TSM03-J. Do not publish partially initialized objects

During initialization of a shared object, the object must only be accessible to the thread constructing it. However, the object can be published safely (that is, made visible to other threads) once it is initialized. The Java Memory Model (JMM) allows multiple threads to observe the object after its initialization has begun, but before it has concluded. Consequently, it is important to ensure that a partially initialized object is not published.

This rule prohibits publishing a reference to a partially initialized member object instance before initialization has concluded. Rule TSM01-J. Do not let the (this) reference escape during object construction prohibits the this reference of the current object from escaping.

Noncompliant Code Example

This noncompliant code example constructs a Helper object in the initialize() method of the Foo class. The Helper object's fields are initialized by its constructor.

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;
  }
  // ...
}

If a thread accesses helper using the getHelper() method before the initialize() method has been called, the thread will observe an uninitialized helper field. Later, if one thread calls initialize(), and another calls getHelper(), the second thread could observe one of the following:

• the helper reference as null
• a fully-initialized Helper object with the n field set to 42
• a partially-initialized Helper object with an uninitialized n, which contains the default value 0

In particular, the JMM permits compilers to allocate memory for the new Helper object and assign it to the helper field before initializing it. In other words, the compiler can reorder the write to the helper instance field with the write that initializes the Helper object (that is, this.n = n) so that the former occurs first. This exposes a race window during which other threads may observe a partially-initialized Helper object instance.

There is a separate issue: if two threads call initialize(), then two Helper objects are created. This is a performance issue and not a correctness issue because n will be properly initialized and the unused Helper objects will be garbage-collected.

Compliant Solution (Synchronization)

The publication of partially constructed object references can be prevented by using method synchronization, as shown in this compliant solution.

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 not execute concurrently. If one thread calls initialize() just before another thread calls getHelper(), the synchronized initialize() method will always finish first. The synchronized keywords establish a happens-before relationship between the two threads. This guarantees that the thread calling getHelper() sees the fully initialized Helper object or none at all (that is, helper contains a null reference). This approach guarantees proper publication for both immutable and mutable members.

Compliant Solution (Final Field)

If the helper field is declared final, 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, this solution requires the assignment of a new Helper instance to helper from Foo's constructor. According to the Java Language Specification, Section 17.5.2, "Reading Final Fields During Construction" [[JLS 2005]]

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 instance should not be published before the Foo class's constructor has finished its initialization. (See rule TSM01-J. Do not let the (this) reference escape during object construction.)

Compliant Solution (Final Field and Thread-safe Composition)

Some collection classes provide thread-safe access to contained elements. If the Helper object is inserted into such a collection, it is guaranteed to be fully initialized before its reference is made visible. This compliant solution encapsulates the helper field in a Vector<Helper>.

class Foo {
  private final Vector<Helper> helper;

  public Foo() {
    helper = new Vector<Helper>();
  }

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

  public synchronized void initialize() {
    if (helper.isEmpty()) {
      helper.add(new Helper(42));
    }
  }
}

The helper field is declared final to guarantee that the vector is created before any accesses take place. It can be initialized safely by invoking the synchronized initialize() method, which ensures that only one Helper object is ever added to the vector. If getHelper() is invoked before initialize(), it calls initialize() to avoid the possibility of a null-pointer de-reference by the client. The getHelper() method does not require synchronization to simply return Helper, and, because the synchronized initialize() method also checks to make sure helper is empty before adding a new Helper object, there is no possibility of exploiting a race condition to add a second object to the vector.

Compliant Solution (Static Initialization)

In this compliant solution, the helper field is statically initialized, ensuring that the object referenced by the field is fully initialized before its reference is visible.

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

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

Although not a requirement, the helper field should be declared final to document the class's immutability.

According to JSR-133, Section 9.2.3, "Static Final Fields" [[JSR-133 2004]]

The rules for class initialization ensure that any thread that reads a static field will be synchronized with the static initialization of that class, which is the only place where static final fields can be set. Thus, no special rules in the JMM are needed for static final fields.

Compliant Solution (Immutable Object - Final Fields, Volatile Reference)

The JMM guarantees that any final fields of an object are fully initialized before a published object becomes visible [[Goetz 2006]]. By declaring n final, the Helper class is made [immutable]. Furthermore, if the helper field is declared volatile in compliance with rule VNA01-J. Ensure visibility of shared references to immutable objects, Helper's reference is guaranteed to be made visible to any thread that calls getHelper() after Helper has been fully initialized.

class Foo {
  private volatile Helper helper;

  public Helper getHelper() {
    return helper;
  }

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

// Immutable Helper
public final class Helper {
  private final int n;

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

This compliant solution requires that helper be declared volatile and class Helper be immutable. If it were not immutable, the code would violate rule VNA06-J. Do not assume that declaring a reference volatile guarantees visibility of the members of the referenced object, and additional synchronization would be necessary. (See the next compliant solution.) And, if the helper field were non-volatile, it would violate rule VNA01-J. Ensure visibility of shared references to immutable objects.

Similarly, a public static factory method that returns a new instance of Helper can be provided in the Helper class. This approach allows the Helper instance to be created in a private constructor.

Compliant Solution (Mutable Thread-safe Object, Volatile Reference)

If Helper is mutable but thread-safe, it can be published safely by declaring the helper field in the Foo class volatile.

class Foo {
  private volatile Helper helper;

  public Helper getHelper() {
    return helper;
  }

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

// Mutable but thread-safe Helper
public class Helper {
  private volatile int n;
  private final Object lock = new Object();

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

  public void setN(int value) {
    synchronized (lock) {
      n = value;
    }
  }
}

Because the Helper object can change state after its construction, synchronization is necessary to ensure the visibility of mutable members after initial publication. Consequently, the setN() method is synchronized to provide the visibility of the n field in this compliant solution. (See rule VNA06-J. Do not assume that declaring a reference volatile guarantees visibility of the members of the referenced object.)

If the Helper class is not synchronized properly, declaring helper volatile in the Foo class only guarantees the visibility of the initial publication of Helper and not of subsequent state changes. Consequently, volatile references alone are inadequate for publishing objects that are not thread-safe.

If the helper field in the Foo class is not declared volatile, the n field should be declared volatile so that a happens-before relationship is established between the initialization of n and the write of Helper to the helper field. This is in compliance with rule VNA06-J. Do not assume that declaring a reference volatile guarantees visibility of the members of the referenced object. This is required only when the caller (class Foo) cannot be trusted to declare helper volatile.

Because the Helper class is declared public, it uses a private lock to handle synchronization in conformance with rule LCK00-J. Use private final lock objects to synchronize classes that may interact with untrusted code.

Exceptions

TSM03-EX1: Classes that prevent partially initialized objects from being used may publish partially initialized objects. This may be implemented, for example, by setting a volatile boolean flag in the last statement of the initializing code and ensuring this flag is set before allowing class methods to execute.

The following compliant solution illustrates this technique:

public class Helper {
  private int n;
  private volatile boolean initialized; // Defaults to false

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

  public void doSomething() {
    if (!initialized) {
      throw new SecurityException("Cannot use partially initialized instance");
    }
    // ...
  }
  // ...
}

This technique ensures that, even if the reference to the Helper object instance is published before its initialization is over, the instance is unusable. The instance is unusable because every method within Helper must check the flag to determine whether the initialization has finished.

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.

Bibliography

      11. Thread-Safety Miscellaneous (TSM)      TSM04-J. Document thread-safety and use annotations where applicable