macroblock.c

linux下编译已经通过

        for( i = 0; i < 4; i++ )
            x264_mb_mc_8x8( h, i );
    }
    else if( h->mb.i_type == B_SKIP || h->mb.i_type == B_DIRECT )
    {
        x264_mb_mc_direct8x8( h, 0, 0 );
        x264_mb_mc_direct8x8( h, 2, 0 );
        x264_mb_mc_direct8x8( h, 0, 2 );
        x264_mb_mc_direct8x8( h, 2, 2 );
    }
    else    /* B_*x* */
    {
        int b_list0[2];
        int b_list1[2];

        int i;

        /* init ref list utilisations */
        for( i = 0; i < 2; i++ )
        {
            b_list0[i] = x264_mb_type_list0_table[h->mb.i_type][i];
            b_list1[i] = x264_mb_type_list1_table[h->mb.i_type][i];
        }
        if( h->mb.i_partition == D_16x16 )
        {
            if( b_list0[0] && b_list1[0] ) x264_mb_mc_01xywh( h, 0, 0, 4, 4 );
            else if( b_list0[0] )          x264_mb_mc_0xywh ( h, 0, 0, 4, 4 );
            else if( b_list1[0] )          x264_mb_mc_1xywh ( h, 0, 0, 4, 4 );
        }
        else if( h->mb.i_partition == D_16x8 )
        {
            if( b_list0[0] && b_list1[0] ) x264_mb_mc_01xywh( h, 0, 0, 4, 2 );
            else if( b_list0[0] )          x264_mb_mc_0xywh ( h, 0, 0, 4, 2 );
            else if( b_list1[0] )          x264_mb_mc_1xywh ( h, 0, 0, 4, 2 );

            if( b_list0[1] && b_list1[1] ) x264_mb_mc_01xywh( h, 0, 2, 4, 2 );
            else if( b_list0[1] )          x264_mb_mc_0xywh ( h, 0, 2, 4, 2 );
            else if( b_list1[1] )          x264_mb_mc_1xywh ( h, 0, 2, 4, 2 );
        }
        else if( h->mb.i_partition == D_8x16 )
        {
            if( b_list0[0] && b_list1[0] ) x264_mb_mc_01xywh( h, 0, 0, 2, 4 );
            else if( b_list0[0] )          x264_mb_mc_0xywh ( h, 0, 0, 2, 4 );
            else if( b_list1[0] )          x264_mb_mc_1xywh ( h, 0, 0, 2, 4 );

            if( b_list0[1] && b_list1[1] ) x264_mb_mc_01xywh( h, 2, 0, 2, 4 );
            else if( b_list0[1] )          x264_mb_mc_0xywh ( h, 2, 0, 2, 4 );
            else if( b_list1[1] )          x264_mb_mc_1xywh ( h, 2, 0, 2, 4 );
        }
    }
}

int x264_macroblock_cache_init( x264_t *h )
{
    int i, j;
    int i_mb_count = h->mb.i_mb_count;

    h->mb.i_mb_stride = h->sps->i_mb_width;
    h->mb.i_b8_stride = h->sps->i_mb_width * 2;
    h->mb.i_b4_stride = h->sps->i_mb_width * 4;

    h->mb.b_interlaced = h->param.b_interlaced;

    CHECKED_MALLOC( h->mb.qp, i_mb_count * sizeof(int8_t) );
    CHECKED_MALLOC( h->mb.cbp, i_mb_count * sizeof(int16_t) );
    CHECKED_MALLOC( h->mb.skipbp, i_mb_count * sizeof(int8_t) );
    CHECKED_MALLOC( h->mb.mb_transform_size, i_mb_count * sizeof(int8_t) );

    /* 0 -> 3 top(4), 4 -> 6 : left(3) */
    CHECKED_MALLOC( h->mb.intra4x4_pred_mode, i_mb_count * 7 * sizeof(int8_t) );

    /* all coeffs */
    CHECKED_MALLOC( h->mb.non_zero_count, i_mb_count * 24 * sizeof(uint8_t) );
    CHECKED_MALLOC( h->mb.nnz_backup, h->sps->i_mb_width * 4 * 16 * sizeof(uint8_t) );

    if( h->param.b_cabac )
    {
        CHECKED_MALLOC( h->mb.chroma_pred_mode, i_mb_count * sizeof(int8_t) );
        CHECKED_MALLOC( h->mb.mvd[0], 2*16 * i_mb_count * sizeof(int16_t) );
        CHECKED_MALLOC( h->mb.mvd[1], 2*16 * i_mb_count * sizeof(int16_t) );
    }

    for( i=0; i<2; i++ )
    {
        int i_refs = X264_MIN(16, (i ? 1 : h->param.i_frame_reference) + h->param.b_bframe_pyramid) << h->param.b_interlaced;
        for( j=0; j < i_refs; j++ )
            CHECKED_MALLOC( h->mb.mvr[i][j], 2 * i_mb_count * sizeof(int16_t) );
    }

    for( i=0; i<=h->param.b_interlaced; i++ )
        for( j=0; j<3; j++ )
        {
            CHECKED_MALLOC( h->mb.intra_border_backup[i][j], h->fdec->i_stride[j] );
            h->mb.intra_border_backup[i][j] += 8;
        }

    /* init with not available (for top right idx=7,15) */
    memset( h->mb.cache.ref[0], -2, X264_SCAN8_SIZE * sizeof( int8_t ) );
    memset( h->mb.cache.ref[1], -2, X264_SCAN8_SIZE * sizeof( int8_t ) );

    return 0;
fail: return -1;
}
void x264_macroblock_cache_end( x264_t *h )
{
    int i, j;
    for( i=0; i<=h->param.b_interlaced; i++ )
        for( j=0; j<3; j++ )
            x264_free( h->mb.intra_border_backup[i][j] - 8 );
    for( i=0; i<2; i++ )
    {
        int i_refs = i ? 1 + h->param.b_bframe_pyramid : h->param.i_frame_reference;
        for( j=0; j < i_refs; j++ )
            x264_free( h->mb.mvr[i][j] );
    }
    if( h->param.b_cabac )
    {
        x264_free( h->mb.chroma_pred_mode );
        x264_free( h->mb.mvd[0] );
        x264_free( h->mb.mvd[1] );
    }
    x264_free( h->mb.intra4x4_pred_mode );
    x264_free( h->mb.non_zero_count );
    x264_free( h->mb.nnz_backup );
    x264_free( h->mb.mb_transform_size );
    x264_free( h->mb.skipbp );
    x264_free( h->mb.cbp );
    x264_free( h->mb.qp );
}
void x264_macroblock_slice_init( x264_t *h )
{
    int i, j;

    h->mb.mv[0] = h->fdec->mv[0];
    h->mb.mv[1] = h->fdec->mv[1];
    h->mb.ref[0] = h->fdec->ref[0];
    h->mb.ref[1] = h->fdec->ref[1];
    h->mb.type = h->fdec->mb_type;

    h->fdec->i_ref[0] = h->i_ref0;
    h->fdec->i_ref[1] = h->i_ref1;
    for( i = 0; i < h->i_ref0; i++ )
        h->fdec->ref_poc[0][i] = h->fref0[i]->i_poc;
    if( h->sh.i_type == SLICE_TYPE_B )
    {
        for( i = 0; i < h->i_ref1; i++ )
            h->fdec->ref_poc[1][i] = h->fref1[i]->i_poc;

        h->mb.map_col_to_list0[-1] = -1;
        h->mb.map_col_to_list0[-2] = -2;
        for( i = 0; i < h->fref1[0]->i_ref[0]; i++ )
        {
            int poc = h->fref1[0]->ref_poc[0][i];
            h->mb.map_col_to_list0[i] = -2;
            for( j = 0; j < h->i_ref0; j++ )
                if( h->fref0[j]->i_poc == poc )
                {
                    h->mb.map_col_to_list0[i] = j;
                    break;
                }
        }
    }
    if( h->sh.i_type == SLICE_TYPE_P )
        memset( h->mb.cache.skip, 0, X264_SCAN8_SIZE * sizeof( int8_t ) );
}

void x264_prefetch_fenc( x264_t *h, x264_frame_t *fenc, int i_mb_x, int i_mb_y )
{
    int stride_y  = fenc->i_stride[0];
    int stride_uv = fenc->i_stride[1];
    int off_y = 16 * (i_mb_x + i_mb_y * stride_y);
    int off_uv = 8 * (i_mb_x + i_mb_y * stride_uv);
    h->mc.prefetch_fenc( fenc->plane[0]+off_y, stride_y,
                         fenc->plane[1+(i_mb_x&1)]+off_uv, stride_uv, i_mb_x );
}

void x264_macroblock_cache_load( x264_t *h, int i_mb_x, int i_mb_y )
{
    int i_mb_xy = i_mb_y * h->mb.i_mb_stride + i_mb_x;
    int i_mb_4x4 = 4*(i_mb_y * h->mb.i_b4_stride + i_mb_x);
    int i_mb_8x8 = 2*(i_mb_y * h->mb.i_b8_stride + i_mb_x);
    int i_top_y = i_mb_y - (1 << h->mb.b_interlaced);
    int i_top_xy = i_top_y * h->mb.i_mb_stride + i_mb_x;
    int i_top_4x4 = (4*i_top_y+3) * h->mb.i_b4_stride + 4*i_mb_x;
    int i_top_8x8 = (2*i_top_y+1) * h->mb.i_b8_stride + 2*i_mb_x;
    int i_left_xy = -1;
    int i_top_type = -1;    /* gcc warn */
    int i_left_type= -1;

    int i;

    assert( h->mb.i_b8_stride == 2*h->mb.i_mb_stride );
    assert( h->mb.i_b4_stride == 4*h->mb.i_mb_stride );

    /* init index */
    h->mb.i_mb_x = i_mb_x;
    h->mb.i_mb_y = i_mb_y;
    h->mb.i_mb_xy = i_mb_xy;
    h->mb.i_b8_xy = i_mb_8x8;
    h->mb.i_b4_xy = i_mb_4x4;
    h->mb.i_mb_top_xy = i_top_xy;
    h->mb.i_neighbour = 0;

    /* load cache */
    if( i_top_xy >= h->sh.i_first_mb )
    {
        h->mb.i_mb_type_top =
        i_top_type= h->mb.type[i_top_xy];

        h->mb.i_neighbour |= MB_TOP;

        /* load intra4x4 */
        h->mb.cache.intra4x4_pred_mode[x264_scan8[0] - 8] = h->mb.intra4x4_pred_mode[i_top_xy][0];
        h->mb.cache.intra4x4_pred_mode[x264_scan8[1] - 8] = h->mb.intra4x4_pred_mode[i_top_xy][1];
        h->mb.cache.intra4x4_pred_mode[x264_scan8[4] - 8] = h->mb.intra4x4_pred_mode[i_top_xy][2];
        h->mb.cache.intra4x4_pred_mode[x264_scan8[5] - 8] = h->mb.intra4x4_pred_mode[i_top_xy][3];

        /* load non_zero_count */
        h->mb.cache.non_zero_count[x264_scan8[0] - 8] = h->mb.non_zero_count[i_top_xy][10];
        h->mb.cache.non_zero_count[x264_scan8[1] - 8] = h->mb.non_zero_count[i_top_xy][11];
        h->mb.cache.non_zero_count[x264_scan8[4] - 8] = h->mb.non_zero_count[i_top_xy][14];
        h->mb.cache.non_zero_count[x264_scan8[5] - 8] = h->mb.non_zero_count[i_top_xy][15];

        h->mb.cache.non_zero_count[x264_scan8[16+0] - 8] = h->mb.non_zero_count[i_top_xy][16+2];
        h->mb.cache.non_zero_count[x264_scan8[16+1] - 8] = h->mb.non_zero_count[i_top_xy][16+3];

        h->mb.cache.non_zero_count[x264_scan8[16+4+0] - 8] = h->mb.non_zero_count[i_top_xy][16+4+2];
        h->mb.cache.non_zero_count[x264_scan8[16+4+1] - 8] = h->mb.non_zero_count[i_top_xy][16+4+3];
    }
    else
    {
        h->mb.i_mb_type_top = -1;
        
        /* load intra4x4 */
        h->mb.cache.intra4x4_pred_mode[x264_scan8[0] - 8] =
        h->mb.cache.intra4x4_pred_mode[x264_scan8[1] - 8] =
        h->mb.cache.intra4x4_pred_mode[x264_scan8[4] - 8] =
        h->mb.cache.intra4x4_pred_mode[x264_scan8[5] - 8] = -1;

        /* load non_zero_count */
        h->mb.cache.non_zero_count[x264_scan8[0] - 8] =
        h->mb.cache.non_zero_count[x264_scan8[1] - 8] =
        h->mb.cache.non_zero_count[x264_scan8[4] - 8] =
        h->mb.cache.non_zero_count[x264_scan8[5] - 8] =
        h->mb.cache.non_zero_count[x264_scan8[16+0] - 8] =
        h->mb.cache.non_zero_count[x264_scan8[16+1] - 8] =
        h->mb.cache.non_zero_count[x264_scan8[16+4+0] - 8] =
        h->mb.cache.non_zero_count[x264_scan8[16+4+1] - 8] = 0x80;

    }

    if( i_mb_x > 0 && i_mb_xy > h->sh.i_first_mb )
    {
        i_left_xy = i_mb_xy - 1;
        h->mb.i_mb_type_left =
        i_left_type = h->mb.type[i_left_xy];

        h->mb.i_neighbour |= MB_LEFT;

        /* load intra4x4 */
        h->mb.cache.intra4x4_pred_mode[x264_scan8[0 ] - 1] = h->mb.intra4x4_pred_mode[i_left_xy][4];
        h->mb.cache.intra4x4_pred_mode[x264_scan8[2 ] - 1] = h->mb.intra4x4_pred_mode[i_left_xy][5];
        h->mb.cache.intra4x4_pred_mode[x264_scan8[8 ] - 1] = h->mb.intra4x4_pred_mode[i_left_xy][6];
        h->mb.cache.intra4x4_pred_mode[x264_scan8[10] - 1] = h->mb.intra4x4_pred_mode[i_left_xy][3];

        /* load non_zero_count */
        h->mb.cache.non_zero_count[x264_scan8[0 ] - 1] = h->mb.non_zero_count[i_left_xy][5];
        h->mb.cache.non_zero_count[x264_scan8[2 ] - 1] = h->mb.non_zero_count[i_left_xy][7];
        h->mb.cache.non_zero_count[x264_scan8[8 ] - 1] = h->mb.non_zero_count[i_left_xy][13];
        h->mb.cache.non_zero_count[x264_scan8[10] - 1] = h->mb.non_zero_count[i_left_xy][15];

        h->mb.cache.non_zero_count[x264_scan8[16+0] - 1] = h->mb.non_zero_count[i_left_xy][16+1];
        h->mb.cache.non_zero_count[x264_scan8[16+2] - 1] = h->mb.non_zero_count[i_left_xy][16+3];

        h->mb.cache.non_zero_count[x264_scan8[16+4+0] - 1] = h->mb.non_zero_count[i_left_xy][16+4+1];
        h->mb.cache.non_zero_count[x264_scan8[16+4+2] - 1] = h->mb.non_zero_count[i_left_xy][16+4+3];
    }
    else
    {
        h->mb.i_mb_type_left = -1;

        h->mb.cache.intra4x4_pred_mode[x264_scan8[0 ] - 1] =
        h->mb.cache.intra4x4_pred_mode[x264_scan8[2 ] - 1] =
        h->mb.cache.intra4x4_pred_mode[x264_scan8[8 ] - 1] =
        h->mb.cache.intra4x4_pred_mode[x264_scan8[10] - 1] = -1;

        /* load non_zero_count */
        h->mb.cache.non_zero_count[x264_scan8[0 ] - 1] =
        h->mb.cache.non_zero_count[x264_scan8[2 ] - 1] =
        h->mb.cache.non_zero_count[x264_scan8[8 ] - 1] =
        h->mb.cache.non_zero_count[x264_scan8[10] - 1] =
        h->mb.cache.non_zero_count[x264_scan8[16+0] - 1] =
        h->mb.cache.non_zero_count[x264_scan8[16+2] - 1] =
        h->mb.cache.non_zero_count[x264_scan8[16+4+0] - 1] =
        h->mb.cache.non_zero_count[x264_scan8[16+4+2] - 1] = 0x80;
    }

    if( i_mb_x < h->sps->i_mb_width - 1 && i_top_xy + 1 >= h->sh.i_first_mb )
    {
        h->mb.i_neighbour |= MB_TOPRIGHT;
        h->mb.i_mb_type_topright = h->mb.type[ i_top_xy + 1 ];
    }
    else
        h->mb.i_mb_type_topright = -1;
    if( i_mb_x > 0 && i_top_xy - 1 >= h->sh.i_first_mb )
    {
        h->mb.i_neighbour |= MB_TOPLEFT;
        h->mb.i_mb_type_topleft = h->mb.type[ i_top_xy - 1 ];
    }
    else
        h->mb.i_mb_type_topleft = -1;

    if( h->pps->b_transform_8x8_mode )
    {
        h->mb.cache.i_neighbour_transform_size =
            ( i_left_type >= 0 && h->mb.mb_transform_size[i_left_xy] )
          + ( i_top_type  >= 0 && h->mb.mb_transform_size[i_top_xy]  );
    }

    if( h->sh.b_mbaff )
    {
        h->mb.pic.i_fref[0] = h->i_ref0 << h->mb.b_interlaced;
        h->mb.pic.i_fref[1] = h->i_ref1 << h->mb.b_interlaced;
        h->mb.cache.i_neighbour_interlaced =
            !!(h->mb.i_neighbour & MB_LEFT)
          + !!(h->mb.i_neighbour & MB_TOP);
    }

    /* fdec:      fenc:
     * yyyyyyy
     * yYYYY      YYYY
     * yYYYY      YYYY
     * yYYYY      YYYY
     * yYYYY      YYYY
     * uuu vvv    UUVV
     * uUU vVV    UUVV
     * uUU vVV
     */
    h->mb.pic.p_fenc[0] = h->mb.pic.fenc_buf;
    h->mb.pic.p_fenc[1] = h->mb.pic.fenc_buf + 16*FENC_STRIDE;
    h->mb.pic.p_fenc[2] = h->mb.pic.fenc_buf + 16*FENC_STRIDE + 8;
    h->mb.pic.p_fdec[0] = h->mb.pic.fdec_buf + 2*FDEC_STRIDE;
    h->mb.pic.p_fdec[1] = h->mb.pic.fdec_buf + 19*FDEC_STRIDE;
    h->mb.pic.p_fdec[2] = h->mb.pic.fdec_buf + 19*FDEC_STRIDE + 16;

    /* load picture pointers */
    for( i = 0; i < 3; i++ )
    {
        const int w = (i == 0 ? 16 : 8);
        const int i_stride = h->fdec->i_stride[i];
        const int i_stride2 = i_stride << h->mb.b_interlaced;
        const int i_pix_offset = h->mb.b_interlaced
                               ? w * (i_mb_x + (i_mb_y&~1) * i_stride) + (i_mb_y&1) * i_stride
                               : w * (i_mb_x + i_mb_y * i_stride);
        int ref_pix_offset[2] = { i_pix_offset, i_pix_offset };
        const uint8_t *plane_fdec = &h->fdec->plane[i][i_pix_offset];
        const uint8_t *intra_fdec = &h->mb.intra_border_backup[i_mb_y & h->sh.b_mbaff][i][i_mb_x*16>>!!i];
        x264_frame_t **fref[2] = { h->fref0, h->fref1 };
        int j, k, l;

        if( h->mb.b_interlaced )
            ref_pix_offset[1] += (1-2*(i_mb_y&1)) * i_stride;

        h->mb.pic.i_stride[i] = i_stride2;

        h->mc.copy[i?PIXEL_8x8:PIXEL_16x16]( h->mb.pic.p_fenc[i], FENC_STRIDE,
            &h->fenc->plane[i][i_pix_offset], i_stride2, w );
        memcpy( &h->mb.pic.p_fdec[i][-1-FDEC_STRIDE], intra_fdec-1, w*3/2+1 );
        for( j = 0; j < w; j++ )
            h->mb.pic.p_fdec[i][-1+j*FDEC_STRIDE] = plane_fdec[-1+j*i_stride2];

        for( l=0; l<2; l++ )
        {
            for( j=0; j<h->mb.pic.i_fref[l]; j++ )
            {
                h->mb.pic.p_fref[l][j][i==0 ? 0:i+3] = &fref[l][j >> h->mb.b_interlaced]->plane[i][ref_pix_offset[j&1]];
                if( i == 0 )
                    for( k = 1; k < 4; k++ )
                        h->mb.pic.p_fref[l][j][k] = &fref[l][j >> h->mb.b_interlaced]->filtered[k][ref_pix_offset[j&1]];
            }
        }
    }

    if( h->fdec->integral )
    {
        assert( !h->mb.b_interlaced );
        for( i = 0; i < h->mb.pic.i_fref[0]; i++ )
            h->mb.pic.p_integral[0][i] = &h->fref0[i]->integral[ 16 * ( i_mb_x + i_mb_y * h->fdec->i_stride[0] )];
        for( i = 0; i < h->mb.pic.i_fref[1]; i++ )
            h->mb.pic.p_integral[1][i] = &h->fref1[i]->integral[ 16 * ( i_mb_x + i_mb_y * h->fdec->i_stride[0] )];
    }

    x264_prefetch_fenc( h, h->fenc, i_mb_x, i_mb_y );

    /* load ref/mv/mvd */
    if( h->sh.i_type != SLICE_TYPE_I )
    {
        const int s8x8 = h->mb.i_b8_stride;
        const int s4x4 = h->mb.i_b4_stride;

        int i_list;

        for( i_list = 0; i_list < (h->sh.i_type == SLICE_TYPE_B ? 2  : 1 ); i_list++ )