Question

i have a multi-threaded application which wants to send a sequence of data to an external device via a serial port. the sequence of data is a typical cmd – response protocol (ie: a given thread sends a sequence of bytes then waits to read a response which is typically an ack and then it might send another sequence).

what we are are looking to do is declare a sequence of code has exclusive access to this resource until it is done and if another thread wants access to the same external resouce, it waits.

this seems like what LOCK does, but all the examples that i have seen show lock being used to protect a specific block of code, not to serialize access to a resource.

so programatically can i have

Object serialPortLock = new Object();

and in different parts of my program use a construct that looks like:

Lock (serialPortLock)
{    
    // my turn to do something that is not the same as what
    // someone else wants to do but it acts on the same resource    
}

the c# documentation talks about using Mutex as a more robust version of Lock. is that whats required here?

Answer 1

Yes, your pattern is correct as long as your program is the only software accessing the serial port.

You have not posted your entire code. If the class that contains serialPortLock has multiple instances, then you MUST make serialPortLock a static. This is normally best practice.

class MySerialPort
{
   static object synchLock = new object();

   public void DoSomething()
   {
      lock (synchLock)
      {
        // whatever
      }
   }
}