threadshare: ts-udpsink changes state from NULL to READY failure due to prepare_socket

When working with pipelines that has ts-udpsink in and its state is changed in sequence: NULL->PLAYING, then PLAYING->NULL, and then back to NULL->PLAYING, the third changing state got failure.

I found that when changing state from NULL to READY, the UdpSink call the method prepare_socket that will process in backtrace:

UdpSink::change_state
	gst::StateChange::NullToReady =>
	   UdpSink::prepare  
	      Udpsink::prepare_socket  
	                let socket = context.enter(|| {  
	                Async::<UdpSocket>::try_from(socket).map_err(|err| {  ==> (1)
	                    error_msg!(  
	                        gst::ResourceError::OpenWrite,    ===> error was raised here.  
	                        ["Failed to setup Async socket: {}", err]  
	                    )  
	                })  
	            })?;

(1)
Async::<UdpSocket>::try_from
    Async::new  
        Reactor::insert_io  
            polling::Poller::add  ===> errror here

This makes me suspect that the socket was added twice leading error, so I checked with a simple rust program that call Poller::add two times for a socket and I got the error:

    thread 'main' panicked at 'called `Result::unwrap()` on an `Err` value: Os { code: 17, kind: AlreadyExists, message: "File exists" }', src/main.rs:11:42