CON26-J. Do not publish partially-constructed objects

Sometimes an object is required to be shared amongst multiple threads. During initialization, the object must remain exclusive to the thread constructing it, However, once the object is initialized, it can be published, that is, made visible to other threads.

The Java Memory Model permits a compiler to modify the order of execution instructions in such a way that a seemingly innocuous publication results in multiple threads observing the object after its initialization has begun, but before it has concluded.

This guideline is similar to [CON14-J. Do not let the "this" reference escape during object construction]. The difference is that in this guideline, a reference to a partially constructed object instance is published instead of the this reference of the current object.

Noncompliant Code Example

This noncompliant code example initializes a Helper object inside class Foo.

class Foo {
  private Helper helper;

  public Helper getHelper() {
    return helper;
  }

  public void initialize() {
    helper = new Helper(42);
  }
}

public class Helper {
  private int n;

  public Helper(int n) {
    this.n = n;
  }

  // ...
}

Suppose two threads have access to the same Foo object through the use of the getHelper() method, and initialize() has not been called yet. Both threads will see the helper field as uninitialized. Subsequently, if one thread calls initialize(), and the other calls getHelper(), the second thread may either see the helper reference as null, observe a fully-initialized Helper object with the n field set to 42, or observe a partially-initialized Helper object with an n that has not yet been initialized, and still contains the default value of 0.

In particular, the [Java Memory Model (JMM)] permits compilers to allocate memory for the new Helper object and assign it to the helper field before initializing it. This introduces a race window during which other threads may see a partially-initialized helper object.

Noncompliant Code Example (immutable)

If the Helper class is [immutable], then it can not be changed after it is initialized. However, this does not prevent it from being read before it is completely initialized.

[[JPL 06]], 14.10.2. Final Fields and Security, states:

... indeed you can use final fields to define immutable objects. There is a common misconception that shared access to immutable objects does not require any synchronization because the state of the object never changes. This is a misconception in general because it relies on the assumption that a thread will be guaranteed to see the initialized state of the immutable object, and that need not be the case. The problem is that, while the shared object is immutable, the reference used to access the shared object is itself shared and often mutable. Consequently, a correctly synchronized program must synchronize access to that shared reference, but often programs do not do this, because programmers do not recognize the need to do it. For example, suppose one thread creates a String object and stores a reference to it in a static field. A second thread then uses that reference to access the string. There is no guarantee, based on what we've discussed so far, that the values written by the first thread when constructing the string will be seen by the second thread when it accesses the string.

public class Helper {
  private final int n;

  public Helper(int n) {
    this.n = n;
  }

  // ...
}

Compliant Solution (volatile)

If the helper field is declared as volatile, it is guaranteed to be fully constructed (that is, its fields are properly initialized) before the reference is made visible.

class Foo {
  private volatile Helper helper;

  public Helper getHelper() {
    return helper;
  }

  public void initialize() {
    helper = new Helper(42);
  }
}

Primitive types can also be safely published by publishing an atomic reference to the corresponding boxed type. For instance, an int field can be safely published by publishing an atomic reference to the equivalent Integer using java.util.concurrent.atomic.AtomicReference<Integer>.

Note that if the Helper object is mutable or not thread-safe, you must add additional locking to protect its fields from being modified concurrently by multiple threads as well. See [CON11-J. Do not assume that declaring an object volatile guarantees visibility of its members] for more information.

Also, if the Foo object allows you to change what the helper fields points to, you will want additional locking to ensure that this ability does not lead to data races.

Compliant Solution (final)

If the helper field is declared as final, then it is guaranteed to be fully constructed before its reference is made visible.

class Foo {
  private final Helper helper;

  public Helper getHelper() {
    return helper;
  }

  public Foo() {
    helper = new Helper(42);
  }
}

However, the compiler now disallows setting the helper field to a new object using the initialize() method. Instead, a constructor must be used to initialize helper.

According to the Java Language Specification [[JLS 05]], section 17.5.2 "Reading Final Fields During Construction":

A read of a final field of an object within the thread that constructs that object is ordered with respect to the initialization of that field within the constructor by the usual happens-before rules. If the read occurs after the field is set in the constructor, it sees the value the final field is assigned, otherwise it sees the default value.

Consequently, the reference to the helper field should not be published before class Foo's constructor has finished its initialization.

Note that if the Helper object is mutable or not thread-safe, you must add additional locking to protect its fields from being modified concurrently by multiple threads as well. See [CON11-J. Do not assume that declaring an object volatile guarantees visibility of its members] for more information.

Compliant Solution (synchronized)

The reference of a partially-constructed object can be prevented from being made visible by using method synchronization.

class Foo {
  private Helper helper;

  public synchronized Helper getHelper() {
    return helper;
  }

  public synchronized void initialize() {
    helper = new Helper(42);
  }
}

Synchronizing both methods guarantees that they will never run simultaneously in different threads. If one thread were to call initialize() just before another thread calls getHelper(), the synchronized initialize() method will always finish first, fully initializing the Helper object on its way. This forbids getHelper() from retrieving a Helper object that is partially initialized.

Note that if the Helper object is mutable or not thread-safe, you must add additional locking to protect its fields from being modified concurrently by multiple threads as well. See [CON11-J. Do not assume that declaring an object volatile guarantees visibility of its members] for more information.

Compliant Solution (thread-safe composition)

Some collection classes provide thread-safety of accesses to contained elements. If the helper field is contained in such a collection, the Helper object is guaranteed to be fully initialized before its reference is made visible. This compliant solution encases the helper field in a Vector.

class Foo {
  private Vector<Helper> helper;

  public Helper getHelper() {
    return helper.elementAt(0);
  }

  public void initialize() {
    helper = new Vector<Helper>();
    helper.add(new Helper(42));
  }
}

Note that if the Helper object is mutable or not thread-safe, you must add additional locking to protect its fields from being modified concurrently by multiple threads as well. See [CON11-J. Do not assume that declaring an object volatile guarantees visibility of its members] for more information.

Also, if the Foo object allows you to change what the helper fields points to, you will want additional locking to ensure that this ability does not lead to data races.

Compliant Solution (static initialization)

In this compliant solution, the helper field is initialized in a static block. When initialized statically, any object is guaranteed to be fully initialized before its reference is made visible.

class Foo {
  private static Helper helper = new Helper(42);

  public static Helper getHelper() {
    return helper;
  } 
}

This requires the helper field to be declared static.

Note that if the Helper object is mutable or not thread-safe, you must add additional locking to protect its fields from being modified concurrently by multiple threads as well. See [CON11-J. Do not assume that declaring an object volatile guarantees visibility of its members] for more information.

Also, if the Foo object allows you to change what the helper fields points to, you will want additional locking to ensure that this ability does not lead to data races.

Risk Assessment

Failing to synchronize access to shared mutable data can cause different threads to observe different states of the object or a partially initialized object.

Automated Detection

TODO

Related Vulnerabilities

Search for vulnerabilities resulting from the violation of this rule on the CERT website.

References

[[API 06]]
[[Bloch 01]] Item 48: "Synchronize access to shared mutable data"
[[Goetz 06]] Section 3.5.3 "Safe Publication Idioms"
[[Goetz 07]] Pattern #2: "one-time safe publication"
[[JPL 06]], 14.10.2. Final Fields and Security:
[[Pugh 04]]

[!The CERT Sun Microsystems Secure Coding Standard for Java^button_arrow_left.png!]      [!The CERT Sun Microsystems Secure Coding Standard for Java^button_arrow_up.png!]      [!The CERT Sun Microsystems Secure Coding Standard for Java^button_arrow_right.png!]