Page History

Declaring a shared mutable variable volatile ensures the visibility of the latest updates on it across other threads but does not guarantee the atomicity of composite operations. For example, the variable increment operation consisting of the sequence read-modify-write is not atomic even when the variable is declared volatile.

In such cases, the java.util.concurrent utilities must be used to atomically manipulate the variable. If the utilities do not provide the required atomic methods, accesses to the variable must be explicitly synchronized. Note that, as with volatile, updated values are immediately visible to other threads when these two techniques are used. Synchronization provides a way to safely share object state across multiple threads without the need to reason about reordering, compiler optimizations and hardware specific behavior.

Noncompliant Code Example (volatile)

In this noncompliant code example, the volatile field itemsInInventory can be accessed by multiple threads. However, when a thread is updating the value of itemsInInventory, it is possible for other threads to read the original value (that is, the value before the update). This is because the post decrement operator is non-atomic.

private volatile int itemsInInventory = 100;

public int removeItem() {
  if(itemsInInventory > 0) {
    return itemsInInventory--;  // Returns new count of items in inventory
  } else {
    return 0;
  }
}

Compliant Solution (1) (java.util.concurrent.atomic classes)

Volatile variables are unsuitable when more than one load/store operation needs to be atomic. There is an alternative method to perform multiple operations atomically. This compliant solution shows a {{java.util.concurrent.atomic.AtomicInteger}} variable. According to the Java API \[[API 06|AA. Java References#API 06]\], Class {{AtomicInteger}} documentation:

...

public class Sync {
  private final AtomicInteger itemsInInventory = new AtomicInteger(100);

  private int removeItem() {
    for (;;) {
      int old = itemsInInventory.get();
      if (old > 0) {
        int next = old - 1;
        if (itemsInInventory.compareAndSet(old, next)) {
          return next;  //returns new count of items in inventory
        }
      } else {
        return 0;
      }
    }
  }
}

Compliant Solution (2) (method synchronization)

This compliant solution uses method synchronization to synchronize access to shared variables. Consequently, access to itemsInInventory is mutually exclusive and consistent across object states.

...

Synchronization is more expensive than using the optimized {{java.util.concurrent}} utilities and should only be used when the utilities do not contain the required method to carry out the atomic operation. When using explicit synchronization, the programmer must also ensure that two or more threads are not mutually accessible from a different set of two or more threads such that each thread holds a lock while trying to obtain another lock that is held by the other thread \[[Lea 00|AA. Java References#Lea 00]\]. Failure to follow this advice results in deadlocks ([CON12-J. Avoid deadlock by requesting locks in the proper order]).

Compliant Solution (3) (block synchronization)

Constructors and methods can use an alternative representation called block synchronization which synchronizes a block of code rather than a method, as highlighted below.

...

Block synchronization is more preferable than method synchronization because it reduces the period for which the lock is held and also protects against denial of service attacks. The variable itemsInInventory still needs to be declared volatile because the check to determine whether it is greater than 0 relies on the latest value of the variable. An alternative to avoid the need to declare the variable volatile is to use block synchronization across the whole if-else block. However, this alternative is more costly.

Noncompliant Code Example (AtomicReference)

This noncompliant code example uses two AtomicReference objects to hold two BigInteger object references. An AtomicReference is an object reference that can be updated atomically. Operations that use these two independently are guaranteed to be atomic, however, if an operation involves using both together, thread-safety issues arise.

public class AtomicAdder {
  private final AtomicReference<BigInteger> first ;	
  private final AtomicReference<BigInteger> second; 

  public AtomicAdder(BigInteger f, BigInteger s) {
    first  = new AtomicReference<BigInteger>(f);
    second = new AtomicReference<BigInteger>(s);
  }

  public void update(BigInteger f, BigInteger s){ // Unsafe
    first.set(f);
    second.set(s);
  }

  public BigInteger add() { // Unsafe
    return first.get().add(second.get()); 
  }
}

An AtomicReference is an object reference that can be updated atomically. Operations that use these two independently are guaranteed to be atomic, however, if an operation involves using both together, thread-safety issues arise. For instance, in this noncompliant code example, adding the two big integers is not thread-safe because it is possible that while the addition is being carried out in a thread, another thread may update the value of one or both big integers, leading to an erroneous result.

Compliant Solution (method synchronization)

This compliant solution declares the update() and add() methods as synchronized to guarantee atomicity.

...