/*
 * Common mpeg video decoding code
 * Copyright (c) 2000,2001 Fabrice Bellard
 * Copyright (c) 2002-2004 Michael Niedermayer <michaelni@gmx.at>
 *
 * This file is part of FFmpeg.
 *
 * FFmpeg is free software; you can redistribute it and/or
 * modify it under the terms of the GNU Lesser General Public
 * License as published by the Free Software Foundation; either
 * version 2.1 of the License, or (at your option) any later version.
 *
 * FFmpeg is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
 * Lesser General Public License for more details.
 *
 * You should have received a copy of the GNU Lesser General Public
 * License along with FFmpeg; if not, write to the Free Software
 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
 */

#include <limits.h>

#include "config_components.h"

#include "libavutil/avassert.h"
#include "libavutil/emms.h"
#include "libavutil/imgutils.h"
#include "libavutil/internal.h"
#include "libavutil/video_enc_params.h"

#include "avcodec.h"
#include "h264chroma.h"
#include "internal.h"
#include "mpegutils.h"
#include "mpegvideo.h"
#include "mpegvideodec.h"
#include "mpeg4videodec.h"
#include "thread.h"
#include "threadframe.h"
#include "wmv2dec.h"

void ff_mpv_decode_init(MpegEncContext *s, AVCodecContext *avctx)
{
    ff_mpv_common_defaults(s);

    s->avctx           = avctx;
    s->width           = avctx->coded_width;
    s->height          = avctx->coded_height;
    s->codec_id        = avctx->codec->id;
    s->workaround_bugs = avctx->workaround_bugs;

    /* convert fourcc to upper case */
    s->codec_tag       = ff_toupper4(avctx->codec_tag);

    ff_h264chroma_init(&s->h264chroma, 8); //for lowres
}

int ff_mpeg_update_thread_context(AVCodecContext *dst,
                                  const AVCodecContext *src)
{
    MpegEncContext *const s1 = src->priv_data;
    MpegEncContext *const s  = dst->priv_data;
    int ret;

    if (dst == src)
        return 0;

    av_assert0(s != s1);

    // FIXME can parameters change on I-frames?
    // in that case dst may need a reinit
    if (!s->context_initialized) {
        void *private_ctx = s->private_ctx;
        int err;
        memcpy(s, s1, sizeof(*s));

        s->avctx                 = dst;
        s->private_ctx           = private_ctx;
        s->bitstream_buffer      = NULL;
        s->bitstream_buffer_size = s->allocated_bitstream_buffer_size = 0;

        if (s1->context_initialized) {
            ff_mpv_idct_init(s);
            if ((err = ff_mpv_common_init(s)) < 0) {
                memset(s, 0, sizeof(*s));
                s->avctx = dst;
                s->private_ctx = private_ctx;
                memcpy(&s->h264chroma, &s1->h264chroma, sizeof(s->h264chroma));
                return err;
            }
        }
    }

    if (s->height != s1->height || s->width != s1->width || s->context_reinit) {
        s->height = s1->height;
        s->width  = s1->width;
        if ((ret = ff_mpv_common_frame_size_change(s)) < 0)
            return ret;
    }

    s->quarter_sample       = s1->quarter_sample;

    s->coded_picture_number = s1->coded_picture_number;
    s->picture_number       = s1->picture_number;

    av_assert0(!s->picture || s->picture != s1->picture);
    if (s->picture)
        for (int i = 0; i < MAX_PICTURE_COUNT; i++) {
            ff_mpeg_unref_picture(s->avctx, &s->picture[i]);
            if (s1->picture && s1->picture[i].f->buf[0] &&
                (ret = ff_mpeg_ref_picture(s->avctx, &s->picture[i], &s1->picture[i])) < 0)
                return ret;
        }

#define UPDATE_PICTURE(pic)\
do {\
    ff_mpeg_unref_picture(s->avctx, &s->pic);\
    if (s1->pic.f && s1->pic.f->buf[0])\
        ret = ff_mpeg_ref_picture(s->avctx, &s->pic, &s1->pic);\
    else\
        ret = ff_update_picture_tables(&s->pic, &s1->pic);\
    if (ret < 0)\
        return ret;\
} while (0)

    UPDATE_PICTURE(current_picture);
    UPDATE_PICTURE(last_picture);
    UPDATE_PICTURE(next_picture);

#define REBASE_PICTURE(pic, new_ctx, old_ctx)                                 \
    ((pic && pic >= old_ctx->picture &&                                       \
      pic < old_ctx->picture + MAX_PICTURE_COUNT) ?                           \
        &new_ctx->picture[pic - old_ctx->picture] : NULL)

    s->last_picture_ptr    = REBASE_PICTURE(s1->last_picture_ptr,    s, s1);
    s->current_picture_ptr = REBASE_PICTURE(s1->current_picture_ptr, s, s1);
    s->next_picture_ptr    = REBASE_PICTURE(s1->next_picture_ptr,    s, s1);

    // Error/bug resilience
    s->workaround_bugs      = s1->workaround_bugs;
    s->padding_bug_score    = s1->padding_bug_score;

    // MPEG-4 timing info
    memcpy(&s->last_time_base, &s1->last_time_base,
           (char *) &s1->pb_field_time + sizeof(s1->pb_field_time) -
           (char *) &s1->last_time_base);

    // B-frame info
    s->max_b_frames = s1->max_b_frames;
    s->low_delay    = s1->low_delay;
    s->droppable    = s1->droppable;

    // DivX handling (doesn't work)
    s->divx_packed  = s1->divx_packed;

    if (s1->bitstream_buffer) {
        av_fast_padded_malloc(&s->bitstream_buffer,
                              &s->allocated_bitstream_buffer_size,
                              s1->bitstream_buffer_size);
        if (!s->bitstream_buffer) {
            s->bitstream_buffer_size = 0;
            return AVERROR(ENOMEM);
        }
        s->bitstream_buffer_size = s1->bitstream_buffer_size;
        memcpy(s->bitstream_buffer, s1->bitstream_buffer,
               s1->bitstream_buffer_size);
    }

    // linesize-dependent scratch buffer allocation
    if (!s->sc.edge_emu_buffer)
        if (s1->linesize) {
            if (ff_mpeg_framesize_alloc(s->avctx, &s->me,
                                        &s->sc, s1->linesize) < 0) {
                av_log(s->avctx, AV_LOG_ERROR, "Failed to allocate context "
                       "scratch buffers.\n");
                return AVERROR(ENOMEM);
            }
        } else {
            av_log(s->avctx, AV_LOG_ERROR, "Context scratch buffers could not "
                   "be allocated due to unknown size.\n");
        }

    // MPEG-2/interlacing info
    memcpy(&s->progressive_sequence, &s1->progressive_sequence,
           (char *) &s1->rtp_mode - (char *) &s1->progressive_sequence);

    return 0;
}

int ff_mpv_common_frame_size_change(MpegEncContext *s)
{
    int err = 0;

    if (!s->context_initialized)
        return AVERROR(EINVAL);

    ff_mpv_free_context_frame(s);

    if (s->picture)
        for (int i = 0; i < MAX_PICTURE_COUNT; i++)
            s->picture[i].needs_realloc = 1;

    s->last_picture_ptr         =
    s->next_picture_ptr         =
    s->current_picture_ptr      = NULL;

    if ((s->width || s->height) &&
        (err = av_image_check_size(s->width, s->height, 0, s->avctx)) < 0)
        goto fail;

    /* set chroma shifts */
    err = av_pix_fmt_get_chroma_sub_sample(s->avctx->pix_fmt,
                                           &s->chroma_x_shift,
                                           &s->chroma_y_shift);
    if (err < 0)
        goto fail;

    if ((err = ff_mpv_init_context_frame(s)))
        goto fail;

    memset(s->thread_context, 0, sizeof(s->thread_context));
    s->thread_context[0]   = s;

    if (s->width && s->height) {
        err = ff_mpv_init_duplicate_contexts(s);
        if (err < 0)
            goto fail;
    }
    s->context_reinit = 0;

    return 0;
 fail:
    ff_mpv_free_context_frame(s);
    s->context_reinit = 1;
    return err;
}

static int alloc_picture(MpegEncContext *s, Picture *pic)
{
    return ff_alloc_picture(s->avctx, pic, &s->me, &s->sc, 0, 0,
                            s->chroma_x_shift, s->chroma_y_shift, s->out_format,
                            s->mb_stride, s->mb_width, s->mb_height, s->b8_stride,
                            &s->linesize, &s->uvlinesize);
}

static void color_frame(AVFrame *frame, int luma)
{
    int h_chroma_shift, v_chroma_shift;

    for (int i = 0; i < frame->height; i++)
        memset(frame->data[0] + frame->linesize[0] * i, luma, frame->width);

    if (!frame->data[1])
        return;
    av_pix_fmt_get_chroma_sub_sample(frame->format, &h_chroma_shift, &v_chroma_shift);
    for (int i = 0; i < AV_CEIL_RSHIFT(frame->height, v_chroma_shift); i++) {
        memset(frame->data[1] + frame->linesize[1] * i,
               0x80, AV_CEIL_RSHIFT(frame->width, h_chroma_shift));
        memset(frame->data[2] + frame->linesize[2] * i,
               0x80, AV_CEIL_RSHIFT(frame->width, h_chroma_shift));
    }
}

/**
 * generic function called after decoding
 * the header and before a frame is decoded.
 */
int ff_mpv_frame_start(MpegEncContext *s, AVCodecContext *avctx)
{
    Picture *pic;
    int idx, ret;

    s->mb_skipped = 0;

    if (!ff_thread_can_start_frame(avctx)) {
        av_log(avctx, AV_LOG_ERROR, "Attempt to start a frame outside SETUP state\n");
        return -1;
    }

    /* mark & release old frames */
    if (s->pict_type != AV_PICTURE_TYPE_B && s->last_picture_ptr &&
        s->last_picture_ptr != s->next_picture_ptr &&
        s->last_picture_ptr->f->buf[0]) {
        ff_mpeg_unref_picture(s->avctx, s->last_picture_ptr);
    }

    /* release non reference/forgotten frames */
    for (int i = 0; i < MAX_PICTURE_COUNT; i++) {
        if (!s->picture[i].reference ||
            (&s->picture[i] != s->last_picture_ptr &&
             &s->picture[i] != s->next_picture_ptr &&
             !s->picture[i].needs_realloc)) {
            ff_mpeg_unref_picture(s->avctx, &s->picture[i]);
        }
    }

    ff_mpeg_unref_picture(s->avctx, &s->current_picture);
    ff_mpeg_unref_picture(s->avctx, &s->last_picture);
    ff_mpeg_unref_picture(s->avctx, &s->next_picture);

    if (s->current_picture_ptr && !s->current_picture_ptr->f->buf[0]) {
        // we already have an unused image
        // (maybe it was set before reading the header)
        pic = s->current_picture_ptr;
    } else {
        idx = ff_find_unused_picture(s->avctx, s->picture, 0);
        if (idx < 0) {
            av_log(s->avctx, AV_LOG_ERROR, "no frame buffer available\n");
            return idx;
        }
        pic = &s->picture[idx];
    }

    pic->reference = 0;
    if (!s->droppable) {
        if (s->pict_type != AV_PICTURE_TYPE_B)
            pic->reference = 3;
    }

#if FF_API_FRAME_PICTURE_NUMBER
FF_DISABLE_DEPRECATION_WARNINGS
    pic->f->coded_picture_number = s->coded_picture_number++;
FF_ENABLE_DEPRECATION_WARNINGS
#endif

    if (alloc_picture(s, pic) < 0)
        return -1;

    s->current_picture_ptr = pic;
    // FIXME use only the vars from current_pic
    s->current_picture_ptr->f->flags |= AV_FRAME_FLAG_TOP_FIELD_FIRST * !!s->top_field_first;
    if (s->codec_id == AV_CODEC_ID_MPEG1VIDEO ||
        s->codec_id == AV_CODEC_ID_MPEG2VIDEO) {
        if (s->picture_structure != PICT_FRAME)
            s->current_picture_ptr->f->flags |= AV_FRAME_FLAG_TOP_FIELD_FIRST *
                ((s->picture_structure == PICT_TOP_FIELD) == s->first_field);
    }
    s->current_picture_ptr->f->flags |= AV_FRAME_FLAG_INTERLACED * (!s->progressive_frame &&
                                                                    !s->progressive_sequence);
    s->current_picture_ptr->field_picture      =  s->picture_structure != PICT_FRAME;

    s->current_picture_ptr->f->pict_type = s->pict_type;
    if (s->pict_type == AV_PICTURE_TYPE_I)
        s->current_picture_ptr->f->flags |= AV_FRAME_FLAG_KEY;
    else
        s->current_picture_ptr->f->flags &= ~AV_FRAME_FLAG_KEY;

    if ((ret = ff_mpeg_ref_picture(s->avctx, &s->current_picture,
                                   s->current_picture_ptr)) < 0)
        return ret;

    if (s->pict_type != AV_PICTURE_TYPE_B) {
        s->last_picture_ptr = s->next_picture_ptr;
        if (!s->droppable)
            s->next_picture_ptr = s->current_picture_ptr;
    }
    ff_dlog(s->avctx, "L%p N%p C%p L%p N%p C%p type:%d drop:%d\n",
            s->last_picture_ptr, s->next_picture_ptr,s->current_picture_ptr,
            s->last_picture_ptr    ? s->last_picture_ptr->f->data[0]    : NULL,
            s->next_picture_ptr    ? s->next_picture_ptr->f->data[0]    : NULL,
            s->current_picture_ptr ? s->current_picture_ptr->f->data[0] : NULL,
            s->pict_type, s->droppable);

    if ((!s->last_picture_ptr || !s->last_picture_ptr->f->buf[0]) &&
        (s->pict_type != AV_PICTURE_TYPE_I)) {
        if (s->pict_type == AV_PICTURE_TYPE_B && s->next_picture_ptr && s->next_picture_ptr->f->buf[0])
            av_log(avctx, AV_LOG_DEBUG,
                   "allocating dummy last picture for B frame\n");
        else if (s->pict_type != AV_PICTURE_TYPE_I)
            av_log(avctx, AV_LOG_ERROR,
                   "warning: first frame is no keyframe\n");

        /* Allocate a dummy frame */
        idx = ff_find_unused_picture(s->avctx, s->picture, 0);
        if (idx < 0) {
            av_log(s->avctx, AV_LOG_ERROR, "no frame buffer available\n");
            return idx;
        }
        s->last_picture_ptr = &s->picture[idx];

        s->last_picture_ptr->reference    = 3;
        s->last_picture_ptr->f->flags &= ~AV_FRAME_FLAG_KEY;
        s->last_picture_ptr->f->pict_type = AV_PICTURE_TYPE_P;

        if (alloc_picture(s, s->last_picture_ptr) < 0) {
            s->last_picture_ptr = NULL;
            return -1;
        }

        if (!avctx->hwaccel) {
            int luma_val = s->codec_id == AV_CODEC_ID_FLV1 || s->codec_id == AV_CODEC_ID_H263 ? 16 : 0x80;
            color_frame(s->last_picture_ptr->f, luma_val);
        }

        ff_thread_report_progress(&s->last_picture_ptr->tf, INT_MAX, 0);
        ff_thread_report_progress(&s->last_picture_ptr->tf, INT_MAX, 1);
    }
    if ((!s->next_picture_ptr || !s->next_picture_ptr->f->buf[0]) &&
        s->pict_type == AV_PICTURE_TYPE_B) {
        /* Allocate a dummy frame */
        idx = ff_find_unused_picture(s->avctx, s->picture, 0);
        if (idx < 0) {
            av_log(s->avctx, AV_LOG_ERROR, "no frame buffer available\n");
            return idx;
        }
        s->next_picture_ptr = &s->picture[idx];

        s->next_picture_ptr->reference   = 3;
        s->next_picture_ptr->f->flags &= ~AV_FRAME_FLAG_KEY;
        s->next_picture_ptr->f->pict_type = AV_PICTURE_TYPE_P;

        if (alloc_picture(s, s->next_picture_ptr) < 0) {
            s->next_picture_ptr = NULL;
            return -1;
        }
        ff_thread_report_progress(&s->next_picture_ptr->tf, INT_MAX, 0);
        ff_thread_report_progress(&s->next_picture_ptr->tf, INT_MAX, 1);
    }

    if (s->last_picture_ptr) {
        if (s->last_picture_ptr->f->buf[0] &&
            (ret = ff_mpeg_ref_picture(s->avctx, &s->last_picture,
                                       s->last_picture_ptr)) < 0)
            return ret;
    }
    if (s->next_picture_ptr) {
        if (s->next_picture_ptr->f->buf[0] &&
            (ret = ff_mpeg_ref_picture(s->avctx, &s->next_picture,
                                       s->next_picture_ptr)) < 0)
            return ret;
    }

    av_assert0(s->pict_type == AV_PICTURE_TYPE_I || (s->last_picture_ptr &&
                                                 s->last_picture_ptr->f->buf[0]));

    if (s->picture_structure != PICT_FRAME) {
        for (int i = 0; i < 4; i++) {
            if (s->picture_structure == PICT_BOTTOM_FIELD) {
                s->current_picture.f->data[i] = FF_PTR_ADD(s->current_picture.f->data[i],
                                                           s->current_picture.f->linesize[i]);
            }
            s->current_picture.f->linesize[i] *= 2;
            s->last_picture.f->linesize[i]    *= 2;
            s->next_picture.f->linesize[i]    *= 2;
        }
    }

    /* set dequantizer, we can't do it during init as
     * it might change for MPEG-4 and we can't do it in the header
     * decode as init is not called for MPEG-4 there yet */
    if (s->mpeg_quant || s->codec_id == AV_CODEC_ID_MPEG2VIDEO) {
        s->dct_unquantize_intra = s->dct_unquantize_mpeg2_intra;
        s->dct_unquantize_inter = s->dct_unquantize_mpeg2_inter;
    } else if (s->out_format == FMT_H263 || s->out_format == FMT_H261) {
        s->dct_unquantize_intra = s->dct_unquantize_h263_intra;
        s->dct_unquantize_inter = s->dct_unquantize_h263_inter;
    } else {
        s->dct_unquantize_intra = s->dct_unquantize_mpeg1_intra;
        s->dct_unquantize_inter = s->dct_unquantize_mpeg1_inter;
    }

    if (s->avctx->debug & FF_DEBUG_NOMC)
        color_frame(s->current_picture_ptr->f, 0x80);

    return 0;
}

/* called after a frame has been decoded. */
void ff_mpv_frame_end(MpegEncContext *s)
{
    emms_c();

    if (s->current_picture.reference)
        ff_thread_report_progress(&s->current_picture_ptr->tf, INT_MAX, 0);
}

void ff_print_debug_info(const MpegEncContext *s, const Picture *p, AVFrame *pict)
{
    ff_print_debug_info2(s->avctx, pict, s->mbskip_table, p->mb_type,
                         p->qscale_table, p->motion_val,
                         s->mb_width, s->mb_height, s->mb_stride, s->quarter_sample);
}

int ff_mpv_export_qp_table(const MpegEncContext *s, AVFrame *f, const Picture *p, int qp_type)
{
    AVVideoEncParams *par;
    int mult = (qp_type == FF_MPV_QSCALE_TYPE_MPEG1) ? 2 : 1;
    unsigned int nb_mb = p->alloc_mb_height * p->alloc_mb_width;

    if (!(s->avctx->export_side_data & AV_CODEC_EXPORT_DATA_VIDEO_ENC_PARAMS))
        return 0;

    par = av_video_enc_params_create_side_data(f, AV_VIDEO_ENC_PARAMS_MPEG2, nb_mb);
    if (!par)
        return AVERROR(ENOMEM);

    for (unsigned y = 0; y < p->alloc_mb_height; y++)
        for (unsigned x = 0; x < p->alloc_mb_width; x++) {
            const unsigned int block_idx = y * p->alloc_mb_width + x;
            const unsigned int     mb_xy = y * p->alloc_mb_stride + x;
            AVVideoBlockParams *const b = av_video_enc_params_block(par, block_idx);

            b->src_x = x * 16;
            b->src_y = y * 16;
            b->w     = 16;
            b->h     = 16;

            b->delta_qp = p->qscale_table[mb_xy] * mult;
        }

    return 0;
}

void ff_mpeg_draw_horiz_band(MpegEncContext *s, int y, int h)
{
    ff_draw_horiz_band(s->avctx, s->current_picture_ptr->f,
                       s->last_picture_ptr ? s->last_picture_ptr->f : NULL,
                       y, h, s->picture_structure,
                       s->first_field, s->low_delay);
}

void ff_mpeg_flush(AVCodecContext *avctx)
{
    MpegEncContext *const s = avctx->priv_data;

    if (!s->picture)
        return;

    for (int i = 0; i < MAX_PICTURE_COUNT; i++)
        ff_mpeg_unref_picture(s->avctx, &s->picture[i]);
    s->current_picture_ptr = s->last_picture_ptr = s->next_picture_ptr = NULL;

    ff_mpeg_unref_picture(s->avctx, &s->current_picture);
    ff_mpeg_unref_picture(s->avctx, &s->last_picture);
    ff_mpeg_unref_picture(s->avctx, &s->next_picture);

    s->mb_x = s->mb_y = 0;

    s->bitstream_buffer_size = 0;
    s->pp_time = 0;
}

void ff_mpv_report_decode_progress(MpegEncContext *s)
{
    if (s->pict_type != AV_PICTURE_TYPE_B && !s->partitioned_frame && !s->er.error_occurred)
        ff_thread_report_progress(&s->current_picture_ptr->tf, s->mb_y, 0);
}


static inline int hpel_motion_lowres(MpegEncContext *s,
                                     uint8_t *dest, const uint8_t *src,
                                     int field_based, int field_select,
                                     int src_x, int src_y,
                                     int width, int height, ptrdiff_t stride,
                                     int h_edge_pos, int v_edge_pos,
                                     int w, int h, const h264_chroma_mc_func *pix_op,
                                     int motion_x, int motion_y)
{
    const int lowres   = s->avctx->lowres;
    const int op_index = FFMIN(lowres, 3);
    const int s_mask   = (2 << lowres) - 1;
    int emu = 0;
    int sx, sy;

    if (s->quarter_sample) {
        motion_x /= 2;
        motion_y /= 2;
    }

    sx = motion_x & s_mask;
    sy = motion_y & s_mask;
    src_x += motion_x >> lowres + 1;
    src_y += motion_y >> lowres + 1;

    src   += src_y * stride + src_x;

    if ((unsigned)src_x > FFMAX( h_edge_pos - (!!sx) - w,                 0) ||
        (unsigned)src_y > FFMAX((v_edge_pos >> field_based) - (!!sy) - h, 0)) {
        s->vdsp.emulated_edge_mc(s->sc.edge_emu_buffer, src,
                                 s->linesize, s->linesize,
                                 w + 1, (h + 1) << field_based,
                                 src_x, src_y * (1 << field_based),
                                 h_edge_pos, v_edge_pos);
        src = s->sc.edge_emu_buffer;
        emu = 1;
    }

    sx = (sx << 2) >> lowres;
    sy = (sy << 2) >> lowres;
    if (field_select)
        src += s->linesize;
    pix_op[op_index](dest, src, stride, h, sx, sy);
    return emu;
}

/* apply one mpeg motion vector to the three components */
static av_always_inline void mpeg_motion_lowres(MpegEncContext *s,
                                                uint8_t *dest_y,
                                                uint8_t *dest_cb,
                                                uint8_t *dest_cr,
                                                int field_based,
                                                int bottom_field,
                                                int field_select,
                                                uint8_t *const *ref_picture,
                                                const h264_chroma_mc_func *pix_op,
                                                int motion_x, int motion_y,
                                                int h, int mb_y)
{
    const uint8_t *ptr_y, *ptr_cb, *ptr_cr;
    int mx, my, src_x, src_y, uvsrc_x, uvsrc_y, sx, sy, uvsx, uvsy;
    ptrdiff_t uvlinesize, linesize;
    const int lowres     = s->avctx->lowres;
    const int op_index   = FFMIN(lowres - 1 + s->chroma_x_shift, 3);
    const int block_s    = 8 >> lowres;
    const int s_mask     = (2 << lowres) - 1;
    const int h_edge_pos = s->h_edge_pos >> lowres;
    const int v_edge_pos = s->v_edge_pos >> lowres;
    int hc = s->chroma_y_shift ? (h+1-bottom_field)>>1 : h;
    linesize   = s->current_picture.f->linesize[0] << field_based;
    uvlinesize = s->current_picture.f->linesize[1] << field_based;

    // FIXME obviously not perfect but qpel will not work in lowres anyway
    if (s->quarter_sample) {
        motion_x /= 2;
        motion_y /= 2;
    }

    if (field_based) {
        motion_y += (bottom_field - field_select)*((1 << lowres)-1);
    }

    sx = motion_x & s_mask;
    sy = motion_y & s_mask;
    src_x = s->mb_x * 2 * block_s + (motion_x >> lowres + 1);
    src_y = (mb_y * 2 * block_s >> field_based) + (motion_y >> lowres + 1);

    if (s->out_format == FMT_H263) {
        uvsx    = ((motion_x >> 1) & s_mask) | (sx & 1);
        uvsy    = ((motion_y >> 1) & s_mask) | (sy & 1);
        uvsrc_x = src_x >> 1;
        uvsrc_y = src_y >> 1;
    } else if (s->out_format == FMT_H261) {
        // even chroma mv's are full pel in H261
        mx      = motion_x / 4;
        my      = motion_y / 4;
        uvsx    = (2 * mx) & s_mask;
        uvsy    = (2 * my) & s_mask;
        uvsrc_x = s->mb_x * block_s + (mx >> lowres);
        uvsrc_y =    mb_y * block_s + (my >> lowres);
    } else {
        if (s->chroma_y_shift) {
            mx      = motion_x / 2;
            my      = motion_y / 2;
            uvsx    = mx & s_mask;
            uvsy    = my & s_mask;
            uvsrc_x = s->mb_x * block_s                 + (mx >> lowres + 1);
            uvsrc_y =   (mb_y * block_s >> field_based) + (my >> lowres + 1);
        } else {
            if (s->chroma_x_shift) {
            //Chroma422
                mx = motion_x / 2;
                uvsx = mx & s_mask;
                uvsy = motion_y & s_mask;
                uvsrc_y = src_y;
                uvsrc_x = s->mb_x*block_s               + (mx >> (lowres+1));
            } else {
            //Chroma444
                uvsx = motion_x & s_mask;
                uvsy = motion_y & s_mask;
                uvsrc_x = src_x;
                uvsrc_y = src_y;
            }
        }
    }

    ptr_y  = ref_picture[0] + src_y   * linesize   + src_x;
    ptr_cb = ref_picture[1] + uvsrc_y * uvlinesize + uvsrc_x;
    ptr_cr = ref_picture[2] + uvsrc_y * uvlinesize + uvsrc_x;

    if ((unsigned) src_x > FFMAX( h_edge_pos - (!!sx) - 2 * block_s,       0) || uvsrc_y<0 ||
        (unsigned) src_y > FFMAX((v_edge_pos >> field_based) - (!!sy) - FFMAX(h, hc<<s->chroma_y_shift), 0)) {
        s->vdsp.emulated_edge_mc(s->sc.edge_emu_buffer, ptr_y,
                                 linesize >> field_based, linesize >> field_based,
                                 17, 17 + field_based,
                                src_x, src_y * (1 << field_based), h_edge_pos,
                                v_edge_pos);
        ptr_y = s->sc.edge_emu_buffer;
        if (!CONFIG_GRAY || !(s->avctx->flags & AV_CODEC_FLAG_GRAY)) {
            uint8_t *ubuf = s->sc.edge_emu_buffer + 18 * s->linesize;
            uint8_t *vbuf =ubuf + 10 * s->uvlinesize;
            if (s->workaround_bugs & FF_BUG_IEDGE)
                vbuf -= s->uvlinesize;
            s->vdsp.emulated_edge_mc(ubuf,  ptr_cb,
                                     uvlinesize >> field_based, uvlinesize >> field_based,
                                     9, 9 + field_based,
                                    uvsrc_x, uvsrc_y * (1 << field_based),
                                    h_edge_pos >> 1, v_edge_pos >> 1);
            s->vdsp.emulated_edge_mc(vbuf,  ptr_cr,
                                     uvlinesize >> field_based,uvlinesize >> field_based,
                                     9, 9 + field_based,
                                    uvsrc_x, uvsrc_y * (1 << field_based),
                                    h_edge_pos >> 1, v_edge_pos >> 1);
            ptr_cb = ubuf;
            ptr_cr = vbuf;
        }
    }

    // FIXME use this for field pix too instead of the obnoxious hack which changes picture.f->data
    if (bottom_field) {
        dest_y  += s->linesize;
        dest_cb += s->uvlinesize;
        dest_cr += s->uvlinesize;
    }

    if (field_select) {
        ptr_y   += s->linesize;
        ptr_cb  += s->uvlinesize;
        ptr_cr  += s->uvlinesize;
    }

    sx = (sx << 2) >> lowres;
    sy = (sy << 2) >> lowres;
    pix_op[lowres - 1](dest_y, ptr_y, linesize, h, sx, sy);

    if (!CONFIG_GRAY || !(s->avctx->flags & AV_CODEC_FLAG_GRAY)) {
        uvsx = (uvsx << 2) >> lowres;
        uvsy = (uvsy << 2) >> lowres;
        if (hc) {
            pix_op[op_index](dest_cb, ptr_cb, uvlinesize, hc, uvsx, uvsy);
            pix_op[op_index](dest_cr, ptr_cr, uvlinesize, hc, uvsx, uvsy);
        }
    }
    // FIXME h261 lowres loop filter
}

static inline void chroma_4mv_motion_lowres(MpegEncContext *s,
                                            uint8_t *dest_cb, uint8_t *dest_cr,
                                            uint8_t *const *ref_picture,
                                            const h264_chroma_mc_func * pix_op,
                                            int mx, int my)
{
    const int lowres     = s->avctx->lowres;
    const int op_index   = FFMIN(lowres, 3);
    const int block_s    = 8 >> lowres;
    const int s_mask     = (2 << lowres) - 1;
    const int h_edge_pos = s->h_edge_pos >> lowres + 1;
    const int v_edge_pos = s->v_edge_pos >> lowres + 1;
    int emu = 0, src_x, src_y, sx, sy;
    ptrdiff_t offset;
    const uint8_t *ptr;

    if (s->quarter_sample) {
        mx /= 2;
        my /= 2;
    }

    /* In case of 8X8, we construct a single chroma motion vector
       with a special rounding */
    mx = ff_h263_round_chroma(mx);
    my = ff_h263_round_chroma(my);

    sx = mx & s_mask;
    sy = my & s_mask;
    src_x = s->mb_x * block_s + (mx >> lowres + 1);
    src_y = s->mb_y * block_s + (my >> lowres + 1);

    offset = src_y * s->uvlinesize + src_x;
    ptr = ref_picture[1] + offset;
    if ((unsigned) src_x > FFMAX(h_edge_pos - (!!sx) - block_s, 0) ||
        (unsigned) src_y > FFMAX(v_edge_pos - (!!sy) - block_s, 0)) {
        s->vdsp.emulated_edge_mc(s->sc.edge_emu_buffer, ptr,
                                 s->uvlinesize, s->uvlinesize,
                                 9, 9,
                                 src_x, src_y, h_edge_pos, v_edge_pos);
        ptr = s->sc.edge_emu_buffer;
        emu = 1;
    }
    sx = (sx << 2) >> lowres;
    sy = (sy << 2) >> lowres;
    pix_op[op_index](dest_cb, ptr, s->uvlinesize, block_s, sx, sy);

    ptr = ref_picture[2] + offset;
    if (emu) {
        s->vdsp.emulated_edge_mc(s->sc.edge_emu_buffer, ptr,
                                 s->uvlinesize, s->uvlinesize,
                                 9, 9,
                                 src_x, src_y, h_edge_pos, v_edge_pos);
        ptr = s->sc.edge_emu_buffer;
    }
    pix_op[op_index](dest_cr, ptr, s->uvlinesize, block_s, sx, sy);
}

/**
 * motion compensation of a single macroblock
 * @param s context
 * @param dest_y luma destination pointer
 * @param dest_cb chroma cb/u destination pointer
 * @param dest_cr chroma cr/v destination pointer
 * @param dir direction (0->forward, 1->backward)
 * @param ref_picture array[3] of pointers to the 3 planes of the reference picture
 * @param pix_op halfpel motion compensation function (average or put normally)
 * the motion vectors are taken from s->mv and the MV type from s->mv_type
 */
static inline void MPV_motion_lowres(MpegEncContext *s,
                                     uint8_t *dest_y, uint8_t *dest_cb,
                                     uint8_t *dest_cr,
                                     int dir, uint8_t *const *ref_picture,
                                     const h264_chroma_mc_func *pix_op)
{
    int mx, my;
    int mb_x, mb_y;
    const int lowres  = s->avctx->lowres;
    const int block_s = 8 >>lowres;

    mb_x = s->mb_x;
    mb_y = s->mb_y;

    switch (s->mv_type) {
    case MV_TYPE_16X16:
        mpeg_motion_lowres(s, dest_y, dest_cb, dest_cr,
                           0, 0, 0,
                           ref_picture, pix_op,
                           s->mv[dir][0][0], s->mv[dir][0][1],
                           2 * block_s, mb_y);
        break;
    case MV_TYPE_8X8:
        mx = 0;
        my = 0;
        for (int i = 0; i < 4; i++) {
            hpel_motion_lowres(s, dest_y + ((i & 1) + (i >> 1) *
                               s->linesize) * block_s,
                               ref_picture[0], 0, 0,
                               (2 * mb_x + (i & 1)) * block_s,
                               (2 * mb_y + (i >> 1)) * block_s,
                               s->width, s->height, s->linesize,
                               s->h_edge_pos >> lowres, s->v_edge_pos >> lowres,
                               block_s, block_s, pix_op,
                               s->mv[dir][i][0], s->mv[dir][i][1]);

            mx += s->mv[dir][i][0];
            my += s->mv[dir][i][1];
        }

        if (!CONFIG_GRAY || !(s->avctx->flags & AV_CODEC_FLAG_GRAY))
            chroma_4mv_motion_lowres(s, dest_cb, dest_cr, ref_picture,
                                     pix_op, mx, my);
        break;
    case MV_TYPE_FIELD:
        if (s->picture_structure == PICT_FRAME) {
            /* top field */
            mpeg_motion_lowres(s, dest_y, dest_cb, dest_cr,
                               1, 0, s->field_select[dir][0],
                               ref_picture, pix_op,
                               s->mv[dir][0][0], s->mv[dir][0][1],
                               block_s, mb_y);
            /* bottom field */
            mpeg_motion_lowres(s, dest_y, dest_cb, dest_cr,
                               1, 1, s->field_select[dir][1],
                               ref_picture, pix_op,
                               s->mv[dir][1][0], s->mv[dir][1][1],
                               block_s, mb_y);
        } else {
            if (   s->picture_structure != s->field_select[dir][0] + 1 && s->pict_type != AV_PICTURE_TYPE_B && !s->first_field
                || !ref_picture[0]) {
                ref_picture = s->current_picture_ptr->f->data;
            }
            mpeg_motion_lowres(s, dest_y, dest_cb, dest_cr,
                               0, 0, s->field_select[dir][0],
                               ref_picture, pix_op,
                               s->mv[dir][0][0],
                               s->mv[dir][0][1], 2 * block_s, mb_y >> 1);
            }
        break;
    case MV_TYPE_16X8:
        for (int i = 0; i < 2; i++) {
            uint8_t *const *ref2picture;

            if ((s->picture_structure == s->field_select[dir][i] + 1 ||
                 s->pict_type == AV_PICTURE_TYPE_B || s->first_field) &&
                ref_picture[0]) {
                ref2picture = ref_picture;
            } else {
                ref2picture = s->current_picture_ptr->f->data;
            }

            mpeg_motion_lowres(s, dest_y, dest_cb, dest_cr,
                               0, 0, s->field_select[dir][i],
                               ref2picture, pix_op,
                               s->mv[dir][i][0], s->mv[dir][i][1] +
                               2 * block_s * i, block_s, mb_y >> 1);

            dest_y  +=  2 * block_s *  s->linesize;
            dest_cb += (2 * block_s >> s->chroma_y_shift) * s->uvlinesize;
            dest_cr += (2 * block_s >> s->chroma_y_shift) * s->uvlinesize;
        }
        break;
    case MV_TYPE_DMV:
        if (s->picture_structure == PICT_FRAME) {
            for (int i = 0; i < 2; i++) {
                for (int j = 0; j < 2; j++) {
                    mpeg_motion_lowres(s, dest_y, dest_cb, dest_cr,
                                       1, j, j ^ i,
                                       ref_picture, pix_op,
                                       s->mv[dir][2 * i + j][0],
                                       s->mv[dir][2 * i + j][1],
                                       block_s, mb_y);
                }
                pix_op = s->h264chroma.avg_h264_chroma_pixels_tab;
            }
        } else {
            if (!ref_picture[0]) {
                ref_picture = s->current_picture_ptr->f->data;
            }
            for (int i = 0; i < 2; i++) {
                mpeg_motion_lowres(s, dest_y, dest_cb, dest_cr,
                                   0, 0, s->picture_structure != i + 1,
                                   ref_picture, pix_op,
                                   s->mv[dir][2 * i][0],s->mv[dir][2 * i][1],
                                   2 * block_s, mb_y >> 1);

                // after put we make avg of the same block
                pix_op = s->h264chroma.avg_h264_chroma_pixels_tab;

                // opposite parity is always in the same
                // frame if this is second field
                if (!s->first_field) {
                    ref_picture = s->current_picture_ptr->f->data;
                }
            }
        }
        break;
    default:
        av_assert2(0);
    }
}

/**
 * find the lowest MB row referenced in the MVs
 */
static int lowest_referenced_row(MpegEncContext *s, int dir)
{
    int my_max = INT_MIN, my_min = INT_MAX, qpel_shift = !s->quarter_sample;
    int off, mvs;

    if (s->picture_structure != PICT_FRAME || s->mcsel)
        goto unhandled;

    switch (s->mv_type) {
        case MV_TYPE_16X16:
            mvs = 1;
            break;
        case MV_TYPE_16X8:
            mvs = 2;
            break;
        case MV_TYPE_8X8:
            mvs = 4;
            break;
        default:
            goto unhandled;
    }

    for (int i = 0; i < mvs; i++) {
        int my = s->mv[dir][i][1];
        my_max = FFMAX(my_max, my);
        my_min = FFMIN(my_min, my);
    }

    off = ((FFMAX(-my_min, my_max) << qpel_shift) + 63) >> 6;

    return av_clip(s->mb_y + off, 0, s->mb_height - 1);
unhandled:
    return s->mb_height - 1;
}

/* add block[] to dest[] */
static inline void add_dct(MpegEncContext *s,
                           int16_t *block, int i, uint8_t *dest, int line_size)
{
    if (s->block_last_index[i] >= 0) {
        s->idsp.idct_add(dest, line_size, block);
    }
}

#define IS_ENCODER 0
#include "mpv_reconstruct_mb_template.c"

void ff_mpv_reconstruct_mb(MpegEncContext *s, int16_t block[12][64])
{
    if (s->avctx->debug & FF_DEBUG_DCT_COEFF) {
       /* print DCT coefficients */
       av_log(s->avctx, AV_LOG_DEBUG, "DCT coeffs of MB at %dx%d:\n", s->mb_x, s->mb_y);
       for (int i = 0; i < 6; i++) {
           for (int j = 0; j < 64; j++) {
               av_log(s->avctx, AV_LOG_DEBUG, "%5d",
                      block[i][s->idsp.idct_permutation[j]]);
           }
           av_log(s->avctx, AV_LOG_DEBUG, "\n");
       }
    }

    if (!s->avctx->lowres) {
#if !CONFIG_SMALL
        if (s->out_format == FMT_MPEG1)
            mpv_reconstruct_mb_internal(s, block, 0, DEFINITELY_MPEG12);
        else
            mpv_reconstruct_mb_internal(s, block, 0, NOT_MPEG12);
#else
        mpv_reconstruct_mb_internal(s, block, 0, MAY_BE_MPEG12);
#endif
    } else
        mpv_reconstruct_mb_internal(s, block, 1, MAY_BE_MPEG12);
}