[FFmpeg-devel] [PATCH v2 1/3] avformat/file: add fd protocol

Rémi Denis-Courmont remi at remlab.net
Tue Dec 13 18:28:44 EET 2022


Le sunnuntaina 11. joulukuuta 2022, 17.17.27 EET Zhao Zhili a écrit :
> From: Zhao Zhili <zhilizhao at tencent.com>
> 
> Unlike the pipe protocol, fd protocol has seek support if it
> corresponding to a regular file.
> ---
> v2: dup the file descriptor for safety
> 
>  doc/protocols.texi      | 24 ++++++++++++++++++
>  libavformat/Makefile    |  1 +
>  libavformat/file.c      | 55 +++++++++++++++++++++++++++++++++++++++++
>  libavformat/protocols.c |  1 +
>  libavformat/version.h   |  4 +--
>  5 files changed, 83 insertions(+), 2 deletions(-)
> 
> diff --git a/doc/protocols.texi b/doc/protocols.texi
> index 5e9198e67c..d9f2d2dec3 100644
> --- a/doc/protocols.texi
> +++ b/doc/protocols.texi
> @@ -275,6 +275,30 @@ For example, to convert a GIF file given inline with
> @command{ffmpeg}: ffmpeg -i
> "data:image/gif;base64,R0lGODdhCAAIAMIEAAAAAAAA//8AAP//AP///////////////ywA
> AAAACAAIAAADF0gEDLojDgdGiJdJqUX02iB4E8Q9jUMkADs=" smiley.png @end example
> 
> + at section fd
> +
> +File descriptor access protocol.
> +
> +The accepted syntax is:
> + at example
> +fd:[@var{number}]
> + at end example
> +
> + at var{number} is the number corresponding to a file descriptor. Unlike the
> pipe +protocol, fd protocol has seek support if it corresponding to a
> regular file. +If @var{number} is not specified, by default the stdout file
> descriptor will +be used for writing, stdin for reading.
> +
> +This protocol accepts the following options:
> +
> + at table @option
> + at item blocksize
> +Set I/O operation maximum block size, in bytes. Default value is
> + at code{INT_MAX}, which results in not limiting the requested block size.
> +Setting this value reasonably low improves user termination request
> reaction +time, which is valuable if data transmission is slow.
> + at end table
> +
>  @section file
> 
>  File access protocol.
> diff --git a/libavformat/Makefile b/libavformat/Makefile
> index d7f198bf39..1452216e29 100644
> --- a/libavformat/Makefile
> +++ b/libavformat/Makefile
> @@ -647,6 +647,7 @@ OBJS-$(CONFIG_DATA_PROTOCOL)             += data_uri.o
>  OBJS-$(CONFIG_FFRTMPCRYPT_PROTOCOL)      += rtmpcrypt.o rtmpdigest.o
> rtmpdh.o OBJS-$(CONFIG_FFRTMPHTTP_PROTOCOL)       += rtmphttp.o
>  OBJS-$(CONFIG_FILE_PROTOCOL)             += file.o
> +OBJS-$(CONFIG_FD_PROTOCOL)               += file.o
>  OBJS-$(CONFIG_FTP_PROTOCOL)              += ftp.o urldecode.o
>  OBJS-$(CONFIG_GOPHER_PROTOCOL)           += gopher.o
>  OBJS-$(CONFIG_GOPHERS_PROTOCOL)          += gopher.o
> diff --git a/libavformat/file.c b/libavformat/file.c
> index 6103c37b34..b8725c1f48 100644
> --- a/libavformat/file.c
> +++ b/libavformat/file.c
> @@ -109,6 +109,13 @@ static const AVClass pipe_class = {
>      .version    = LIBAVUTIL_VERSION_INT,
>  };
> 
> +static const AVClass fd_class = {
> +    .class_name = "fd",
> +    .item_name  = av_default_item_name,
> +    .option     = pipe_options,
> +    .version    = LIBAVUTIL_VERSION_INT,
> +};
> +
>  static int file_read(URLContext *h, unsigned char *buf, int size)
>  {
>      FileContext *c = h->priv_data;
> @@ -412,3 +419,51 @@ const URLProtocol ff_pipe_protocol = {
>  };
> 
>  #endif /* CONFIG_PIPE_PROTOCOL */
> +
> +#if CONFIG_FD_PROTOCOL
> +
> +static int fd_open(URLContext *h, const char *filename, int flags)
> +{
> +    FileContext *c = h->priv_data;
> +    int fd;
> +    char *final;
> +    struct stat st;
> +
> +    av_strstart(filename, "fd:", &filename);
> +
> +    fd = strtol(filename, &final, 10);
> +    if ((filename == final) || *final ) {
> +        if (flags & AVIO_FLAG_WRITE) {
> +            fd = 1;
> +        } else {
> +            fd = 0;
> +        }
> +    }
> +#if HAVE_SETMODE
> +    setmode(fd, O_BINARY);
> +#endif
> +    if (fstat(fd, &st) < 0)
> +        return AVERROR(errno);
> +    h->is_streamed = !(S_ISREG(st.st_mode) || S_ISBLK(st.st_mode));
> +    c->fd = dup(fd);

This leaks c->fd into child processes. You need atomic close-on-exec.


-- 
Реми Дёни-Курмон
http://www.remlab.net/





More information about the ffmpeg-devel mailing list