ffmpeg源码跟踪笔记之avformat_open_input

1、函数调用图

2、带注释的源码

int avformat_open_input(AVFormatContext **ps, const char *filename,
                        AVInputFormat *fmt, AVDictionary **options)
{
    AVFormatContext *s = *ps;
    int ret = 0;
    AVDictionary *tmp = NULL;
    ID3v2ExtraMeta *id3v2_extra_meta = NULL;
    //为context申请空间,并赋初值，为internal申请空间
    if (!s && !(s = avformat_alloc_context()))
        return AVERROR(ENOMEM);
    if (!s->av_class) {
        av_log(NULL, AV_LOG_ERROR, "Input context has not been properly allocated by avformat_alloc_context() and is not NULL either\n");
        return AVERROR(EINVAL);
    }
    if (fmt)//如果指定了AVInputFormat则使用指定的
        s->iformat = fmt;

    if (options)//把输入参数拷贝一份
        av_dict_copy(&tmp, *options, 0);
    //如果AVFormatContext的AVClass中有这个参数的话，那就设置参数，AVFormatContext具体有哪些参数可以查看options_table.h里的avformat_options
    if ((ret = av_opt_set_dict(s, &tmp)) < 0)
        goto fail;
    //Open input file and probe the format if necessary
    if ((ret = init_input(s, filename, &tmp)) < 0)
        goto fail;
    s->probe_score = ret;

    if (s->format_whitelist && av_match_list(s->iformat->name, s->format_whitelist, ',') <= 0) {
        av_log(s, AV_LOG_ERROR, "Format not on whitelist\n");
        ret = AVERROR(EINVAL);
        goto fail;
    }

    avio_skip(s->pb, s->skip_initial_bytes);

    /* Check filename in case an image number is expected. */
    if (s->iformat->flags & AVFMT_NEEDNUMBER) {
        if (!av_filename_number_test(filename)) {
            ret = AVERROR(EINVAL);
            goto fail;
        }
    }

    s->duration = s->start_time = AV_NOPTS_VALUE;
    av_strlcpy(s->filename, filename ? filename : "", sizeof(s->filename));
    //为priv_data申请空间并根据传入的输入参数赋值，其实这个priv_data大多数是为了给AVInputFormat的read_header函数传递参数用的
    //AVInputFormat具体支持哪些参数就需要具体看AVInputFormat的定义了，拿yuv来举例，查看rawvideodec.c的rawvideo_options就知道yuv具体支持哪些参数了
    if (s->iformat->priv_data_size > 0) {
        if (!(s->priv_data = av_mallocz(s->iformat->priv_data_size))) {
            ret = AVERROR(ENOMEM);
            goto fail;
        }
        if (s->iformat->priv_class) {
            *(const AVClass **) s->priv_data = s->iformat->priv_class;
            av_opt_set_defaults(s->priv_data);
            if ((ret = av_opt_set_dict(s->priv_data, &tmp)) < 0)
                goto fail;
        }
    }

    /* e.g. AVFMT_NOFILE formats will not have a AVIOContext */
    if (s->pb)//还没研究过为什么老是有关于id3v2的代码
        ff_id3v2_read(s, ID3v2_DEFAULT_MAGIC, &id3v2_extra_meta, 0);
    //如果有这个read_header函数指针那么就调用，其实就是设置某些初始值，拿yuv来说就是设置宽、高、像素格式这些东西，参数会通过AVFormatContext的priv_data来传递
    if (!(s->flags&AVFMT_FLAG_PRIV_OPT) && s->iformat->read_header)
        if ((ret = s->iformat->read_header(s)) < 0)
            goto fail;

    if (id3v2_extra_meta) {
        if (!strcmp(s->iformat->name, "mp3") || !strcmp(s->iformat->name, "aac") ||
            !strcmp(s->iformat->name, "tta")) {
            if ((ret = ff_id3v2_parse_apic(s, &id3v2_extra_meta)) < 0)
                goto fail;
        } else
            av_log(s, AV_LOG_DEBUG, "demuxer does not support additional id3 data, skipping\n");
    }
    ff_id3v2_free_extra_meta(&id3v2_extra_meta);

    if ((ret = avformat_queue_attached_pictures(s)) < 0)
        goto fail;

    if (!(s->flags&AVFMT_FLAG_PRIV_OPT) && s->pb && !s->internal->data_offset)
        s->internal->data_offset = avio_tell(s->pb);

    s->internal->raw_packet_buffer_remaining_size = RAW_PACKET_BUFFER_SIZE;

    if (options) {
        av_dict_free(options);
        *options = tmp;
    }
    *ps = s;
    return 0;

fail:
    ff_id3v2_free_extra_meta(&id3v2_extra_meta);
    av_dict_free(&tmp);
    if (s->pb && !(s->flags & AVFMT_FLAG_CUSTOM_IO))
        avio_closep(&s->pb);
    avformat_free_context(s);
    *ps = NULL;
    return ret;
}

3、简单解释

此函数最重要的也就5个点。

第一个点：

avformat_alloc_context申请空间并设默认值。

第二个点：

av_opt_set_dict为AVFormatContext设置输入参数。

第 三个点：

init_input打开文件并把格式匹配出来

第四个点：

av_opt_set_dict为为priv_data申请空间并根据传入的输入参数赋值

第五个点：

read_header为InputFormat初始化，并new stream

4、av_opt_set_dict如何设置值

根据上面的函数调用图知道此函数是通过av_opt_set来设置值的。

看看av_opt_set的代码：

int av_opt_set(void *obj, const char *name, const char *val, int search_flags)
{
    int ret = 0;
    void *dst, *target_obj;
    const AVOption *o = av_opt_find2(obj, name, NULL, 0, search_flags, &target_obj);
    if (!o || !target_obj)
        return AVERROR_OPTION_NOT_FOUND;
    if (!val && (o->type != AV_OPT_TYPE_STRING &&
                 o->type != AV_OPT_TYPE_PIXEL_FMT && o->type != AV_OPT_TYPE_SAMPLE_FMT &&
                 o->type != AV_OPT_TYPE_IMAGE_SIZE && o->type != AV_OPT_TYPE_VIDEO_RATE &&
                 o->type != AV_OPT_TYPE_DURATION && o->type != AV_OPT_TYPE_COLOR &&
                 o->type != AV_OPT_TYPE_CHANNEL_LAYOUT))
        return AVERROR(EINVAL);

    if (o->flags & AV_OPT_FLAG_READONLY)
        return AVERROR(EINVAL);

    dst = ((uint8_t*)target_obj) + o->offset;
    switch (o->type) {
    case AV_OPT_TYPE_STRING:   return set_string(obj, o, val, dst);
    case AV_OPT_TYPE_BINARY:   return set_string_binary(obj, o, val, dst);
    case AV_OPT_TYPE_FLAGS:
    case AV_OPT_TYPE_INT:
    case AV_OPT_TYPE_INT64:
    case AV_OPT_TYPE_FLOAT:
    case AV_OPT_TYPE_DOUBLE:
    case AV_OPT_TYPE_RATIONAL: return set_string_number(obj, target_obj, o, val, dst);
    case AV_OPT_TYPE_IMAGE_SIZE: return set_string_image_size(obj, o, val, dst);
    case AV_OPT_TYPE_VIDEO_RATE: return set_string_video_rate(obj, o, val, dst);
    case AV_OPT_TYPE_PIXEL_FMT:  return set_string_pixel_fmt(obj, o, val, dst);
    case AV_OPT_TYPE_SAMPLE_FMT: return set_string_sample_fmt(obj, o, val, dst);
    case AV_OPT_TYPE_DURATION:
        if (!val) {
            *(int64_t *)dst = 0;
            return 0;
        } else {
            if ((ret = av_parse_time(dst, val, 1)) < 0)
                av_log(obj, AV_LOG_ERROR, "Unable to parse option value \"%s\" as duration\n", val);
            return ret;
        }
        break;
    case AV_OPT_TYPE_COLOR:      return set_string_color(obj, o, val, dst);
    case AV_OPT_TYPE_CHANNEL_LAYOUT:
        if (!val || !strcmp(val, "none")) {
            *(int64_t *)dst = 0;
        } else {
#if FF_API_GET_CHANNEL_LAYOUT_COMPAT
            int64_t cl = ff_get_channel_layout(val, 0);
#else
            int64_t cl = av_get_channel_layout(val);
#endif
            if (!cl) {
                av_log(obj, AV_LOG_ERROR, "Unable to parse option value \"%s\" as channel layout\n", val);
                ret = AVERROR(EINVAL);
            }
            *(int64_t *)dst = cl;
            return ret;
        }
        break;
    }

    av_log(obj, AV_LOG_ERROR, "Invalid option type.\n");
    return AVERROR(EINVAL);
}

代码大致流程是，先调用av_opt_find2找到与输入参数对应的AVOption和目标对象指针也就是target_obj，具体设置那个值是由offset来确定的：dst = ((uint8_t*)target_obj) + o->offset;根据o->type类型来调用不同的参数达到为目标对象的指定字段设置值的目的。

看一下yuv的option：

static const AVOption rawvideo_options[] = {
    { "video_size", "set frame size", OFFSET(width), AV_OPT_TYPE_IMAGE_SIZE, {.str = NULL}, 0, 0, DEC },
    { "pixel_format", "set pixel format", OFFSET(pixel_format), AV_OPT_TYPE_STRING, {.str = "yuv420p"}, 0, 0, DEC },
    { "framerate", "set frame rate", OFFSET(framerate), AV_OPT_TYPE_VIDEO_RATE, {.str = "25"}, 0, 0, DEC },
    { NULL },
};

如果我们传入video_size 1280x720，那么av_opt_find2会返回&rawvideo_options[0]，然后AV_OPT_TYPE_IMAGE_SIZE类型会调用set_string_image_size函数来设置值，看看set_string_image_size函数：

static int set_string_image_size(void *obj, const AVOption *o, const char *val, int *dst)
{
    int ret;

    if (!val || !strcmp(val, "none")) {
        dst[0] =
        dst[1] = 0;
        return 0;
    }
    ret = av_parse_video_size(dst, dst + 1, val);
    if (ret < 0)
        av_log(obj, AV_LOG_ERROR, "Unable to parse option value \"%s\" as image size\n", val);
    return ret;
}

av_parse_video_size函数：

int av_parse_video_size(int *width_ptr, int *height_ptr, const char *str)
{
    int i;
    int n = FF_ARRAY_ELEMS(video_size_abbrs);
    const char *p;
    int width = 0, height = 0;

    for (i = 0; i < n; i++) {
        if (!strcmp(video_size_abbrs[i].abbr, str)) {
            width  = video_size_abbrs[i].width;
            height = video_size_abbrs[i].height;
            break;
        }
    }
    if (i == n) {
        width = strtol(str, (void*)&p, 10);
        if (*p)
            p++;
        height = strtol(p, (void*)&p, 10);

        /* trailing extraneous data detected, like in 123x345foobar */
        if (*p)
            return AVERROR(EINVAL);
    }
    if (width <= 0 || height <= 0)
        return AVERROR(EINVAL);
    *width_ptr  = width;
    *height_ptr = height;
    return 0;
}

那么这个目标对象是如何来的？

在yuv demuxer定义中：

AVInputFormat ff_rawvideo_demuxer = {
    .name           = "rawvideo",
    .long_name      = NULL_IF_CONFIG_SMALL("raw video"),
    .priv_data_size = sizeof(RawVideoDemuxerContext),
    .read_header    = rawvideo_read_header,
    .read_packet    = rawvideo_read_packet,
    .flags          = AVFMT_GENERIC_INDEX,
    .extensions     = "yuv,cif,qcif,rgb",
    .raw_codec_id   = AV_CODEC_ID_RAWVIDEO,
    .priv_class     = &rawvideo_demuxer_class,
};

我们发现有一个RawVideoDemuxerContext的结构，定义如下：

typedef struct RawVideoDemuxerContext {
    const AVClass *class;     /**< Class for private options. */
    int width, height;        /**< Integers describing video size, set by a private option. */
    char *pixel_format;       /**< Set by a private option. */
    AVRational framerate;     /**< AVRational describing framerate, set by a private option. */
} RawVideoDemuxerContext;

这个结构就包含了yuv demuxer需要的输入参数，来在avformat_open_input的时候，会根据此结构来初始化priv_data，传给av_opt_set_dict的参数就是此结构的指针：

//为priv_data申请空间并根据传入的输入参数赋值，其实这个priv_data大多数是为了给AVInputFormat的read_header函数传递参数用的
    //AVInputFormat具体支持哪些参数就需要具体看AVInputFormat的定义了，拿yuv来举例，查看rawvideodec.c的rawvideo_options就知道yuv具体支持哪些参数了
    if (s->iformat->priv_data_size > 0) {
        if (!(s->priv_data = av_mallocz(s->iformat->priv_data_size))) {
            ret = AVERROR(ENOMEM);
            goto fail;
        }
        if (s->iformat->priv_class) {
            *(const AVClass **) s->priv_data = s->iformat->priv_class;
            av_opt_set_defaults(s->priv_data);
            if ((ret = av_opt_set_dict(s->priv_data, &tmp)) < 0)
                goto fail;
        }
    }

这样就成功的把参数设置到priv_data并且传入InputFormatr的read_header函数了，

看看yuv demuxer的rawvideo_read_header代码：

static int rawvideo_read_header(AVFormatContext *ctx)
{
    RawVideoDemuxerContext *s = ctx->priv_data;
    enum AVPixelFormat pix_fmt;
    AVStream *st;

    st = avformat_new_stream(ctx, NULL);
    if (!st)
        return AVERROR(ENOMEM);

    st->codec->codec_type = AVMEDIA_TYPE_VIDEO;

    st->codec->codec_id = ctx->iformat->raw_codec_id;

    if ((pix_fmt = av_get_pix_fmt(s->pixel_format)) == AV_PIX_FMT_NONE) {
        av_log(ctx, AV_LOG_ERROR, "No such pixel format: %s.\n",
               s->pixel_format);
        return AVERROR(EINVAL);
    }

    avpriv_set_pts_info(st, 64, s->framerate.den, s->framerate.num);

    st->codec->width  = s->width;
    st->codec->height = s->height;
    st->codec->pix_fmt = pix_fmt;
    st->codec->bit_rate = av_rescale_q(avpicture_get_size(st->codec->pix_fmt, s->width, s->height),
                                       (AVRational){8,1}, st->time_base);

    return 0;
}

成功的把AVFormatContext 的priv_data转为RawVideoDemuxerContext结构来做参数了。