[FFmpeg-devel] [PATCH v6 2/2] avformat: add demuxer for Pro Pinball Series' Soundbanks

[Anton Khirnov]

Quoting Zane van Iperen (2020-03-29 19:18:20)
> Signed-off-by: Zane van Iperen <zane at zanevaniperen.com>
> +static int pp_bnk_read_header(AVFormatContext *s)
> +{
> +    int ret;
> +    AVStream *st;
> +    AVCodecParameters *par;
> +    PPBnkCtx *ctx = s->priv_data;
> +    uint8_t buf[FFMAX(PP_BNK_FILE_HEADER_SIZE, PP_BNK_TRACK_SIZE)];
> +    PPBnkHeader hdr;
> +
> +    if ((ret = avio_read(s->pb, buf, PP_BNK_FILE_HEADER_SIZE)) < 0)
> +        return ret;
> +    else if (ret != PP_BNK_FILE_HEADER_SIZE)
> +        return AVERROR(EIO);
> +
> +    pp_bnk_parse_header(&hdr, buf);
> +
> +    if (hdr.track_count == 0 || hdr.track_count > INT_MAX)
> +        return AVERROR_INVALIDDATA;
> +
> +    if (hdr.sample_rate == 0 || hdr.sample_rate > INT_MAX)
> +        return AVERROR_INVALIDDATA;
> +
> +    if (hdr.always1 != 1) {
> +        avpriv_request_sample(s, "Non-one header value");
> +        return AVERROR_PATCHWELCOME;
> +    }
> +
> +    ctx->track_count = hdr.track_count;
> +    if ((ret = av_reallocp_array(&ctx->tracks, hdr.track_count, sizeof(PPBnkCtxTrack))))

Why realloc? Seems this is only allocated once.
> +        return ret;
> +
> +    /* Parse and validate each track. */
> +    for (int i = 0; i < hdr.track_count; i++) {
> +        PPBnkTrack e;
> +
> +        if ((ret = avio_read(s->pb, buf, PP_BNK_TRACK_SIZE)) < 0) {
> +            av_freep(&ctx->tracks);

You are duplicating this free at too many places. Would be better to
have a cleanup block at the end and jump to that.

> +            return ret;
> +        } else if (ret != PP_BNK_TRACK_SIZE) {
> +            av_freep(&ctx->tracks);
> +            return AVERROR(EIO);
> +        }
> +
> +        pp_bnk_parse_track(&e, buf);
> +
> +        /* The individual sample rates of all tracks must match that of the file header. */
> +        if (e.sample_rate != hdr.sample_rate) {
> +            av_freep(&ctx->tracks);
> +            return AVERROR_INVALIDDATA;
> +        }
> +
> +        ctx->tracks[i].data_offset = avio_tell(s->pb);
> +        ctx->tracks[i].data_size   = e.size;
> +
> +        /* Skip over the data to the next stream header. */
> +        avio_skip(s->pb, e.size);
> +    }
> +
> +    /* Build the streams. */
> +    for (int i = 0; i < hdr.track_count; i++) {
> +

nit: unnecessary empty line

> +        if (!(st = avformat_new_stream(s, NULL))) {
> +            av_freep(&ctx->tracks);
> +            return AVERROR(ENOMEM);
> +        }
> +
> +        par                         = st->codecpar;
> +        par->codec_type             = AVMEDIA_TYPE_AUDIO;
> +        par->codec_id               = AV_CODEC_ID_ADPCM_IMA_CUNNING;
> +        par->format                 = AV_SAMPLE_FMT_S16;
> +        par->channel_layout         = AV_CH_LAYOUT_MONO;
> +        par->channels               = 1;
> +        par->sample_rate            = hdr.sample_rate;
> +        par->bits_per_coded_sample  = 4;
> +        par->bits_per_raw_sample    = 16;
> +        par->block_align            = 1;
> +        par->bit_rate               = par->sample_rate * par->bits_per_coded_sample;
> +
> +        avpriv_set_pts_info(st, 64, 1, par->sample_rate);
> +        st->start_time              = 0;
> +        st->duration                = ctx->tracks[i].data_size * 2;
> +    }
> +
> +    /* Seek to the start of the first stream. */
> +    if ((ret = avio_seek(s->pb, ctx->tracks[0].data_offset, SEEK_SET)) < 0) {
> +        av_freep(&ctx->tracks);
> +        return ret;
> +    } else if (ret != ctx->tracks[0].data_offset) {
> +        av_freep(&ctx->tracks);
> +        return AVERROR(EIO);
> +    }
> +
> +    return 0;
> +}
> +
> +static int pp_bnk_read_packet(AVFormatContext *s, AVPacket *pkt)
> +{
> +    int64_t ret;
> +    int size;
> +    PPBnkCtx *ctx = s->priv_data;
> +    PPBnkCtxTrack *trk = ctx->tracks + ctx->current_track;
> +
> +    av_assert0(ctx->bytes_read <= trk->data_size);
> +
> +    if (ctx->bytes_read == trk->data_size) {
> +

nit: unnecessary empty line

Otherwise looks ok, but would be nice to have some tests.

-- 
Anton Khirnov