FFmpeg/libavcodec/hapdec.c

/*
 * Vidvox Hap decoder
 * Copyright (C) 2015 Vittorio Giovara <vittorio.giovara@gmail.com>
 * Copyright (C) 2015 Tom Butterworth <bangnoise@gmail.com>
 *
 * This file is part of Libav.
 *
 * Libav 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.
 *
 * Libav 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 Libav; if not, write to the Free Software
 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
 */

/**
 * @file
 * Hap decoder
 *
 * Fourcc: Hap1, Hap5, HapY
 *
 * https://github.com/Vidvox/hap/blob/master/documentation/HapVideoDRAFT.md
 */

#include <stdint.h>

#include "libavutil/imgutils.h"

#include "avcodec.h"
#include "bytestream.h"
#include "hap.h"
#include "internal.h"
#include "snappy.h"
#include "texturedsp.h"
#include "thread.h"

/* The first three bytes are the size of the section past the header, or zero
 * if the length is stored in the next long word. The fourth byte in the first
 * long word indicates the type of the current section. */
static int parse_section_header(GetByteContext *gbc, int *section_size,
                                enum HapSectionType *section_type)
{
    if (bytestream2_get_bytes_left(gbc) < 4)
        return AVERROR_INVALIDDATA;

    *section_size = bytestream2_get_le24(gbc);
    *section_type = bytestream2_get_byte(gbc);

    if (*section_size == 0) {
        if (bytestream2_get_bytes_left(gbc) < 4)
            return AVERROR_INVALIDDATA;

        *section_size = bytestream2_get_le32(gbc);
    }

    if (*section_size > bytestream2_get_bytes_left(gbc))
        return AVERROR_INVALIDDATA;
    else
        return 0;
}

static int hap_parse_decode_instructions(HapContext *ctx, int size)
{
    GetByteContext *gbc = &ctx->gbc;
    int section_size;
    enum HapSectionType section_type;
    int is_first_table = 1, had_offsets = 0, had_compressors = 0, had_sizes = 0;
    int i, ret;

    while (size > 0) {
        int stream_remaining = bytestream2_get_bytes_left(gbc);
        ret = parse_section_header(gbc, &section_size, &section_type);
        if (ret != 0)
            return ret;

        size -= stream_remaining - bytestream2_get_bytes_left(gbc);

        switch (section_type) {
            case HAP_ST_COMPRESSOR_TABLE:
                ret = ff_hap_set_chunk_count(ctx, section_size, is_first_table);
                if (ret != 0)
                    return ret;
                for (i = 0; i < section_size; i++) {
                    ctx->chunks[i].compressor = bytestream2_get_byte(gbc) << 4;
                }
                had_compressors = 1;
                is_first_table = 0;
                break;
            case HAP_ST_SIZE_TABLE:
                ret = ff_hap_set_chunk_count(ctx, section_size / 4, is_first_table);
                if (ret != 0)
                    return ret;
                for (i = 0; i < section_size / 4; i++) {
                    ctx->chunks[i].compressed_size = bytestream2_get_le32(gbc);
                }
                had_sizes = 1;
                is_first_table = 0;
                break;
            case HAP_ST_OFFSET_TABLE:
                ret = ff_hap_set_chunk_count(ctx, section_size / 4, is_first_table);
                if (ret != 0)
                    return ret;
                for (i = 0; i < section_size / 4; i++) {
                    ctx->chunks[i].compressed_offset = bytestream2_get_le32(gbc);
                }
                had_offsets = 1;
                is_first_table = 0;
                break;
            default:
                break;
        }
        size -= section_size;
    }

    if (!had_sizes || !had_compressors)
        return AVERROR_INVALIDDATA;

    /* The offsets table is optional. If not present than calculate offsets by
     * summing the sizes of preceding chunks. */
    if (!had_offsets) {
        size_t running_size = 0;
        for (i = 0; i < ctx->chunk_count; i++) {
            ctx->chunks[i].compressed_offset = running_size;
            running_size += ctx->chunks[i].compressed_size;
        }
    }

    return 0;
}

static int hap_can_use_tex_in_place(HapContext *ctx)
{
    int i;
    size_t running_offset = 0;
    for (i = 0; i < ctx->chunk_count; i++) {
        if (ctx->chunks[i].compressed_offset != running_offset
            || ctx->chunks[i].compressor != HAP_COMP_NONE)
            return 0;
        running_offset += ctx->chunks[i].compressed_size;
    }
    return 1;
}

static int hap_parse_frame_header(AVCodecContext *avctx)
{
    HapContext *ctx = avctx->priv_data;
    GetByteContext *gbc = &ctx->gbc;
    int section_size;
    enum HapSectionType section_type;
    const char *compressorstr;
    int i, ret;

    ret = parse_section_header(gbc, &section_size, &section_type);
    if (ret != 0)
        return ret;

    if ((avctx->codec_tag == MKTAG('H','a','p','1') && (section_type & 0x0F) != HAP_FMT_RGBDXT1) ||
        (avctx->codec_tag == MKTAG('H','a','p','5') && (section_type & 0x0F) != HAP_FMT_RGBADXT5) ||
        (avctx->codec_tag == MKTAG('H','a','p','Y') && (section_type & 0x0F) != HAP_FMT_YCOCGDXT5)) {
        av_log(avctx, AV_LOG_ERROR,
               "Invalid texture format %#04x.\n", section_type & 0x0F);
        return AVERROR_INVALIDDATA;
    }

    switch (section_type & 0xF0) {
        case HAP_COMP_NONE:
        case HAP_COMP_SNAPPY:
            ret = ff_hap_set_chunk_count(ctx, 1, 1);
            if (ret == 0) {
                ctx->chunks[0].compressor = section_type & 0xF0;
                ctx->chunks[0].compressed_offset = 0;
                ctx->chunks[0].compressed_size = section_size;
            }
            if (ctx->chunks[0].compressor == HAP_COMP_NONE) {
                compressorstr = "none";
            } else {
                compressorstr = "snappy";
            }
            break;
        case HAP_COMP_COMPLEX:
            ret = parse_section_header(gbc, &section_size, &section_type);
            if (ret == 0 && section_type != HAP_ST_DECODE_INSTRUCTIONS)
                ret = AVERROR_INVALIDDATA;
            if (ret == 0)
                ret = hap_parse_decode_instructions(ctx, section_size);
            compressorstr = "complex";
            break;
        default:
            ret = AVERROR_INVALIDDATA;
            break;
    }

    if (ret != 0)
        return ret;

    /* Check the frame is valid and read the uncompressed chunk sizes */
    ctx->tex_size = 0;
    for (i = 0; i < ctx->chunk_count; i++) {
        HapChunk *chunk = &ctx->chunks[i];

        /* Check the compressed buffer is valid */
        if (chunk->compressed_offset + chunk->compressed_size > bytestream2_get_bytes_left(gbc))
            return AVERROR_INVALIDDATA;

        /* Chunks are unpacked sequentially, ctx->tex_size is the uncompressed
         * size thus far */
        chunk->uncompressed_offset = ctx->tex_size;

        /* Fill out uncompressed size */
        if (chunk->compressor == HAP_COMP_SNAPPY) {
            GetByteContext gbc_tmp;
            int64_t uncompressed_size;
            bytestream2_init(&gbc_tmp, gbc->buffer + chunk->compressed_offset,
                             chunk->compressed_size);
            uncompressed_size = ff_snappy_peek_uncompressed_length(&gbc_tmp);
            if (uncompressed_size < 0) {
                return uncompressed_size;
            }
            chunk->uncompressed_size = uncompressed_size;
        } else if (chunk->compressor == HAP_COMP_NONE) {
            chunk->uncompressed_size = chunk->compressed_size;
        } else {
            return AVERROR_INVALIDDATA;
        }
        ctx->tex_size += chunk->uncompressed_size;
    }

    av_log(avctx, AV_LOG_DEBUG, "%s compressor\n", compressorstr);

    return ret;
}

static int decompress_chunks_thread(AVCodecContext *avctx, void *arg,
                                    int chunk_nb, int thread_nb)
{
    HapContext *ctx = avctx->priv_data;

    HapChunk *chunk = &ctx->chunks[chunk_nb];
    GetByteContext gbc;
    uint8_t *dst = ctx->tex_buf + chunk->uncompressed_offset;

    bytestream2_init(&gbc, ctx->gbc.buffer + chunk->compressed_offset, chunk->compressed_size);

    if (chunk->compressor == HAP_COMP_SNAPPY) {
        int ret;
        int64_t uncompressed_size = ctx->tex_size;

        /* Uncompress the frame */
        ret = ff_snappy_uncompress(&gbc, dst, &uncompressed_size);
        if (ret < 0) {
             av_log(avctx, AV_LOG_ERROR, "Snappy uncompress error\n");
             return ret;
        }
    } else if (chunk->compressor == HAP_COMP_NONE) {
        bytestream2_get_buffer(&gbc, dst, chunk->compressed_size);
    }

    return 0;
}

static int decompress_texture_thread(AVCodecContext *avctx, void *arg,
                                     int slice, int thread_nb)
{
    HapContext *ctx = avctx->priv_data;
    AVFrame *frame = arg;
    const uint8_t *d = ctx->tex_data;
    int w_block = avctx->coded_width / TEXTURE_BLOCK_W;
    int h_block = avctx->coded_height / TEXTURE_BLOCK_H;
    int x, y;
    int start_slice, end_slice;
    int base_blocks_per_slice = h_block / ctx->slice_count;
    int remainder_blocks = h_block % ctx->slice_count;

    /* When the frame height (in blocks) doesn't divide evenly between the
     * number of slices, spread the remaining blocks evenly between the first
     * operations */
    start_slice = slice * base_blocks_per_slice;
    /* Add any extra blocks (one per slice) that have been added before this slice */
    start_slice += FFMIN(slice, remainder_blocks);

    end_slice = start_slice + base_blocks_per_slice;
    /* Add an extra block if there are still remainder blocks to be accounted for */
    if (slice < remainder_blocks)
        end_slice++;

    for (y = start_slice; y < end_slice; y++) {
        uint8_t *p = frame->data[0] + y * frame->linesize[0] * TEXTURE_BLOCK_H;
        int off  = y * w_block;
        for (x = 0; x < w_block; x++) {
            ctx->tex_fun(p + x * 16, frame->linesize[0],
                         d + (off + x) * ctx->tex_rat);
        }
    }

    return 0;
}

static int hap_decode(AVCodecContext *avctx, void *data,
                      int *got_frame, AVPacket *avpkt)
{
    HapContext *ctx = avctx->priv_data;
    ThreadFrame tframe;
    int ret, i;

    bytestream2_init(&ctx->gbc, avpkt->data, avpkt->size);

    /* Check for section header */
    ret = hap_parse_frame_header(avctx);
    if (ret < 0)
        return ret;

    /* Get the output frame ready to receive data */
    tframe.f = data;
    ret = ff_thread_get_buffer(avctx, &tframe, 0);
    if (ret < 0)
        return ret;
    ff_thread_finish_setup(avctx);

    /* Unpack the DXT texture */
    if (hap_can_use_tex_in_place(ctx)) {
        /* Only DXTC texture compression in a contiguous block */
        ctx->tex_data = ctx->gbc.buffer;
    } else {
        /* Perform the second-stage decompression */
        ret = av_reallocp(&ctx->tex_buf, ctx->tex_size);
        if (ret < 0)
            return ret;

        avctx->execute2(avctx, decompress_chunks_thread, NULL,
                        ctx->chunk_results, ctx->chunk_count);

        for (i = 0; i < ctx->chunk_count; i++) {
            if (ctx->chunk_results[i] < 0)
                return ctx->chunk_results[i];
        }

        ctx->tex_data = ctx->tex_buf;
    }

    /* Use the decompress function on the texture, one block per thread */
    avctx->execute2(avctx, decompress_texture_thread, tframe.f, NULL, ctx->slice_count);

    /* Frame is ready to be output */
    tframe.f->pict_type = AV_PICTURE_TYPE_I;
    tframe.f->key_frame = 1;
    *got_frame = 1;

    return avpkt->size;
}

static av_cold int hap_init(AVCodecContext *avctx)
{
    HapContext *ctx = avctx->priv_data;
    const char *texture_name;
    int ret = av_image_check_size(avctx->width, avctx->height, 0, avctx);

    if (ret < 0) {
        av_log(avctx, AV_LOG_ERROR, "Invalid video size %dx%d.\n",
               avctx->width, avctx->height);
        return ret;
    }

    /* Since codec is based on 4x4 blocks, size is aligned to 4 */
    avctx->coded_width  = FFALIGN(avctx->width,  TEXTURE_BLOCK_W);
    avctx->coded_height = FFALIGN(avctx->height, TEXTURE_BLOCK_H);

    /* Technically only one mode has alpha, but 32 bits are easier to handle */
    avctx->pix_fmt = AV_PIX_FMT_RGBA;

    ff_texturedsp_init(&ctx->dxtc);

    switch (avctx->codec_tag) {
    case MKTAG('H','a','p','1'):
        texture_name = "DXT1";
        ctx->tex_rat = 8;
        ctx->tex_fun = ctx->dxtc.dxt1_block;
        break;
    case MKTAG('H','a','p','5'):
        texture_name = "DXT5";
        ctx->tex_rat = 16;
        ctx->tex_fun = ctx->dxtc.dxt5_block;
        break;
    case MKTAG('H','a','p','Y'):
        texture_name = "DXT5-YCoCg-scaled";
        ctx->tex_rat = 16;
        ctx->tex_fun = ctx->dxtc.dxt5ys_block;
        break;
    default:
        return AVERROR_DECODER_NOT_FOUND;
    }

    av_log(avctx, AV_LOG_DEBUG, "%s texture\n", texture_name);

    ctx->slice_count = av_clip(avctx->thread_count, 1,
                               avctx->coded_height / TEXTURE_BLOCK_H);

    return 0;
}

static av_cold int hap_close(AVCodecContext *avctx)
{
    HapContext *ctx = avctx->priv_data;

    ff_hap_free_context(ctx);

    return 0;
}

AVCodec ff_hap_decoder = {
    .name           = "hap",
    .long_name      = NULL_IF_CONFIG_SMALL("Vidvox Hap decoder"),
    .type           = AVMEDIA_TYPE_VIDEO,
    .id             = AV_CODEC_ID_HAP,
    .init           = hap_init,
    .decode         = hap_decode,
    .close          = hap_close,
    .priv_data_size = sizeof(HapContext),
    .capabilities   = AV_CODEC_CAP_FRAME_THREADS | AV_CODEC_CAP_SLICE_THREADS |
                      AV_CODEC_CAP_DR1,
    .caps_internal  = FF_CODEC_CAP_INIT_THREADSAFE |
                      FF_CODEC_CAP_INIT_CLEANUP,
};