/src/libvpx/vp8/common/reconinter.c

Source
/*
 *  Copyright (c) 2010 The WebM project authors. All Rights Reserved.
 *
 *  Use of this source code is governed by a BSD-style license
 *  that can be found in the LICENSE file in the root of the source
 *  tree. An additional intellectual property rights grant can be found
 *  in the file PATENTS.  All contributing project authors may
 *  be found in the AUTHORS file in the root of the source tree.
 */

#include <limits.h>
#include <string.h>

#include "vpx_config.h"
#include "vp8_rtcd.h"
#include "vpx/vpx_integer.h"
#include "blockd.h"
#include "reconinter.h"
#if CONFIG_RUNTIME_CPU_DETECT
#include "onyxc_int.h"
#endif

void vp8_copy_mem16x16_c(unsigned char *src, int src_stride, unsigned char *dst,
                         int dst_stride) {
  int r;

  for (r = 0; r < 16; ++r) {
    memcpy(dst, src, 16);

    src += src_stride;
    dst += dst_stride;
  }
}

void vp8_copy_mem8x8_c(unsigned char *src, int src_stride, unsigned char *dst,
                       int dst_stride) {
  int r;

  for (r = 0; r < 8; ++r) {
    memcpy(dst, src, 8);

    src += src_stride;
    dst += dst_stride;
  }
}

void vp8_copy_mem8x4_c(unsigned char *src, int src_stride, unsigned char *dst,
                       int dst_stride) {
  int r;

  for (r = 0; r < 4; ++r) {
    memcpy(dst, src, 8);

    src += src_stride;
    dst += dst_stride;
  }
}

void vp8_build_inter_predictors_b(BLOCKD *d, int pitch, unsigned char *base_pre,
                                  int pre_stride, vp8_subpix_fn_t sppf) {
  int r;
  unsigned char *pred_ptr = d->predictor;
  unsigned char *ptr;
  ptr = base_pre + d->offset + (d->bmi.mv.as_mv.row >> 3) * pre_stride +
        (d->bmi.mv.as_mv.col >> 3);

  if (d->bmi.mv.as_mv.row & 7 || d->bmi.mv.as_mv.col & 7) {
    sppf(ptr, pre_stride, d->bmi.mv.as_mv.col & 7, d->bmi.mv.as_mv.row & 7,
         pred_ptr, pitch);
  } else {
    for (r = 0; r < 4; ++r) {
      pred_ptr[0] = ptr[0];
      pred_ptr[1] = ptr[1];
      pred_ptr[2] = ptr[2];
      pred_ptr[3] = ptr[3];
      pred_ptr += pitch;
      ptr += pre_stride;
    }
  }
}

static void build_inter_predictors4b(MACROBLOCKD *x, BLOCKD *d,
                                     unsigned char *dst, int dst_stride,
                                     unsigned char *base_pre, int pre_stride) {
  unsigned char *ptr;
  ptr = base_pre + d->offset + (d->bmi.mv.as_mv.row >> 3) * pre_stride +
        (d->bmi.mv.as_mv.col >> 3);

  if (d->bmi.mv.as_mv.row & 7 || d->bmi.mv.as_mv.col & 7) {
    x->subpixel_predict8x8(ptr, pre_stride, d->bmi.mv.as_mv.col & 7,
                           d->bmi.mv.as_mv.row & 7, dst, dst_stride);
  } else {
    vp8_copy_mem8x8(ptr, pre_stride, dst, dst_stride);
  }
}

static void build_inter_predictors2b(MACROBLOCKD *x, BLOCKD *d,
                                     unsigned char *dst, int dst_stride,
                                     unsigned char *base_pre, int pre_stride) {
  unsigned char *ptr;
  ptr = base_pre + d->offset + (d->bmi.mv.as_mv.row >> 3) * pre_stride +
        (d->bmi.mv.as_mv.col >> 3);

  if (d->bmi.mv.as_mv.row & 7 || d->bmi.mv.as_mv.col & 7) {
    x->subpixel_predict8x4(ptr, pre_stride, d->bmi.mv.as_mv.col & 7,
                           d->bmi.mv.as_mv.row & 7, dst, dst_stride);
  } else {
    vp8_copy_mem8x4(ptr, pre_stride, dst, dst_stride);
  }
}

static void build_inter_predictors_b(BLOCKD *d, unsigned char *dst,
                                     int dst_stride, unsigned char *base_pre,
                                     int pre_stride, vp8_subpix_fn_t sppf) {
  int r;
  unsigned char *ptr;
  ptr = base_pre + d->offset + (d->bmi.mv.as_mv.row >> 3) * pre_stride +
        (d->bmi.mv.as_mv.col >> 3);

  if (d->bmi.mv.as_mv.row & 7 || d->bmi.mv.as_mv.col & 7) {
    sppf(ptr, pre_stride, d->bmi.mv.as_mv.col & 7, d->bmi.mv.as_mv.row & 7, dst,
         dst_stride);
  } else {
    for (r = 0; r < 4; ++r) {
      dst[0] = ptr[0];
      dst[1] = ptr[1];
      dst[2] = ptr[2];
      dst[3] = ptr[3];
      dst += dst_stride;
      ptr += pre_stride;
    }
  }
}

/*encoder only*/
void vp8_build_inter16x16_predictors_mbuv(MACROBLOCKD *x) {
  unsigned char *uptr, *vptr;
  unsigned char *upred_ptr = &x->predictor[256];
  unsigned char *vpred_ptr = &x->predictor[320];

  int mv_row = x->mode_info_context->mbmi.mv.as_mv.row;
  int mv_col = x->mode_info_context->mbmi.mv.as_mv.col;
  int offset;
  int pre_stride = x->pre.uv_stride;

  /* calc uv motion vectors */
  mv_row += 1 | (mv_row >> (sizeof(int) * CHAR_BIT - 1));
  mv_col += 1 | (mv_col >> (sizeof(int) * CHAR_BIT - 1));
  mv_row /= 2;
  mv_col /= 2;
  mv_row &= x->fullpixel_mask;
  mv_col &= x->fullpixel_mask;

  offset = (mv_row >> 3) * pre_stride + (mv_col >> 3);
  uptr = x->pre.u_buffer + offset;
  vptr = x->pre.v_buffer + offset;

  if ((mv_row | mv_col) & 7) {
    x->subpixel_predict8x8(uptr, pre_stride, mv_col & 7, mv_row & 7, upred_ptr,
                           8);
    x->subpixel_predict8x8(vptr, pre_stride, mv_col & 7, mv_row & 7, vpred_ptr,
                           8);
  } else {
    vp8_copy_mem8x8(uptr, pre_stride, upred_ptr, 8);
    vp8_copy_mem8x8(vptr, pre_stride, vpred_ptr, 8);
  }
}

/*encoder only*/
void vp8_build_inter4x4_predictors_mbuv(MACROBLOCKD *x) {
  int i, j;
  int pre_stride = x->pre.uv_stride;
  unsigned char *base_pre;

  /* build uv mvs */
  for (i = 0; i < 2; ++i) {
    for (j = 0; j < 2; ++j) {
      int yoffset = i * 8 + j * 2;
      int uoffset = 16 + i * 2 + j;
      int voffset = 20 + i * 2 + j;

      int temp;

      temp = x->block[yoffset].bmi.mv.as_mv.row +
             x->block[yoffset + 1].bmi.mv.as_mv.row +
             x->block[yoffset + 4].bmi.mv.as_mv.row +
             x->block[yoffset + 5].bmi.mv.as_mv.row;

      temp += 4 + ((temp >> (sizeof(temp) * CHAR_BIT - 1)) * 8);

      x->block[uoffset].bmi.mv.as_mv.row = (temp / 8) & x->fullpixel_mask;

      temp = x->block[yoffset].bmi.mv.as_mv.col +
             x->block[yoffset + 1].bmi.mv.as_mv.col +
             x->block[yoffset + 4].bmi.mv.as_mv.col +
             x->block[yoffset + 5].bmi.mv.as_mv.col;

      temp += 4 + ((temp >> (sizeof(temp) * CHAR_BIT - 1)) * 8);

      x->block[uoffset].bmi.mv.as_mv.col = (temp / 8) & x->fullpixel_mask;

      x->block[voffset].bmi.mv.as_int = x->block[uoffset].bmi.mv.as_int;
    }
  }

  base_pre = x->pre.u_buffer;
  for (i = 16; i < 20; i += 2) {
    BLOCKD *d0 = &x->block[i];
    BLOCKD *d1 = &x->block[i + 1];

    if (d0->bmi.mv.as_int == d1->bmi.mv.as_int) {
      build_inter_predictors2b(x, d0, d0->predictor, 8, base_pre, pre_stride);
    } else {
      vp8_build_inter_predictors_b(d0, 8, base_pre, pre_stride,
                                   x->subpixel_predict);
      vp8_build_inter_predictors_b(d1, 8, base_pre, pre_stride,
                                   x->subpixel_predict);
    }
  }

  base_pre = x->pre.v_buffer;
  for (i = 20; i < 24; i += 2) {
    BLOCKD *d0 = &x->block[i];
    BLOCKD *d1 = &x->block[i + 1];

    if (d0->bmi.mv.as_int == d1->bmi.mv.as_int) {
      build_inter_predictors2b(x, d0, d0->predictor, 8, base_pre, pre_stride);
    } else {
      vp8_build_inter_predictors_b(d0, 8, base_pre, pre_stride,
                                   x->subpixel_predict);
      vp8_build_inter_predictors_b(d1, 8, base_pre, pre_stride,
                                   x->subpixel_predict);
    }
  }
}

/*encoder only*/
void vp8_build_inter16x16_predictors_mby(MACROBLOCKD *x, unsigned char *dst_y,
                                         int dst_ystride) {
  unsigned char *ptr_base;
  unsigned char *ptr;
  int mv_row = x->mode_info_context->mbmi.mv.as_mv.row;
  int mv_col = x->mode_info_context->mbmi.mv.as_mv.col;
  int pre_stride = x->pre.y_stride;

  ptr_base = x->pre.y_buffer;
  ptr = ptr_base + (mv_row >> 3) * pre_stride + (mv_col >> 3);

  if ((mv_row | mv_col) & 7) {
    x->subpixel_predict16x16(ptr, pre_stride, mv_col & 7, mv_row & 7, dst_y,
                             dst_ystride);
  } else {
    vp8_copy_mem16x16(ptr, pre_stride, dst_y, dst_ystride);
  }
}

static void clamp_mv_to_umv_border(MV *mv, const MACROBLOCKD *xd) {
  /* If the MV points so far into the UMV border that no visible pixels
   * are used for reconstruction, the subpel part of the MV can be
   * discarded and the MV limited to 16 pixels with equivalent results.
   *
   * This limit kicks in at 19 pixels for the top and left edges, for
   * the 16 pixels plus 3 taps right of the central pixel when subpel
   * filtering. The bottom and right edges use 16 pixels plus 2 pixels
   * left of the central pixel when filtering.
   */
  if (mv->col < (xd->mb_to_left_edge - (19 << 3))) {
    mv->col = xd->mb_to_left_edge - (16 << 3);
  } else if (mv->col > xd->mb_to_right_edge + (18 << 3)) {
    mv->col = xd->mb_to_right_edge + (16 << 3);
  }

  if (mv->row < (xd->mb_to_top_edge - (19 << 3))) {
    mv->row = xd->mb_to_top_edge - (16 << 3);
  } else if (mv->row > xd->mb_to_bottom_edge + (18 << 3)) {
    mv->row = xd->mb_to_bottom_edge + (16 << 3);
  }
}

/* A version of the above function for chroma block MVs.*/
static void clamp_uvmv_to_umv_border(MV *mv, const MACROBLOCKD *xd) {
  mv->col = (2 * mv->col < (xd->mb_to_left_edge - (19 << 3)))
                ? (xd->mb_to_left_edge - (16 << 3)) >> 1
                : mv->col;
  mv->col = (2 * mv->col > xd->mb_to_right_edge + (18 << 3))
                ? (xd->mb_to_right_edge + (16 << 3)) >> 1
                : mv->col;

  mv->row = (2 * mv->row < (xd->mb_to_top_edge - (19 << 3)))
                ? (xd->mb_to_top_edge - (16 << 3)) >> 1
                : mv->row;
  mv->row = (2 * mv->row > xd->mb_to_bottom_edge + (18 << 3))
                ? (xd->mb_to_bottom_edge + (16 << 3)) >> 1
                : mv->row;
}

void vp8_build_inter16x16_predictors_mb(MACROBLOCKD *x, unsigned char *dst_y,
                                        unsigned char *dst_u,
                                        unsigned char *dst_v, int dst_ystride,
                                        int dst_uvstride) {
  int offset;
  unsigned char *ptr;
  unsigned char *uptr, *vptr;

  int_mv _16x16mv;

  unsigned char *ptr_base = x->pre.y_buffer;
  int pre_stride = x->pre.y_stride;

  _16x16mv.as_int = x->mode_info_context->mbmi.mv.as_int;

  if (x->mode_info_context->mbmi.need_to_clamp_mvs) {
    clamp_mv_to_umv_border(&_16x16mv.as_mv, x);
  }

  ptr = ptr_base + (_16x16mv.as_mv.row >> 3) * pre_stride +
        (_16x16mv.as_mv.col >> 3);

  if (_16x16mv.as_int & 0x00070007) {
    x->subpixel_predict16x16(ptr, pre_stride, _16x16mv.as_mv.col & 7,
                             _16x16mv.as_mv.row & 7, dst_y, dst_ystride);
  } else {
    vp8_copy_mem16x16(ptr, pre_stride, dst_y, dst_ystride);
  }

  /* calc uv motion vectors */
  _16x16mv.as_mv.row +=
      1 | (_16x16mv.as_mv.row >> (sizeof(int) * CHAR_BIT - 1));
  _16x16mv.as_mv.col +=
      1 | (_16x16mv.as_mv.col >> (sizeof(int) * CHAR_BIT - 1));
  _16x16mv.as_mv.row /= 2;
  _16x16mv.as_mv.col /= 2;
  _16x16mv.as_mv.row &= x->fullpixel_mask;
  _16x16mv.as_mv.col &= x->fullpixel_mask;

  /* Rounding to full-pel can move the chroma MV outside the border tap
   * window even though the luma MV was already clamped, so clamp again to
   * keep the reference read in bounds and ensure dst_u/dst_v are written. */
  clamp_uvmv_to_umv_border(&_16x16mv.as_mv, x);

  pre_stride >>= 1;
  offset = (_16x16mv.as_mv.row >> 3) * pre_stride + (_16x16mv.as_mv.col >> 3);
  uptr = x->pre.u_buffer + offset;
  vptr = x->pre.v_buffer + offset;

  if (_16x16mv.as_int & 0x00070007) {
    x->subpixel_predict8x8(uptr, pre_stride, _16x16mv.as_mv.col & 7,
                           _16x16mv.as_mv.row & 7, dst_u, dst_uvstride);
    x->subpixel_predict8x8(vptr, pre_stride, _16x16mv.as_mv.col & 7,
                           _16x16mv.as_mv.row & 7, dst_v, dst_uvstride);
  } else {
    vp8_copy_mem8x8(uptr, pre_stride, dst_u, dst_uvstride);
    vp8_copy_mem8x8(vptr, pre_stride, dst_v, dst_uvstride);
  }
}

static void build_inter4x4_predictors_mb(MACROBLOCKD *x) {
  int i;
  unsigned char *base_dst = x->dst.y_buffer;
  unsigned char *base_pre = x->pre.y_buffer;

  if (x->mode_info_context->mbmi.partitioning < 3) {
    BLOCKD *b;
    int dst_stride = x->dst.y_stride;

    x->block[0].bmi = x->mode_info_context->bmi[0];
    x->block[2].bmi = x->mode_info_context->bmi[2];
    x->block[8].bmi = x->mode_info_context->bmi[8];
    x->block[10].bmi = x->mode_info_context->bmi[10];
    if (x->mode_info_context->mbmi.need_to_clamp_mvs) {
      clamp_mv_to_umv_border(&x->block[0].bmi.mv.as_mv, x);
      clamp_mv_to_umv_border(&x->block[2].bmi.mv.as_mv, x);
      clamp_mv_to_umv_border(&x->block[8].bmi.mv.as_mv, x);
      clamp_mv_to_umv_border(&x->block[10].bmi.mv.as_mv, x);
    }

    b = &x->block[0];
    build_inter_predictors4b(x, b, base_dst + b->offset, dst_stride, base_pre,
                             dst_stride);
    b = &x->block[2];
    build_inter_predictors4b(x, b, base_dst + b->offset, dst_stride, base_pre,
                             dst_stride);
    b = &x->block[8];
    build_inter_predictors4b(x, b, base_dst + b->offset, dst_stride, base_pre,
                             dst_stride);
    b = &x->block[10];
    build_inter_predictors4b(x, b, base_dst + b->offset, dst_stride, base_pre,
                             dst_stride);
  } else {
    for (i = 0; i < 16; i += 2) {
      BLOCKD *d0 = &x->block[i];
      BLOCKD *d1 = &x->block[i + 1];
      int dst_stride = x->dst.y_stride;

      x->block[i + 0].bmi = x->mode_info_context->bmi[i + 0];
      x->block[i + 1].bmi = x->mode_info_context->bmi[i + 1];
      if (x->mode_info_context->mbmi.need_to_clamp_mvs) {
        clamp_mv_to_umv_border(&x->block[i + 0].bmi.mv.as_mv, x);
        clamp_mv_to_umv_border(&x->block[i + 1].bmi.mv.as_mv, x);
      }

      if (d0->bmi.mv.as_int == d1->bmi.mv.as_int) {
        build_inter_predictors2b(x, d0, base_dst + d0->offset, dst_stride,
                                 base_pre, dst_stride);
      } else {
        build_inter_predictors_b(d0, base_dst + d0->offset, dst_stride,
                                 base_pre, dst_stride, x->subpixel_predict);
        build_inter_predictors_b(d1, base_dst + d1->offset, dst_stride,
                                 base_pre, dst_stride, x->subpixel_predict);
      }
    }
  }
  base_dst = x->dst.u_buffer;
  base_pre = x->pre.u_buffer;
  for (i = 16; i < 20; i += 2) {
    BLOCKD *d0 = &x->block[i];
    BLOCKD *d1 = &x->block[i + 1];
    int dst_stride = x->dst.uv_stride;

    /* Note: uv mvs already clamped in build_4x4uvmvs() */

    if (d0->bmi.mv.as_int == d1->bmi.mv.as_int) {
      build_inter_predictors2b(x, d0, base_dst + d0->offset, dst_stride,
                               base_pre, dst_stride);
    } else {
      build_inter_predictors_b(d0, base_dst + d0->offset, dst_stride, base_pre,
                               dst_stride, x->subpixel_predict);
      build_inter_predictors_b(d1, base_dst + d1->offset, dst_stride, base_pre,
                               dst_stride, x->subpixel_predict);
    }
  }

  base_dst = x->dst.v_buffer;
  base_pre = x->pre.v_buffer;
  for (i = 20; i < 24; i += 2) {
    BLOCKD *d0 = &x->block[i];
    BLOCKD *d1 = &x->block[i + 1];
    int dst_stride = x->dst.uv_stride;

    /* Note: uv mvs already clamped in build_4x4uvmvs() */

    if (d0->bmi.mv.as_int == d1->bmi.mv.as_int) {
      build_inter_predictors2b(x, d0, base_dst + d0->offset, dst_stride,
                               base_pre, dst_stride);
    } else {
      build_inter_predictors_b(d0, base_dst + d0->offset, dst_stride, base_pre,
                               dst_stride, x->subpixel_predict);
      build_inter_predictors_b(d1, base_dst + d1->offset, dst_stride, base_pre,
                               dst_stride, x->subpixel_predict);
    }
  }
}

static void build_4x4uvmvs(MACROBLOCKD *x) {
  int i, j;

  for (i = 0; i < 2; ++i) {
    for (j = 0; j < 2; ++j) {
      int yoffset = i * 8 + j * 2;
      int uoffset = 16 + i * 2 + j;
      int voffset = 20 + i * 2 + j;

      int temp;

      temp = x->mode_info_context->bmi[yoffset + 0].mv.as_mv.row +
             x->mode_info_context->bmi[yoffset + 1].mv.as_mv.row +
             x->mode_info_context->bmi[yoffset + 4].mv.as_mv.row +
             x->mode_info_context->bmi[yoffset + 5].mv.as_mv.row;

      temp += 4 + ((temp >> (sizeof(temp) * CHAR_BIT - 1)) * 8);

      x->block[uoffset].bmi.mv.as_mv.row = (temp / 8) & x->fullpixel_mask;

      temp = x->mode_info_context->bmi[yoffset + 0].mv.as_mv.col +
             x->mode_info_context->bmi[yoffset + 1].mv.as_mv.col +
             x->mode_info_context->bmi[yoffset + 4].mv.as_mv.col +
             x->mode_info_context->bmi[yoffset + 5].mv.as_mv.col;

      temp += 4 + ((temp >> (sizeof(temp) * CHAR_BIT - 1)) * 8);

      x->block[uoffset].bmi.mv.as_mv.col = (temp / 8) & x->fullpixel_mask;

      if (x->mode_info_context->mbmi.need_to_clamp_mvs) {
        clamp_uvmv_to_umv_border(&x->block[uoffset].bmi.mv.as_mv, x);
      }

      x->block[voffset].bmi.mv.as_int = x->block[uoffset].bmi.mv.as_int;
    }
  }
}

void vp8_build_inter_predictors_mb(MACROBLOCKD *xd) {
  if (xd->mode_info_context->mbmi.mode != SPLITMV) {
    vp8_build_inter16x16_predictors_mb(xd, xd->dst.y_buffer, xd->dst.u_buffer,
                                       xd->dst.v_buffer, xd->dst.y_stride,
                                       xd->dst.uv_stride);
  } else {
    build_4x4uvmvs(xd);
    build_inter4x4_predictors_mb(xd);
  }
}

Coverage Report

Created: 2026-10-28 08:13

Line	Count	Source
1		/*
2		* Copyright (c) 2010 The WebM project authors. All Rights Reserved.
3		*
4		* Use of this source code is governed by a BSD-style license
5		* that can be found in the LICENSE file in the root of the source
6		* tree. An additional intellectual property rights grant can be found
7		* in the file PATENTS. All contributing project authors may
8		* be found in the AUTHORS file in the root of the source tree.
9		*/
10
11		#include <limits.h>
12		#include <string.h>
13
14		#include "vpx_config.h"
15		#include "vp8_rtcd.h"
16		#include "vpx/vpx_integer.h"
17		#include "blockd.h"
18		#include "reconinter.h"
19		#if CONFIG_RUNTIME_CPU_DETECT
20		#include "onyxc_int.h"
21		#endif
22
23		void vp8_copy_mem16x16_c(unsigned char src, int src_stride, unsigned char dst,
24	0	int dst_stride) {
25	0	int r;
26
27	0	for (r = 0; r < 16; ++r) {
28	0	memcpy(dst, src, 16);
29
30	0	src += src_stride;
31	0	dst += dst_stride;
32	0	}
33	0	}
34
35		void vp8_copy_mem8x8_c(unsigned char src, int src_stride, unsigned char dst,
36	0	int dst_stride) {
37	0	int r;
38
39	0	for (r = 0; r < 8; ++r) {
40	0	memcpy(dst, src, 8);
41
42	0	src += src_stride;
43	0	dst += dst_stride;
44	0	}
45	0	}
46
47		void vp8_copy_mem8x4_c(unsigned char src, int src_stride, unsigned char dst,
48	0	int dst_stride) {
49	0	int r;
50
51	0	for (r = 0; r < 4; ++r) {
52	0	memcpy(dst, src, 8);
53
54	0	src += src_stride;
55	0	dst += dst_stride;
56	0	}
57	0	}
58
59		void vp8_build_inter_predictors_b(BLOCKD d, int pitch, unsigned char base_pre,
60	85.9M	int pre_stride, vp8_subpix_fn_t sppf) {
61	85.9M	int r;
62	85.9M	unsigned char *pred_ptr = d->predictor;
63	85.9M	unsigned char *ptr;
64	85.9M	ptr = base_pre + d->offset + (d->bmi.mv.as_mv.row >> 3) * pre_stride +
65	85.9M	(d->bmi.mv.as_mv.col >> 3);
66
67	85.9M	if (d->bmi.mv.as_mv.row & 7 \|\| d->bmi.mv.as_mv.col & 7) {
68	22.3M	sppf(ptr, pre_stride, d->bmi.mv.as_mv.col & 7, d->bmi.mv.as_mv.row & 7,
69	22.3M	pred_ptr, pitch);
70	63.6M	} else {
71	318M	for (r = 0; r < 4; ++r) {
72	254M	pred_ptr[0] = ptr[0];
73	254M	pred_ptr[1] = ptr[1];
74	254M	pred_ptr[2] = ptr[2];
75	254M	pred_ptr[3] = ptr[3];
76	254M	pred_ptr += pitch;
77	254M	ptr += pre_stride;
78	254M	}
79	63.6M	}
80	85.9M	}
81
82		static void build_inter_predictors4b(MACROBLOCKD x, BLOCKD d,
83		unsigned char *dst, int dst_stride,
84	2.80M	unsigned char *base_pre, int pre_stride) {
85	2.80M	unsigned char *ptr;
86	2.80M	ptr = base_pre + d->offset + (d->bmi.mv.as_mv.row >> 3) * pre_stride +
87	2.80M	(d->bmi.mv.as_mv.col >> 3);
88
89	2.80M	if (d->bmi.mv.as_mv.row & 7 \|\| d->bmi.mv.as_mv.col & 7) {
90	1.39M	x->subpixel_predict8x8(ptr, pre_stride, d->bmi.mv.as_mv.col & 7,
91	1.39M	d->bmi.mv.as_mv.row & 7, dst, dst_stride);
92	1.41M	} else {
93	1.41M	vp8_copy_mem8x8(ptr, pre_stride, dst, dst_stride);
94	1.41M	}
95	2.80M	}
96
97		static void build_inter_predictors2b(MACROBLOCKD x, BLOCKD d,
98		unsigned char *dst, int dst_stride,
99	6.31M	unsigned char *base_pre, int pre_stride) {
100	6.31M	unsigned char *ptr;
101	6.31M	ptr = base_pre + d->offset + (d->bmi.mv.as_mv.row >> 3) * pre_stride +
102	6.31M	(d->bmi.mv.as_mv.col >> 3);
103
104	6.31M	if (d->bmi.mv.as_mv.row & 7 \|\| d->bmi.mv.as_mv.col & 7) {
105	2.43M	x->subpixel_predict8x4(ptr, pre_stride, d->bmi.mv.as_mv.col & 7,
106	2.43M	d->bmi.mv.as_mv.row & 7, dst, dst_stride);
107	3.87M	} else {
108	3.87M	vp8_copy_mem8x4(ptr, pre_stride, dst, dst_stride);
109	3.87M	}
110	6.31M	}
111
112		static void build_inter_predictors_b(BLOCKD d, unsigned char dst,
113		int dst_stride, unsigned char *base_pre,
114	6.64M	int pre_stride, vp8_subpix_fn_t sppf) {
115	6.64M	int r;
116	6.64M	unsigned char *ptr;
117	6.64M	ptr = base_pre + d->offset + (d->bmi.mv.as_mv.row >> 3) * pre_stride +
118	6.64M	(d->bmi.mv.as_mv.col >> 3);
119
120	6.64M	if (d->bmi.mv.as_mv.row & 7 \|\| d->bmi.mv.as_mv.col & 7) {
121	2.53M	sppf(ptr, pre_stride, d->bmi.mv.as_mv.col & 7, d->bmi.mv.as_mv.row & 7, dst,
122	2.53M	dst_stride);
123	4.11M	} else {
124	20.5M	for (r = 0; r < 4; ++r) {
125	16.4M	dst[0] = ptr[0];
126	16.4M	dst[1] = ptr[1];
127	16.4M	dst[2] = ptr[2];
128	16.4M	dst[3] = ptr[3];
129	16.4M	dst += dst_stride;
130	16.4M	ptr += pre_stride;
131	16.4M	}
132	4.11M	}
133	6.64M	}
134
135		/encoder only/
136	2.61M	void vp8_build_inter16x16_predictors_mbuv(MACROBLOCKD *x) {
137	2.61M	unsigned char uptr, vptr;
138	2.61M	unsigned char *upred_ptr = &x->predictor[256];
139	2.61M	unsigned char *vpred_ptr = &x->predictor[320];
140
141	2.61M	int mv_row = x->mode_info_context->mbmi.mv.as_mv.row;
142	2.61M	int mv_col = x->mode_info_context->mbmi.mv.as_mv.col;
143	2.61M	int offset;
144	2.61M	int pre_stride = x->pre.uv_stride;
145
146		/* calc uv motion vectors */
147	2.61M	mv_row += 1 \| (mv_row >> (sizeof(int) * CHAR_BIT - 1));
148	2.61M	mv_col += 1 \| (mv_col >> (sizeof(int) * CHAR_BIT - 1));
149	2.61M	mv_row /= 2;
150	2.61M	mv_col /= 2;
151	2.61M	mv_row &= x->fullpixel_mask;
152	2.61M	mv_col &= x->fullpixel_mask;
153
154	2.61M	offset = (mv_row >> 3) * pre_stride + (mv_col >> 3);
155	2.61M	uptr = x->pre.u_buffer + offset;
156	2.61M	vptr = x->pre.v_buffer + offset;
157
158	2.61M	if ((mv_row \| mv_col) & 7) {
159	1.20M	x->subpixel_predict8x8(uptr, pre_stride, mv_col & 7, mv_row & 7, upred_ptr,
160	1.20M	8);
161	1.20M	x->subpixel_predict8x8(vptr, pre_stride, mv_col & 7, mv_row & 7, vpred_ptr,
162	1.20M	8);
163	1.40M	} else {
164	1.40M	vp8_copy_mem8x8(uptr, pre_stride, upred_ptr, 8);
165	1.40M	vp8_copy_mem8x8(vptr, pre_stride, vpred_ptr, 8);
166	1.40M	}
167	2.61M	}
168
169		/encoder only/
170	354k	void vp8_build_inter4x4_predictors_mbuv(MACROBLOCKD *x) {
171	354k	int i, j;
172	354k	int pre_stride = x->pre.uv_stride;
173	354k	unsigned char *base_pre;
174
175		/* build uv mvs */
176	1.06M	for (i = 0; i < 2; ++i) {
177	2.12M	for (j = 0; j < 2; ++j) {
178	1.41M	int yoffset = i * 8 + j * 2;
179	1.41M	int uoffset = 16 + i * 2 + j;
180	1.41M	int voffset = 20 + i * 2 + j;
181
182	1.41M	int temp;
183
184	1.41M	temp = x->block[yoffset].bmi.mv.as_mv.row +
185	1.41M	x->block[yoffset + 1].bmi.mv.as_mv.row +
186	1.41M	x->block[yoffset + 4].bmi.mv.as_mv.row +
187	1.41M	x->block[yoffset + 5].bmi.mv.as_mv.row;
188
189	1.41M	temp += 4 + ((temp >> (sizeof(temp) * CHAR_BIT - 1)) * 8);
190
191	1.41M	x->block[uoffset].bmi.mv.as_mv.row = (temp / 8) & x->fullpixel_mask;
192
193	1.41M	temp = x->block[yoffset].bmi.mv.as_mv.col +
194	1.41M	x->block[yoffset + 1].bmi.mv.as_mv.col +
195	1.41M	x->block[yoffset + 4].bmi.mv.as_mv.col +
196	1.41M	x->block[yoffset + 5].bmi.mv.as_mv.col;
197
198	1.41M	temp += 4 + ((temp >> (sizeof(temp) * CHAR_BIT - 1)) * 8);
199
200	1.41M	x->block[uoffset].bmi.mv.as_mv.col = (temp / 8) & x->fullpixel_mask;
201
202	1.41M	x->block[voffset].bmi.mv.as_int = x->block[uoffset].bmi.mv.as_int;
203	1.41M	}
204	709k	}
205
206	354k	base_pre = x->pre.u_buffer;
207	1.06M	for (i = 16; i < 20; i += 2) {
208	709k	BLOCKD *d0 = &x->block[i];
209	709k	BLOCKD *d1 = &x->block[i + 1];
210
211	709k	if (d0->bmi.mv.as_int == d1->bmi.mv.as_int) {
212	398k	build_inter_predictors2b(x, d0, d0->predictor, 8, base_pre, pre_stride);
213	398k	} else {
214	311k	vp8_build_inter_predictors_b(d0, 8, base_pre, pre_stride,
215	311k	x->subpixel_predict);
216	311k	vp8_build_inter_predictors_b(d1, 8, base_pre, pre_stride,
217	311k	x->subpixel_predict);
218	311k	}
219	709k	}
220
221	354k	base_pre = x->pre.v_buffer;
222	1.06M	for (i = 20; i < 24; i += 2) {
223	709k	BLOCKD *d0 = &x->block[i];
224	709k	BLOCKD *d1 = &x->block[i + 1];
225
226	709k	if (d0->bmi.mv.as_int == d1->bmi.mv.as_int) {
227	398k	build_inter_predictors2b(x, d0, d0->predictor, 8, base_pre, pre_stride);
228	398k	} else {
229	311k	vp8_build_inter_predictors_b(d0, 8, base_pre, pre_stride,
230	311k	x->subpixel_predict);
231	311k	vp8_build_inter_predictors_b(d1, 8, base_pre, pre_stride,
232	311k	x->subpixel_predict);
233	311k	}
234	709k	}
235	354k	}
236
237		/encoder only/
238		void vp8_build_inter16x16_predictors_mby(MACROBLOCKD x, unsigned char dst_y,
239	2.61M	int dst_ystride) {
240	2.61M	unsigned char *ptr_base;
241	2.61M	unsigned char *ptr;
242	2.61M	int mv_row = x->mode_info_context->mbmi.mv.as_mv.row;
243	2.61M	int mv_col = x->mode_info_context->mbmi.mv.as_mv.col;
244	2.61M	int pre_stride = x->pre.y_stride;
245
246	2.61M	ptr_base = x->pre.y_buffer;
247	2.61M	ptr = ptr_base + (mv_row >> 3) * pre_stride + (mv_col >> 3);
248
249	2.61M	if ((mv_row \| mv_col) & 7) {
250	640k	x->subpixel_predict16x16(ptr, pre_stride, mv_col & 7, mv_row & 7, dst_y,
251	640k	dst_ystride);
252	1.96M	} else {
253	1.96M	vp8_copy_mem16x16(ptr, pre_stride, dst_y, dst_ystride);
254	1.96M	}
255	2.61M	}
256
257	4.57M	static void clamp_mv_to_umv_border(MV mv, const MACROBLOCKD xd) {
258		/* If the MV points so far into the UMV border that no visible pixels
259		* are used for reconstruction, the subpel part of the MV can be
260		* discarded and the MV limited to 16 pixels with equivalent results.
261		*
262		* This limit kicks in at 19 pixels for the top and left edges, for
263		* the 16 pixels plus 3 taps right of the central pixel when subpel
264		* filtering. The bottom and right edges use 16 pixels plus 2 pixels
265		* left of the central pixel when filtering.
266		*/
267	4.57M	if (mv->col < (xd->mb_to_left_edge - (19 << 3))) {
268	338k	mv->col = xd->mb_to_left_edge - (16 << 3);
269	4.23M	} else if (mv->col > xd->mb_to_right_edge + (18 << 3)) {
270	328k	mv->col = xd->mb_to_right_edge + (16 << 3);
271	328k	}
272
273	4.57M	if (mv->row < (xd->mb_to_top_edge - (19 << 3))) {
274	833k	mv->row = xd->mb_to_top_edge - (16 << 3);
275	3.74M	} else if (mv->row > xd->mb_to_bottom_edge + (18 << 3)) {
276	446k	mv->row = xd->mb_to_bottom_edge + (16 << 3);
277	446k	}
278	4.57M	}
279
280		/* A version of the above function for chroma block MVs.*/
281	3.96M	static void clamp_uvmv_to_umv_border(MV mv, const MACROBLOCKD xd) {
282	3.96M	mv->col = (2 * mv->col < (xd->mb_to_left_edge - (19 << 3)))
283	3.96M	? (xd->mb_to_left_edge - (16 << 3)) >> 1
284	3.96M	: mv->col;
285	3.96M	mv->col = (2 * mv->col > xd->mb_to_right_edge + (18 << 3))
286	3.96M	? (xd->mb_to_right_edge + (16 << 3)) >> 1
287	3.96M	: mv->col;
288
289	3.96M	mv->row = (2 * mv->row < (xd->mb_to_top_edge - (19 << 3)))
290	3.96M	? (xd->mb_to_top_edge - (16 << 3)) >> 1
291	3.96M	: mv->row;
292	3.96M	mv->row = (2 * mv->row > xd->mb_to_bottom_edge + (18 << 3))
293	3.96M	? (xd->mb_to_bottom_edge + (16 << 3)) >> 1
294	3.96M	: mv->row;
295	3.96M	}
296
297		void vp8_build_inter16x16_predictors_mb(MACROBLOCKD x, unsigned char dst_y,
298		unsigned char *dst_u,
299		unsigned char *dst_v, int dst_ystride,
300	2.40M	int dst_uvstride) {
301	2.40M	int offset;
302	2.40M	unsigned char *ptr;
303	2.40M	unsigned char uptr, vptr;
304
305	2.40M	int_mv _16x16mv;
306
307	2.40M	unsigned char *ptr_base = x->pre.y_buffer;
308	2.40M	int pre_stride = x->pre.y_stride;
309
310	2.40M	_16x16mv.as_int = x->mode_info_context->mbmi.mv.as_int;
311
312	2.40M	if (x->mode_info_context->mbmi.need_to_clamp_mvs) {
313	291k	clamp_mv_to_umv_border(&_16x16mv.as_mv, x);
314	291k	}
315
316	2.40M	ptr = ptr_base + (_16x16mv.as_mv.row >> 3) * pre_stride +
317	2.40M	(_16x16mv.as_mv.col >> 3);
318
319	2.40M	if (_16x16mv.as_int & 0x00070007) {
320	1.39M	x->subpixel_predict16x16(ptr, pre_stride, _16x16mv.as_mv.col & 7,
321	1.39M	_16x16mv.as_mv.row & 7, dst_y, dst_ystride);
322	1.39M	} else {
323	1.01M	vp8_copy_mem16x16(ptr, pre_stride, dst_y, dst_ystride);
324	1.01M	}
325
326		/* calc uv motion vectors */
327	2.40M	_16x16mv.as_mv.row +=
328	2.40M	1 \| (_16x16mv.as_mv.row >> (sizeof(int) * CHAR_BIT - 1));
329	2.40M	_16x16mv.as_mv.col +=
330	2.40M	1 \| (_16x16mv.as_mv.col >> (sizeof(int) * CHAR_BIT - 1));
331	2.40M	_16x16mv.as_mv.row /= 2;
332	2.40M	_16x16mv.as_mv.col /= 2;
333	2.40M	_16x16mv.as_mv.row &= x->fullpixel_mask;
334	2.40M	_16x16mv.as_mv.col &= x->fullpixel_mask;
335
336		/* Rounding to full-pel can move the chroma MV outside the border tap
337		* window even though the luma MV was already clamped, so clamp again to
338		* keep the reference read in bounds and ensure dst_u/dst_v are written. */
339	2.40M	clamp_uvmv_to_umv_border(&_16x16mv.as_mv, x);
340
341	2.40M	pre_stride >>= 1;
342	2.40M	offset = (_16x16mv.as_mv.row >> 3) * pre_stride + (_16x16mv.as_mv.col >> 3);
343	2.40M	uptr = x->pre.u_buffer + offset;
344	2.40M	vptr = x->pre.v_buffer + offset;
345
346	2.40M	if (_16x16mv.as_int & 0x00070007) {
347	404k	x->subpixel_predict8x8(uptr, pre_stride, _16x16mv.as_mv.col & 7,
348	404k	_16x16mv.as_mv.row & 7, dst_u, dst_uvstride);
349	404k	x->subpixel_predict8x8(vptr, pre_stride, _16x16mv.as_mv.col & 7,
350	404k	_16x16mv.as_mv.row & 7, dst_v, dst_uvstride);
351	2.00M	} else {
352	2.00M	vp8_copy_mem8x8(uptr, pre_stride, dst_u, dst_uvstride);
353	2.00M	vp8_copy_mem8x8(vptr, pre_stride, dst_v, dst_uvstride);
354	2.00M	}
355	2.40M	}
356
357	1.20M	static void build_inter4x4_predictors_mb(MACROBLOCKD *x) {
358	1.20M	int i;
359	1.20M	unsigned char *base_dst = x->dst.y_buffer;
360	1.20M	unsigned char *base_pre = x->pre.y_buffer;
361
362	1.20M	if (x->mode_info_context->mbmi.partitioning < 3) {
363	701k	BLOCKD *b;
364	701k	int dst_stride = x->dst.y_stride;
365
366	701k	x->block[0].bmi = x->mode_info_context->bmi[0];
367	701k	x->block[2].bmi = x->mode_info_context->bmi[2];
368	701k	x->block[8].bmi = x->mode_info_context->bmi[8];
369	701k	x->block[10].bmi = x->mode_info_context->bmi[10];
370	701k	if (x->mode_info_context->mbmi.need_to_clamp_mvs) {
371	163k	clamp_mv_to_umv_border(&x->block[0].bmi.mv.as_mv, x);
372	163k	clamp_mv_to_umv_border(&x->block[2].bmi.mv.as_mv, x);
373	163k	clamp_mv_to_umv_border(&x->block[8].bmi.mv.as_mv, x);
374	163k	clamp_mv_to_umv_border(&x->block[10].bmi.mv.as_mv, x);
375	163k	}
376
377	701k	b = &x->block[0];
378	701k	build_inter_predictors4b(x, b, base_dst + b->offset, dst_stride, base_pre,
379	701k	dst_stride);
380	701k	b = &x->block[2];
381	701k	build_inter_predictors4b(x, b, base_dst + b->offset, dst_stride, base_pre,
382	701k	dst_stride);
383	701k	b = &x->block[8];
384	701k	build_inter_predictors4b(x, b, base_dst + b->offset, dst_stride, base_pre,
385	701k	dst_stride);
386	701k	b = &x->block[10];
387	701k	build_inter_predictors4b(x, b, base_dst + b->offset, dst_stride, base_pre,
388	701k	dst_stride);
389	701k	} else {
390	4.52M	for (i = 0; i < 16; i += 2) {
391	4.01M	BLOCKD *d0 = &x->block[i];
392	4.01M	BLOCKD *d1 = &x->block[i + 1];
393	4.01M	int dst_stride = x->dst.y_stride;
394
395	4.01M	x->block[i + 0].bmi = x->mode_info_context->bmi[i + 0];
396	4.01M	x->block[i + 1].bmi = x->mode_info_context->bmi[i + 1];
397	4.01M	if (x->mode_info_context->mbmi.need_to_clamp_mvs) {
398	1.81M	clamp_mv_to_umv_border(&x->block[i + 0].bmi.mv.as_mv, x);
399	1.81M	clamp_mv_to_umv_border(&x->block[i + 1].bmi.mv.as_mv, x);
400	1.81M	}
401
402	4.01M	if (d0->bmi.mv.as_int == d1->bmi.mv.as_int) {
403	2.69M	build_inter_predictors2b(x, d0, base_dst + d0->offset, dst_stride,
404	2.69M	base_pre, dst_stride);
405	2.69M	} else {
406	1.32M	build_inter_predictors_b(d0, base_dst + d0->offset, dst_stride,
407	1.32M	base_pre, dst_stride, x->subpixel_predict);
408	1.32M	build_inter_predictors_b(d1, base_dst + d1->offset, dst_stride,
409	1.32M	base_pre, dst_stride, x->subpixel_predict);
410	1.32M	}
411	4.01M	}
412	502k	}
413	1.20M	base_dst = x->dst.u_buffer;
414	1.20M	base_pre = x->pre.u_buffer;
415	3.61M	for (i = 16; i < 20; i += 2) {
416	2.40M	BLOCKD *d0 = &x->block[i];
417	2.40M	BLOCKD *d1 = &x->block[i + 1];
418	2.40M	int dst_stride = x->dst.uv_stride;
419
420		/* Note: uv mvs already clamped in build_4x4uvmvs() */
421
422	2.40M	if (d0->bmi.mv.as_int == d1->bmi.mv.as_int) {
423	1.40M	build_inter_predictors2b(x, d0, base_dst + d0->offset, dst_stride,
424	1.40M	base_pre, dst_stride);
425	1.40M	} else {
426	1.00M	build_inter_predictors_b(d0, base_dst + d0->offset, dst_stride, base_pre,
427	1.00M	dst_stride, x->subpixel_predict);
428	1.00M	build_inter_predictors_b(d1, base_dst + d1->offset, dst_stride, base_pre,
429	1.00M	dst_stride, x->subpixel_predict);
430	1.00M	}
431	2.40M	}
432
433	1.20M	base_dst = x->dst.v_buffer;
434	1.20M	base_pre = x->pre.v_buffer;
435	3.61M	for (i = 20; i < 24; i += 2) {
436	2.40M	BLOCKD *d0 = &x->block[i];
437	2.40M	BLOCKD *d1 = &x->block[i + 1];
438	2.40M	int dst_stride = x->dst.uv_stride;
439
440		/* Note: uv mvs already clamped in build_4x4uvmvs() */
441
442	2.40M	if (d0->bmi.mv.as_int == d1->bmi.mv.as_int) {
443	1.40M	build_inter_predictors2b(x, d0, base_dst + d0->offset, dst_stride,
444	1.40M	base_pre, dst_stride);
445	1.40M	} else {
446	1.00M	build_inter_predictors_b(d0, base_dst + d0->offset, dst_stride, base_pre,
447	1.00M	dst_stride, x->subpixel_predict);
448	1.00M	build_inter_predictors_b(d1, base_dst + d1->offset, dst_stride, base_pre,
449	1.00M	dst_stride, x->subpixel_predict);
450	1.00M	}
451	2.40M	}
452	1.20M	}
453
454	1.20M	static void build_4x4uvmvs(MACROBLOCKD *x) {
455	1.20M	int i, j;
456
457	3.61M	for (i = 0; i < 2; ++i) {
458	7.22M	for (j = 0; j < 2; ++j) {
459	4.81M	int yoffset = i * 8 + j * 2;
460	4.81M	int uoffset = 16 + i * 2 + j;
461	4.81M	int voffset = 20 + i * 2 + j;
462
463	4.81M	int temp;
464
465	4.81M	temp = x->mode_info_context->bmi[yoffset + 0].mv.as_mv.row +
466	4.81M	x->mode_info_context->bmi[yoffset + 1].mv.as_mv.row +
467	4.81M	x->mode_info_context->bmi[yoffset + 4].mv.as_mv.row +
468	4.81M	x->mode_info_context->bmi[yoffset + 5].mv.as_mv.row;
469
470	4.81M	temp += 4 + ((temp >> (sizeof(temp) * CHAR_BIT - 1)) * 8);
471
472	4.81M	x->block[uoffset].bmi.mv.as_mv.row = (temp / 8) & x->fullpixel_mask;
473
474	4.81M	temp = x->mode_info_context->bmi[yoffset + 0].mv.as_mv.col +
475	4.81M	x->mode_info_context->bmi[yoffset + 1].mv.as_mv.col +
476	4.81M	x->mode_info_context->bmi[yoffset + 4].mv.as_mv.col +
477	4.81M	x->mode_info_context->bmi[yoffset + 5].mv.as_mv.col;
478
479	4.81M	temp += 4 + ((temp >> (sizeof(temp) * CHAR_BIT - 1)) * 8);
480
481	4.81M	x->block[uoffset].bmi.mv.as_mv.col = (temp / 8) & x->fullpixel_mask;
482
483	4.81M	if (x->mode_info_context->mbmi.need_to_clamp_mvs) {
484	1.56M	clamp_uvmv_to_umv_border(&x->block[uoffset].bmi.mv.as_mv, x);
485	1.56M	}
486
487	4.81M	x->block[voffset].bmi.mv.as_int = x->block[uoffset].bmi.mv.as_int;
488	4.81M	}
489	2.40M	}
490	1.20M	}
491
492	3.60M	void vp8_build_inter_predictors_mb(MACROBLOCKD *xd) {
493	3.60M	if (xd->mode_info_context->mbmi.mode != SPLITMV) {
494	2.40M	vp8_build_inter16x16_predictors_mb(xd, xd->dst.y_buffer, xd->dst.u_buffer,
495	2.40M	xd->dst.v_buffer, xd->dst.y_stride,
496	2.40M	xd->dst.uv_stride);
497	2.40M	} else {
498	1.20M	build_4x4uvmvs(xd);
499	1.20M	build_inter4x4_predictors_mb(xd);
500	1.20M	}
501	3.60M	}