[FFmpeg-devel] [PATCH 1/2] avformat/network: fix timeout inaccurate in wait_fd_timeout

[Marton Balint]

> The wait_start was about POLLING_TIME larger which leads to timeout
> 100ms late than the option setting.
> ---
> libavformat/network.c | 13 ++++++-------
> 1 file changed, 6 insertions(+), 7 deletions(-)
>
> diff --git a/libavformat/network.c b/libavformat/network.c
> index 0f5a575f77..7a9a4be5bb 100644
> --- a/libavformat/network.c
> +++ b/libavformat/network.c
> @@ -78,7 +78,10 @@ int ff_network_wait_fd(int fd, int write)
> int ff_network_wait_fd_timeout(int fd, int write, int64_t timeout, AVIOInterruptCB *int_cb)
> {
>     int ret;
> -    int64_t wait_start = 0;
> +    int64_t wait_start;
> +
> +    if (timeout > 0)
> +        wait_start = av_gettime_relative();

I think we intentionally wanted to avoid calling gettime on the fast path.

>
>     while (1) {
>         if (ff_check_interrupt(int_cb))
> @@ -86,12 +89,8 @@ int ff_network_wait_fd_timeout(int fd, int write, int64_t timeout, AVIOInterrupt
>         ret = ff_network_wait_fd(fd, write);
>         if (ret != AVERROR(EAGAIN))
>             return ret;
> -        if (timeout > 0) {
> -            if (!wait_start)
> -                wait_start = av_gettime_relative();

Why not simply wait_start = av_gettime_relative() - POLLING_TIME? It seems 
to achieve the same result.

Thanks,
Marton

> -            else if (av_gettime_relative() - wait_start > timeout)
> -                return AVERROR(ETIMEDOUT);
> -        }
> +        if (timeout > 0 && (av_gettime_relative() - wait_start > timeout))
> +            return AVERROR(ETIMEDOUT);
>     }
> }
> 
> -- 
> 2.28.0
>
> _______________________________________________
> ffmpeg-devel mailing list
> ffmpeg-devel at ffmpeg.org
> https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
>
> To unsubscribe, visit link above, or email
> ffmpeg-devel-request at ffmpeg.org with subject "unsubscribe".