Best practices for lock (or Enter/Exit) in C#

I ha开发者_如何学Pythonve a measuring instrument object:

public class Instrument
{
  public double Measure() 
  { 
    return 0; 
  }
}

I have a device that needs to do some measuring:

public class Device
{
  public Instrument MeasuringInstrument { get; set; }
  public void DoMeasuring()
  {
    var result = this.MeasuringInstrument.Measure();
  }
}

The measuring instrument can only operate on one device at a time, yet many devices may use the same instrument. I'm new to threading, and from what I understand, both of the following solutions have caveats.

public class Instrument
{
  public double Measure() 
  { 
    lock(this)
    {
      return 0; 
    }
  }
}

public class Device
{
  public Instrument MeasuringInstrument { get; set; }
  public void DoMeasuring()
  {
    lock(this.MeasurementInstrument)
    {
      var result = this.MeasuringInstrument.Measure();
    }
  }
}

I've read it's best to lock on private objects, but I don't know how to do that while still allowing the MeasuringInstrument to be get/set on the Device. Any suggestions?

Thanks much,

Ken

---

if your instrument is used by multiple devices the best practice is to set lock in your instrument class. so the fist solution works better.

but its better to create a new lock object and use it in instrument class.

public class Instrument
{
  Object lockKey = new Object();
  public double Measure() 
  { 
    lock(lockKey)
    {
      return 0; 
    }
  }
}

---

the usual pattern is to create your own private object just for locking in the case where the obvious choice might be exposed outside of the class, for example:

public class Instrument
{
    private object thisLock = new object();

    public double Measure() 
    { 
        lock(this.thisLock)
        {
            return 0; 
        }
    }
}

public class Device
{
    public Instrument MeasuringInstrument { get; set; }
    private object measuringInstrumentLock = new object();

    public void DoMeasuring()
    {
        lock(this.measuringInstrumentLock)
        {
            var result = this.MeasuringInstrument.Measure();
        }
    }
}

Also, I suspect that you only need one of those two locks (either the one in DoMeasuring or the one in Measure) although that does depend on the missing bits.