh6 = vec_sra(idctv, sixv);                 \
 419     vec_u16 dst16 = (vec_u16)vec_mergeh(zero_u8v, dstv);   \
 420     vec_s16 idstsum = vec_adds(idct_sh6, (vec_s16)dst16);  \
 421     vec_u8 idstsum8 = vec_packsu(zero_s16v, idstsum);        \
 422     vec_u8 edgehv;                                           \
 423     /* unaligned store */                                      \
 424     vec_u8 bodyv  = vec_perm( idstsum8, idstsum8, perm_stv );\
 425     vec_u8 edgelv = vec_perm( sel, zero_u8v, perm_stv );     \
 426     lv    = vec_sel( lv, bodyv, edgelv );                      \
 427     vec_st( lv, 7, dest );                                     \
 428     hv    = vec_ld( 0, dest );                                 \
 429     edgehv = vec_perm( zero_u8v, sel, perm_stv );              \
 430     hv    = vec_sel( hv, bodyv, edgehv );                      \
 431     vec_st( hv, 0, dest );                                     \
 432  }
 433
 434 static void ff_h264_idct8_add_altivec( uint8_t *dst, DCTELEM *dct, int stride ) {
 435     vec_s16 s0, s1, s2, s3, s4, s5, s6, s7;
 436     vec_s16 d0, d1, d2, d3, d4, d5, d6, d7;
 437     vec_s16 idct0, idct1, idct2, idct3, idct4, idct5, idct6, idct7;
 438
 439     vec_u8 perm_ldv = vec_lvsl(0, dst);
 440     vec_u8 perm_stv = vec_lvsr(8, dst);
 441
 442     const vec_u16 onev = vec_splat_u16(1);
 443     const vec_u16 twov = vec_splat_u16(2);
 444     const vec_u16 sixv = vec_splat_u16(6);
 445
 446     const vec_u8 sel = (vec_u8) {0,0,0,0,0,0,0,0,-1,-1,-1,-1,-1,-1,-1,-1};
 447     LOAD_ZERO;
 448
 449     dct[0] += 32; // rounding for the >>6 at the end
 450
 451     s0 = vec_ld(0x00, (int16_t*)dct);
 452     s1 = vec_ld(0x10, (int16_t*)dct);
 453     s2 = vec_ld(0x20, (int16_t*)dct);
 454     s3 = vec_ld(0x30, (int16_t*)dct);
 455     s4 = vec_ld(0x40, (int16_t*)dct);
 456     s5 = vec_ld(0x50, (int16_t*)dct);
 457     s6 = vec_ld(0x60, (int16_t*)dct);
 458     s7 = vec_ld(0x70, (int16_t*)dct);
 459
 460     IDCT8_1D_ALTIVEC(s0, s1, s2, s3, s4, s5, s6, s7,
 461                      d0, d1, d2, d3, d4, d5, d6, d7);
 462
 463     TRANSPOSE8( d0,  d1,  d2,  d3,  d4,  d5,  d6, d7 );
 464
 465     IDCT8_1D_ALTIVEC(d0,  d1,  d2,  d3,  d4,  d5,  d6, d7,
 466                      idct0, idct1, idct2, idct3, idct4, idct5, idct6, idct7);
 467
 468     ALTIVEC_STORE_SUM_CLIP(&dst[0*stride], idct0, perm_ldv, perm_stv, sel);
 469     ALTIVEC_STORE_SUM_CLIP(&dst[1*stride], idct1, perm_ldv, perm_stv, sel);
 470     ALTIVEC_STORE_SUM_CLIP(&dst[2*stride], idct2, perm_ldv, perm_stv, sel);
 471     ALTIVEC_STORE_SUM_CLIP(&dst[3*stride], idct3, perm_ldv, perm_stv, sel);
 472     ALTIVEC_STORE_SUM_CLIP(&dst[4*stride], idct4, perm_ldv, perm_stv, sel);
 473     ALTIVEC_STORE_SUM_CLIP(&dst[5*stride], idct5, perm_ldv, perm_stv, sel);
 474     ALTIVEC_STORE_SUM_CLIP(&dst[6*stride], idct6, perm_ldv, perm_stv, sel);
 475     ALTIVEC_STORE_SUM_CLIP(&dst[7*stride], idct7, perm_ldv, perm_stv, sel);
 476 }
 477
 478 static av_always_inline void h264_idct_dc_add_internal(uint8_t *dst, DCTELEM *block, int stride, int size)
 479 {
 480     vec_s16 dc16;
 481     vec_u8 dcplus, dcminus, v0, v1, v2, v3, aligner;
 482     LOAD_ZERO;
 483     DECLARE_ALIGNED(16, int, dc);
 484     int i;
 485
 486     dc = (block[0] + 32) >> 6;
 487     dc16 = vec_splat((vec_s16) vec_lde(0, &dc), 1);
 488
 489     if (size == 4)
 490         dc16 = vec_sld(dc16, zero_s16v, 8);
 491     dcplus = vec_packsu(dc16, zero_s16v);
 492     dcminus = vec_packsu(vec_sub(zero_s16v, dc16), zero_s16v);
 493
 494     aligner = vec_lvsr(0, dst);
 495     dcplus = vec_perm(dcplus, dcplus, aligner);
 496     dcminus = vec_perm(dcminus, dcminus, aligner);
 497
 498     for (i = 0; i < size; i += 4) {
 499         v0 = vec_ld(0, dst+0*stride);
 500         v1 = vec_ld(0, dst+1*stride);
 501         v2 = vec_ld(0, dst+2*stride);
 502         v3 = vec_ld(0, dst+3*stride);
 503
 504         v0 = vec_adds(v0, dcplus);
 505         v1 = vec_adds(v1, dcplus);
 506         v2 = vec_adds(v2, dcplus);
 507         v3 = vec_adds(v3, dcplus);
 508
 509         v0 = vec_subs(v0, dcminus);
 510         v1 = vec_subs(v1, dcminus);
 511         v2 = vec_subs(v2, dcminus);
 512         v3 = vec_subs(v3, dcminus);
 513
 514         vec_st(v0, 0, dst+0*stride);
 515         vec_st(v1, 0, dst+1*stride);
 516         vec_st(v2, 0, dst+2*stride);
 517         vec_st(v3, 0, dst+3*stride);
 518
 519         dst += 4*stride;
 520     }
 521 }
 522
 523 static void h264_idct_dc_add_altivec(uint8_t *dst, DCTELEM *block, int stride)
 524 {
 525     h264_idct_dc_add_internal(dst, block, stride, 4);
 526 }
 527
 528 static void ff_h264_idct8_dc_add_altivec(uint8_t *dst, DCTELEM *block, int stride)
 529 {
 530     h264_idct_dc_add_internal(dst, block, stride, 8);
 531 }
 532
 533 static void ff_h264_idct_add16_altivec(uint8_t *dst, const int *block_offset, DCTELEM *block, int stride, const uint8_t nnzc[6*8]){
 534     int i;
 535     for(i=0; i<16; i++){
 536         int nnz = nnzc[ scan8[i] ];
 537         if(nnz){
 538             if(nnz==1 && block[i*16]) h264_idct_dc_add_altivec(dst + block_offset[i], block + i*16, stride);
 539             else                      ff_h264_idct_add_altivec(dst + block_offset[i], block + i*16, stride);
 540         }
 541     }
 542 }
 543
 544 static void ff_h264_idct_add16intra_altivec(uint8_t *dst, const int *block_offset, DCTELEM *block, int stride, const uint8_t nnzc[6*8]){
 545     int i;
 546     for(i=0; i<16; i++){
 547         if(nnzc[ scan8[i] ]) ff_h264_idct_add_altivec(dst + block_offset[i], block + i*16, stride);
 548         else if(block[i*16]) h264_idct_dc_add_altivec(dst + block_offset[i], block + i*16, stride);
 549     }
 550 }
 551
 552 static void ff_h264_idct8_add4_altivec(uint8_t *dst, const int *block_offset, DCTELEM *block, int stride, const uint8_t nnzc[6*8]){
 553     int i;
 554     for(i=0; i<16; i+=4){
 555         int nnz = nnzc[ scan8[i] ];
 556         if(nnz){
 557             if(nnz==1 && block[i*16]) ff_h264_idct8_dc_add_altivec(dst + block_offset[i], block + i*16, stride);
 558             else                      ff_h264_idct8_add_altivec   (dst + block_offset[i], block + i*16, stride);
 559         }
 560     }
 561 }
 562
 563 static void ff_h264_idct_add8_altivec(uint8_t **dest, const int *block_offset, DCTELEM *block, int stride, const uint8_t nnzc[6*8]){
 564     int i;
 565     for(i=16; i<16+8; i++){
 566         if(nnzc[ scan8[i] ])
 567             ff_h264_idct_add_altivec(dest[(i&4)>>2] + block_offset[i], block + i*16, stride);
 568         else if(block[i*16])
 569             h264_idct_dc_add_altivec(dest[(i&4)>>2] + block_offset[i], block + i*16, stride);
 570     }
 571 }
 572
 573 #define transpose4x16(r0, r1, r2, r3) {      \
 574     register vec_u8 r4;                    \
 575     register vec_u8 r5;                    \
 576     register vec_u8 r6;                    \
 577     register vec_u8 r7;                    \
 578                                              \
 579     r4 = vec_mergeh(r0, r2);  /*0, 2 set 0*/ \
 580     r5 = vec_mergel(r0, r2);  /*0, 2 set 1*/ \
 581     r6 = vec_mergeh(r1, r3);  /*1, 3 set 0*/ \
 582     r7 = vec_mergel(r1, r3);  /*1, 3 set 1*/ \
 583                                              \
 584     r0 = vec_mergeh(r4, r6);  /*all set 0*/  \
 585     r1 = vec_mergel(r4, r6);  /*all set 1*/  \
 586     r2 = vec_mergeh(r5, r7);  /*all set 2*/  \
 587     r3 = vec_mergel(r5, r7);  /*all set 3*/  \
 588 }
 589
 590 static inline void write16x4(uint8_t *dst, int dst_stride,
 591                              register vec_u8 r0, register vec_u8 r1,
 592                              register vec_u8 r2, register vec_u8 r3) {
 593     DECLARE_ALIGNED(16, unsigned char, result)[64];
 594     uint32_t *src_int = (uint32_t *)result, *dst_int = (uint32_t *)dst;
 595     int int_dst_stride = dst_stride/4;
 596
 597     vec_st(r0, 0, result);
 598     vec_st(r1, 16, result);
 599     vec_st(r2, 32, result);
 600     vec_st(r3, 48, result);
 601     /* FIXME: there has to be a better way!!!! */
 602     *dst_int = *src_int;
 603     *(dst_int+   int_dst_stride) = *(src_int + 1);
 604     *(dst_int+ 2*int_dst_stride) = *(src_int + 2);
 605     *(dst_int+ 3*int_dst_stride) = *(src_int + 3);
 606     *(dst_int+ 4*int_dst_stride) = *(src_int + 4);
 607     *(dst_int+ 5*int_dst_stride) = *(src_int + 5);
 608     *(dst_int+ 6*int_dst_stride) = *(src_int + 6);
 609     *(dst_int+ 7*int_dst_stride) = *(src_int + 7);
 610     *(dst_int+ 8*int_dst_stride) = *(src_int + 8);
 611     *(dst_int+ 9*int_dst_stride) = *(src_int + 9);
 612     *(dst_int+10*int_dst_stride) = *(src_int + 10);
 613     *(dst_int+11*int_dst_stride) = *(src_int + 11);
 614     *(dst_int+12*int_dst_stride) = *(src_int + 12);
 615     *(dst_int+13*int_dst_stride) = *(src_int + 13);
 616     *(dst_int+14*int_dst_stride) = *(src_int + 14);
 617     *(dst_int+15*int_dst_stride) = *(src_int + 15);
 618 }
 619
 620 /** \brief performs a 6x16 transpose of data in src, and stores it to dst
 621     \todo FIXME: see if we can't spare some vec_lvsl() by them factorizing
 622     out of unaligned_load() */
 623 #define readAndTranspose16x6(src, src_stride, r8, r9, r10, r11, r12, r13) {\
 624     register vec_u8 r0  = unaligned_load(0,             src);            \
 625     register vec_u8 r1  = unaligned_load(   src_stride, src);            \
 626     register vec_u8 r2  = unaligned_load(2* src_stride, src);            \
 627     register vec_u8 r3  = unaligned_load(3* src_stride, src);            \
 628     register vec_u8 r4  = unaligned_load(4* src_stride, src);            \
 629     register vec_u8 r5  = unaligned_load(5* src_stride, src);            \
 630     register vec_u8 r6  = unaligned_load(6* src_stride, src);            \
 631     register vec_u8 r7  = unaligned_load(7* src_stride, src);            \
 632     register vec_u8 r14 = unaligned_load(14*src_stride, src);            \
 633     register vec_u8 r15 = unaligned_load(15*src_stride, src);            \
 634                                                                            \
 635     r8  = unaligned_load( 8*src_stride, src);                              \
 636     r9  = unaligned_load( 9*src_stride, src);                              \
 637     r10 = unaligned_load(10*src_stride, src);                              \
 638     r11 = unaligned_load(11*src_stride, src);                              \
 639     r12 = unaligned_load(12*src_stride, src);                              \
 640     r13 = unaligned_load(13*src_stride, src);                              \
 641                                                                            \
 642     /*Merge first pairs*/                                                  \
 643     r0 = vec_mergeh(r0, r8);    /*0, 8*/                                   \
 644     r1 = vec_mergeh(r1, r9);    /*1, 9*/                                   \
 645     r2 = vec_mergeh(r2, r10);   /*2,10*/                                   \
 646     r3 = vec_mergeh(r3, r11);   /*3,11*/                                   \
 647     r4 = vec_mergeh(r4, r12);   /*4,12*/                                   \
 648     r5 = vec_mergeh(r5, r13);   /*5,13*/                                   \
 649     r6 = vec_mergeh(r6, r14);   /*6,14*/                                   \
 650     r7 = vec_mergeh(r7, r15);   /*7,15*/                                   \
 651                                                                            \
 652     /*Merge second pairs*/                                                 \
 653     r8  = vec_mergeh(r0, r4);   /*0,4, 8,12 set 0*/                        \
 654     r9  = vec_mergel(r0, r4);   /*0,4, 8,12 set 1*/                        \
 655     r10 = vec_mergeh(r1, r5);   /*1,5, 9,13 set 0*/                        \
 656     r11 = vec_mergel(r1, r5);   /*1,5, 9,13 set 1*/                        \
 657     r12 = vec_mergeh(r2, r6);   /*2,6,10,14 set 0*/                        \
 658     r13 = vec_mergel(r2, r6);   /*2,6,10,14 set 1*/                        \
 659     r14 = vec_mergeh(r3, r7);   /*3,7,11,15 set 0*/                        \
 660     r15 = vec_mergel(r3, r7);   /*3,7,11,15 set 1*/                        \
 661                                                                            \
 662     /*Third merge*/                                                        \
 663     r0 = vec_mergeh(r8,  r12);  /*0,2,4,6,8,10,12,14 set 0*/               \
 664     r1 = vec_mergel(r8,  r12);  /*0,2,4,6,8,10,12,14 set 1*/               \
 665     r2 = vec_mergeh(r9,  r13);  /*0,2,4,6,8,10,12,14 set 2*/               \
 666     r4 = vec_mergeh(r10, r14);  /*1,3,5,7,9,11,13,15 set 0*/               \
 667     r5 = vec_mergel(r10, r14);  /*1,3,5,7,9,11,13,15 set 1*/               \
 668     r6 = vec_mergeh(r11, r15);  /*1,3,5,7,9,11,13,15 set 2*/               \
 669     /* Don't need to compute 3 and 7*/                                     \
 670                                                                            \
 671     /*Final merge*/                                                        \
 672     r8  = vec_mergeh(r0, r4);   /*all set 0*/                              \
 673     r9  = vec_mergel(r0, r4);   /*all set 1*/                              \
 674     r10 = vec_mergeh(r1, r5);   /*all set 2*/                              \
 675     r11 = vec_mergel(r1, r5);   /*all set 3*/                              \
 676     r12 = vec_mergeh(r2, r6);   /*all set 4*/                              \
 677     r13 = vec_mergel(r2, r6);   /*all set 5*/                              \
 678     /* Don't need to compute 14 and 15*/                                   \
 679                                                                            \
 680 }
 681
 682 // out: o = |x-y| < a
 683 static inline vec_u8 diff_lt_altivec ( register vec_u8 x,
 684                                          register vec_u8 y,
 685                                          register vec_u8 a) {
 686
 687     register vec_u8 diff = vec_subs(x, y);
 688     register vec_u8 diffneg = vec_subs(y, x);
 689     register vec_u8 o = vec_or(diff, diffneg); /* |x-y| */
 690     o = (vec_u8)vec_cmplt(o, a);
 691     return o;
 692 }
 693
 694 static inline vec_u8 h264_deblock_mask ( register vec_u8 p0,
 695                                            register vec_u8 p1,
 696                                            register vec_u8 q0,
 697                                            register vec_u8 q1,
 698                                            register vec_u8 alpha,
 699                                            register vec_u8 beta) {
 700
 701     register vec_u8 mask;
 702     register vec_u8 tempmask;
 703
 704     mask = diff_lt_altivec(p0, q0, alpha);
 705     tempmask = diff_lt_altivec(p1, p0, beta);
 706     mask = vec_and(mask, tempmask);
 707     tempmask = diff_lt_altivec(q1, q0, beta);
 708     mask = vec_and(mask, tempmask);
 709
 710     return mask;
 711 }
 712
 713 // out: newp1 = clip((p2 + ((p0 + q0 + 1) >> 1)) >> 1, p1-tc0, p1+tc0)
 714 static inline vec_u8 h264_deblock_q1(register vec_u8 p0,
 715                                        register vec_u8 p1,
 716                                        register vec_u8 p2,
 717                                        register vec_u8 q0,
 718                                        register vec_u8 tc0) {
 719
 720     register vec_u8 average = vec_avg(p0, q0);
 721     register vec_u8 temp;
 722     register vec_u8 uncliped;
 723     register vec_u8 ones;
 724     register vec_u8 max;
 725     register vec_u8 min;
 726     register vec_u8 newp1;
 727
 728     temp = vec_xor(average, p2);
 729     average = vec_avg(average, p2);     /*avg(p2, avg(p0, q0)) */
 730     ones = vec_splat_u8(1);
 731     temp = vec_and(temp, ones);         /*(p2^avg(p0, q0)) & 1 */
 732     uncliped = vec_subs(average, temp); /*(p2+((p0+q0+1)>>1))>>1 */
 733     max = vec_adds(p1, tc0);
 734     min = vec_subs(p1, tc0);
 735     newp1 = vec_max(min, uncliped);
 736     newp1 = vec_min(max, newp1);
 737     return newp1;
 738 }
 739
 740 #define h264_deblock_p0_q0(p0, p1, q0, q1, tc0masked) {                                           \
 741                                                                                                   \
 742     const vec_u8 A0v = vec_sl(vec_splat_u8(10), vec_splat_u8(4));                               \
 743                                                                                                   \
 744     register vec_u8 pq0bit = vec_xor(p0,q0);                                                    \
 745     register vec_u8 q1minus;                                                                    \
 746     register vec_u8 p0minus;                                                                    \
 747     register vec_u8 stage1;                                                                     \
 748     register vec_u8 stage2;                                                                     \
 749     register vec_u8 vec160;                                                                     \
 750     register vec_u8 delta;                                                                      \
 751     register vec_u8 deltaneg;                                                                   \
 752                                                                                                   \
 753     q1minus = vec_nor(q1, q1);                 /* 255 - q1 */                                     \
 754     stage1 = vec_avg(p1, q1minus);             /* (p1 - q1 + 256)>>1 */                           \
 755     stage2 = vec_sr(stage1, vec_splat_u8(1));  /* (p1 - q1 + 256)>>2 = 64 + (p1 - q1) >> 2 */     \
 756     p0minus = vec_nor(p0, p0);                 /* 255 - p0 */                                     \
 757     stage1 = vec_avg(q0, p0minus);             /* (q0 - p0 + 256)>>1 */                           \
 758     pq0bit = vec_and(pq0bit, vec_splat_u8(1));                                                    \
 759     stage2 = vec_avg(stage2, pq0bit);          /* 32 + ((q0 - p0)&1 + (p1 - q1) >> 2 + 1) >> 1 */ \
 760     stage2 = vec_adds(stage2, stage1);         /* 160 + ((p0 - q0) + (p1 - q1) >> 2 + 1) >> 1 */  \
 761     vec160 = vec_ld(0, &A0v);                                                                     \
 762     deltaneg = vec_subs(vec160, stage2);       /* -d */                                           \
 763     delta = vec_subs(stage2, vec160);          /* d */                                            \
 764     deltaneg = vec_min(tc0masked, deltaneg);                                                      \
 765     delta = vec_min(tc0masked, delta);                                                            \
 766     p0 = vec_subs(p0, deltaneg);                                                                  \
 767     q0 = vec_subs(q0, delta);                                                                     \
 768     p0 = vec_adds(p0, delta);                                                                     \
 769     q0 = vec_adds(q0, deltaneg);                                                                  \
 770 }
 771
 772 #define h264_loop_filter_luma_altivec(p2, p1, p0, q0, q1, q2, alpha, beta, tc0) {            \
 773     DECLARE_ALIGNED(16, unsigned char, temp)[16];                                             \
 774     register vec_u8 alphavec;                                                              \
 775     register vec_u8 betavec;                                                               \
 776     register vec_u8 mask;                                                                  \
 777     register vec_u8 p1mask;                                                                \
 778     register vec_u8 q1mask;                                                                \
 779     register vector signed   char tc0vec;                                                    \
 780     register vec_u8 finaltc0;                                                              \
 781     register vec_u8 tc0masked;                                                             \
 782     register vec_u8 newp1;                                                                 \
 783     register vec_u8 newq1;                                                                 \
 784                                                                                              \
 785     temp[0] = alpha;                                                                         \
 786     temp[1] = beta;                                                                          \
 787     alphavec = vec_ld(0, temp);                                                              \
 788     betavec = vec_splat(alphavec, 0x1);                                                      \
 789     alphavec = vec_splat(alphavec, 0x0);                                                     \
 790     mask = h264_deblock_mask(p0, p1, q0, q1, alphavec, betavec); /*if in block */            \
 791                                                                                              \
 792     *((int *)temp) = *((int *)tc0);                                                          \
 793     tc0vec = vec_ld(0, (signed char*)temp);                                                  \
 794     tc0vec = vec_mergeh(tc0vec, tc0vec);                                                     \
 795     tc0vec = vec_mergeh(tc0vec, tc0vec);                                                     \
 796     mask = vec_and(mask, vec_cmpgt(tc0vec, vec_splat_s8(-1)));  /* if tc0[i] >= 0 */         \
 797     finaltc0 = vec_and((vec_u8)tc0vec, mask);     /* tc = tc0 */                           \
 798                                                                                              \
 799     p1mask = diff_lt_altivec(p2, p0, betavec);                                               \
 800     p1mask = vec_and(p1mask, mask);                             /* if ( |p2 - p0| < beta) */ \
 801     tc0masked = vec_and(p1mask, (vec_u8)tc0vec);                                           \
 802     finaltc0 = vec_sub(finaltc0, p1mask);                       /* tc++ */                   \
 803     newp1 = h264_deblock_q1(p0, p1, p2, q0, tc0masked);                                      \
 804     /*end if*/                                                                               \
 805                                                                                              \
 806     q1mask = diff_lt_altivec(q2, q0, betavec);                                               \
 807     q1mask = vec_and(q1mask, mask);                             /* if ( |q2 - q0| < beta ) */\
 808     tc0masked = vec_and(q1mask, (vec_u8)tc0vec);                                           \
 809     finaltc0 = vec_sub(finaltc0, q1mask);                       /* tc++ */                   \
 810     newq1 = h264_deblock_q1(p0, q1, q2, q0, tc0masked);                                      \
 811     /*end if*/                                                                               \
 812                                                                                              \
 813     h264_deblock_p0_q0(p0, p1, q0, q1, finaltc0);                                            \
 814     p1 = newp1;                                                                              \
 815     q1 = newq1;                                                                              \
 816 }
 817
 818 static void h264_v_loop_filter_luma_altivec(uint8_t *pix, int stride, int alpha, int beta, int8_t *tc0) {
 819
 820     if ((tc0[0] & tc0[1] & tc0[2] & tc0[3]) >= 0) {
 821         register vec_u8 p2 = vec_ld(-3*stride, pix);
 822         register vec_u8 p1 = vec_ld(-2*stride, pix);
 823         register vec_u8 p0 = vec_ld(-1*stride, pix);
 824         register vec_u8 q0 = vec_ld(0, pix);
 825         register vec_u8 q1 = vec_ld(stride, pix);
 826         register vec_u8 q2 = vec_ld(2*stride, pix);
 827         h264_loop_filter_luma_altivec(p2, p1, p0, q0, q1, q2, alpha, beta, tc0);
 828         vec_st(p1, -2*stride, pix);
 829         vec_st(p0, -1*stride, pix);
 830         vec_st(q0, 0, pix);
 831         vec_st(q1, stride, pix);
 832     }
 833 }
 834
 835 static void h264_h_loop_filter_luma_altivec(uint8_t *pix, int stride, int alpha, int beta, int8_t *tc0) {
 836
 837     register vec_u8 line0, line1, line2, line3, line4, line5;
 838     if ((tc0[0] & tc0[1] & tc0[2] & tc0[3]) < 0)
 839         return;
 840     readAndTranspose16x6(pix-3, stride, line0, line1, line2, line3, line4, line5);
 841     h264_loop_filter_luma_altivec(line0, line1, line2, line3, line4, line5, alpha, beta, tc0);
 842     transpose4x16(line1, line2, line3, line4);
 843     write16x4(pix-2, stride, line1, line2, line3, line4);
 844 }
 845
 846 static av_always_inline
 847 void weight_h264_WxH_altivec(uint8_t *block, int stride, int log2_denom, int weight, int offset, int w, int h)
 848 {
 849     int y, aligned;
 850     vec_u8 vblock;
 851     vec_s16 vtemp, vweight, voffset, v0, v1;
 852     vec_u16 vlog2_denom;
 853     DECLARE_ALIGNED(16, int32_t, temp)[4];
 854     LOAD_ZERO;
 855
 856     offset <<= log2_denom;
 857     if(log2_denom) offset += 1<<(log2_denom-1);
 858     temp[0] = log2_denom;
 859     temp[1] = weight;
 860     temp[2] = offset;
 861
 862     vtemp = (vec_s16)vec_ld(0, temp);
 863     vlog2_denom = (vec_u16)vec_splat(vtemp, 1);
 864     vweight = vec_splat(vtemp, 3);
 865     voffset = vec_splat(vtemp, 5);
 866     aligned = !((unsigned long)block & 0xf);
 867
 868     for (y=0; y<h; y++) {
 869         vblock = vec_ld(0, block);
 870
 871         v0 = (vec_s16)vec_mergeh(zero_u8v, vblock);
 872         v1 = (vec_s16)vec_mergel(zero_u8v, vblock);
 873
 874         if (w == 16 || aligned) {
 875             v0 = vec_mladd(v0, vweight, zero_s16v);
 876             v0 = vec_adds(v0, voffset);
 877             v0 = vec_sra(v0, vlog2_denom);
 878         }
 879         if (w == 16 || !aligned) {
 880             v1 = vec_mladd(v1, vweight, zero_s16v);
 881             v1 = vec_adds(v1, voffset);
 882             v1 = vec_sra(v1, vlog2_denom);
 883         }
 884         vblock = vec_packsu(v0, v1);
 885         vec_st(vblock, 0, block);
 886
 887         block += stride;
 888     }
 889 }
 890
 891 static av_always_inline
 892 void biweight_h264_WxH_altivec(uint8_t *dst, uint8_t *src, int stride, int log2_denom,
 893                                int weightd, int weights, int offset, int w, int h)
 894 {
 895     int y, dst_aligned, src_aligned;
 896     vec_u8 vsrc, vdst;
 897     vec_s16 vtemp, vweights, vweightd, voffset, v0, v1, v2, v3;
 898     vec_u16 vlog2_denom;
 899     DECLARE_ALIGNED(16, int32_t, temp)[4];
 900     LOAD_ZERO;
 901
 902     offset = ((offset + 1) | 1) << log2_denom;
 903     temp[0] = log2_denom+1;
 904     temp[1] = weights;
 905     temp[2] = weightd;
 906     temp[3] = offset;
 907
 908     vtemp = (vec_s16)vec_ld(0, temp);
 909     vlog2_denom = (vec_u16)vec_splat(vtemp, 1);
 910     vweights = vec_splat(vtemp, 3);
 911     vweightd = vec_splat(vtemp, 5);
 912     voffset = vec_splat(vtemp, 7);
 913     dst_aligned = !((unsigned long)dst & 0xf);
 914     src_aligned = !((unsigned long)src & 0xf);
 915
 916     for (y=0; y<h; y++) {
 917         vdst = vec_ld(0, dst);
 918         vsrc = vec_ld(0, src);
 919
 920         v0 = (vec_s16)vec_mergeh(zero_u8v, vdst);
 921         v1 = (vec_s16)vec_mergel(zero_u8v, vdst);
 922         v2 = (vec_s16)vec_mergeh(zero_u8v, vsrc);
 923         v3 = (vec_s16)vec_mergel(zero_u8v, vsrc);
 924
 925         if (w == 8) {
 926             if (src_aligned)
 927                 v3 = v2;
 928             else
 929                 v2 = v3;
 930         }
 931
 932         if (w == 16 || dst_aligned) {
 933             v0 = vec_mladd(v0, vweightd, zero_s16v);
 934             v2 = vec_mladd(v2, vweights, zero_s16v);
 935
 936             v0 = vec_adds(v0, voffset);
 937             v0 = vec_adds(v0, v2);
 938             v0 = vec_sra(v0, vlog2_denom);
 939         }
 940         if (w == 16 || !dst_aligned) {
 941             v1 = vec_mladd(v1, vweightd, zero_s16v);
 942             v3 = vec_mladd(v3, vweights, zero_s16v);
 943
 944             v1 = vec_adds(v1, voffset);
 945             v1 = vec_adds(v1, v3);
 946             v1 = vec_sra(v1, vlog2_denom);
 947         }
 948         vdst = vec_packsu(v0, v1);
 949         vec_st(vdst, 0, dst);
 950
 951         dst += stride;
 952         src += stride;
 953     }
 954 }
 955
 956 #define H264_WEIGHT(W,H) \
 957 static void ff_weight_h264_pixels ## W ## x ## H ## _altivec(uint8_t *block, int stride, int log2_denom, int weight, int offset){ \
 958     weight_h264_WxH_altivec(block, stride, log2_denom, weight, offset, W, H); \
 959 }\
 960 static void ff_biweight_h264_pixels ## W ## x ## H ## _altivec(uint8_t *dst, uint8_t *src, int stride, int log2_denom, int weightd, int weights, int offset){ \
 961     biweight_h264_WxH_altivec(dst, src, stride, log2_denom, weightd, weights, offset, W, H); \
 962 }
 963
 964 H264_WEIGHT(16,16)
 965 H264_WEIGHT(16, 8)
 966 H264_WEIGHT( 8,16)
 967 H264_WEIGHT( 8, 8)
 968 H264_WEIGHT( 8, 4)
 969
 970 void dsputil_h264_init_ppc(DSPContext* c, AVCodecContext *avctx) {
 971
 972     if (has_altivec()) {
 973         c->put_h264_chroma_pixels_tab[0] = put_h264_chroma_mc8_altivec;
 974         c->avg_h264_chroma_pixels_tab[0] = avg_h264_chroma_mc8_altivec;
 975         c->put_no_rnd_vc1_chroma_pixels_tab[0] = put_no_rnd_vc1_chroma_mc8_altivec;
 976         c->avg_no_rnd_vc1_chroma_pixels_tab[0] = avg_no_rnd_vc1_chroma_mc8_altivec;
 977
 978 #define dspfunc(PFX, IDX, NUM) \
 979         c->PFX ## _pixels_tab[IDX][ 0] = PFX ## NUM ## _mc00_altivec; \
 980         c->PFX ## _pixels_tab[IDX][ 1] = PFX ## NUM ## _mc10_altivec; \
 981         c->PFX ## _pixels_tab[IDX][ 2] = PFX ## NUM ## _mc20_altivec; \
 982         c->PFX ## _pixels_tab[IDX][ 3] = PFX ## NUM ## _mc30_altivec; \
 983         c->PFX ## _pixels_tab[IDX][ 4] = PFX ## NUM ## _mc01_altivec; \
 984         c->PFX ## _pixels_tab[IDX][ 5] = PFX ## NUM ## _mc11_altivec; \
 985         c->PFX ## _pixels_tab[IDX][ 6] = PFX ## NUM ## _mc21_altivec; \
 986         c->PFX ## _pixels_tab[IDX][ 7] = PFX ## NUM ## _mc31_altivec; \
 987         c->PFX ## _pixels_tab[IDX][ 8] = PFX ## NUM ## _mc02_altivec; \
 988         c->PFX ## _pixels_tab[IDX][ 9] = PFX ## NUM ## _mc12_altivec; \
 989         c->PFX ## _pixels_tab[IDX][10] = PFX ## NUM ## _mc22_altivec; \
 990         c->PFX ## _pixels_tab[IDX][11] = PFX ## NUM ## _mc32_altivec; \
 991         c->PFX ## _pixels_tab[IDX][12] = PFX ## NUM ## _mc03_altivec; \
 992         c->PFX ## _pixels_tab[IDX][13] = PFX ## NUM ## _mc13_altivec; \
 993         c->PFX ## _pixels_tab[IDX][14] = PFX ## NUM ## _mc23_altivec; \
 994         c->PFX ## _pixels_tab[IDX][15] = PFX ## NUM ## _mc33_altivec
 995
 996         dspfunc(put_h264_qpel, 0, 16);
 997         dspfunc(avg_h264_qpel, 0, 16);
 998 #undef dspfunc
 999     }
1000 }
1001
1002 void ff_h264dsp_init_ppc(H264DSPContext *c)
1003 {
1004     if (has_altivec()) {
1005         c->h264_idct_add = ff_h264_idct_add_altivec;
1006         c->h264_idct_add8 = ff_h264_idct_add8_altivec;
1007         c->h264_idct_add16 = ff_h264_idct_add16_altivec;
1008         c->h264_idct_add16intra = ff_h264_idct_add16intra_altivec;
1009         c->h264_idct_dc_add= h264_idct_dc_add_altivec;
1010         c->h264_idct8_dc_add = ff_h264_idct8_dc_add_altivec;
1011         c->h264_idct8_add = ff_h264_idct8_add_altivec;
1012         c->h264_idct8_add4 = ff_h264_idct8_add4_altivec;
1013         c->h264_v_loop_filter_luma= h264_v_loop_filter_luma_altivec;
1014         c->h264_h_loop_filter_luma= h264_h_loop_filter_luma_altivec;
1015
1016         c->weight_h264_pixels_tab[0] = ff_weight_h264_pixels16x16_altivec;
1017         c->weight_h264_pixels_tab[1] = ff_weight_h264_pixels16x8_altivec;
1018         c->weight_h264_pixels_tab[2] = ff_weight_h264_pixels8x16_altivec;
1019         c->weight_h264_pixels_tab[3] = ff_weight_h264_pixels8x8_altivec;
1020         c->weight_h264_pixels_tab[4] = ff_weight_h264_pixels8x4_altivec;
1021         c->biweight_h264_pixels_tab[0] = ff_biweight_h264_pixels16x16_altivec;
1022         c->biweight_h264_pixels_tab[1] = ff_biweight_h264_pixels16x8_altivec;
1023         c->biweight_h264_pixels_tab[2] = ff_biweight_h264_pixels8x16_altivec;
1024         c->biweight_h264_pixels_tab[3] = ff_biweight_h264_pixels8x8_altivec;
1025         c->biweight_h264_pixels_tab[4] = ff_biweight_h264_pixels8x4_altivec;
1026     }
1027 }