CON17-J. Avoid using ThreadGroup APIs

A ThreadGroup is a group of threads as defined by the class java.lang.ThreadGroup. A thread is assign to a thread group upon the thread's creation. If the group name is not specified explicitly, the default group called main is assigned by the Java Virtual Machine (JVM). The convenience methods of the ThreadGroup class can be used to operate on all threads in its group at once. For instance, the ThreadGroup.interrupt() method will invoke interrupt() on all threads in the thread group. Another use for thread groups is to reinforce layered security by confining threads into groups so that they do not interfere with each other. [[JavaThreads 04]]

ThreadGroups are still used by Java to organize threads into a hierarchy; consequently they must be used when managing the thread hierarchy. However, many of the methods of ThreadGroup are deprecated, including allowThreadSuspension, resume, stop, suspend. Furthermore, many non-deprecated methods of ThreadGroup are seldom used because they offer little desirable functionality. Ironically, a few ThreadGroup methods are not even thread-safe. [[Bloch 01]]

Some of the insecure, yet non-deprecated methods are:

ThreadGroup.activeCount(): A programmer may sometimes wish to enumerate all the threads in a group as a precursor to other operations. This is accomplished by using the activeCount() method that "Returns an estimate of the number of active threads in this thread group. " [[API 06]]. Note that there is no conclusive information on whether it returns the exact count or not; the definition of active also has a different connotation here. A thread that is constructed and not started is still counted by the activeCount() method as active.

ThreadGroup.enumerate(): According to the Java API [[API 06]], class ThreadGroup documentation:

[The enumerate() method] Copies into the specified array every active thread in this thread group and its subgroups. An application should use the activeCount method to get an estimate of how big the array should be. If the array is too short to hold all the threads, the extra threads are silently ignored.

Threads are removed from the thread array either when they are stopped or when their run() method has finished executing. As a result, if a thread is not started, it continues to reside in the array despite the loss of the original reference. [[JavaThreads 99]]

Using the ThreadGroup APIs to facilitate thread shutdown also has pitfalls. Because the stop() method is deprecated, alternative ways are required to stop threads. "One way is for the thread initiating the termination to join the other threads and so know when those threads have terminated. However, an application may have to maintain its own list of the threads it creates because simply inspecting the ThreadGroup may return library threads that do not terminate and for which join will not return." [[JPL 06]].

Thread pools, as exemplified by classes in java.util.concurrent such as ThreadPoolExecutor, provide a better API for managing groups of threads, with more secure facilities for handling shutdown and thread exceptions. If a logical grouping of threads is desired, use thread pools [[Bloch 08]].

Noncompliant Code Example

This noncompliant code example shows a NetworkHandler class that maintains a controller thread controller. This thread is responsible for delegating every network request to be handled to other threads. For the sake of brevity, it is assumed that the thread controller services these requests by starting three threads in succession. All threads are defined to belong to the same group, Chief.

final class HandleRequest implements Runnable {
  public void run() {
    // Do something
  }
}


public final class NetworkHandler implements Runnable {
  private static ThreadGroup tg = new ThreadGroup("Chief");

  public void run() {
    new Thread(tg, new HandleRequest(), "thread1").start(); // Start thread 1
    new Thread(tg, new HandleRequest(), "thread2").start(); // Start thread 2	
    new Thread(tg, new HandleRequest(), "thread3").start(); // Start thread 3
  }
	
  public static void printActiveCount(int point) {
    System.out.println("Active Threads in Thread Group " + tg.getName() + 
      " at point(" + point + "):" + " " + tg.activeCount());
  }

  public static void printEnumeratedThreads(Thread[] ta, int len) {
    System.out.println("Enumerating all threads...");
    for(int i = 0; i < len; i++) {
      System.out.println("Thread " + i + " = " + ta[i].getName());
    }   
  }

  public static void main(String[] args) throws InterruptedException {
    Thread thread = new Thread(tg, new NetworkHandler(), "controller"); // start thread t
    thread.start();
	     
    Thread[] ta = new Thread[tg.activeCount()]; // Gets the active count (insecure) 		

    printActiveCount(1);           // At point 1 
    Thread.sleep(1000);            // Delay to demonstrate TOCTOU condition (race window)   
    printActiveCount(2);           // At point 2, the thread count changes as new threads are initiated
    int n = tg.enumerate(ta);      // Incorrectly uses the (now stale) thread count obtained at point 1	
    printEnumeratedThreads(ta, n); // Silently ignores printing newly initiated threads
                                   // (between point 1 and point 2)   

    // This code destroys the thread group if it does not have any alive threads
    for (Thread thr : ta) { 
      thr.interrupt();
      while(thr.isAlive());
    }
    tg.destroy();
  }
}

There is a Time of Check-Time of Use (TOCTOU) vulnerability in this implementation because obtaining the count and enumerating the list do not constitute an atomic operation. If new requests come in during the time interval between calling activeCount() and enumerate() in the main() method, the total number of threads in the group will most likely increase but the enumerated list ta will contain only the initial number, that is two thread references (main and controller). Consequently, the program will operate on stale data instead of taking into account the newly started threads in the thread group Chief.

Different runs of the program produce different values of the number of threads in the thread group Chief because of the race conditions and demonstrate how a few incoming requests can get completely ignored. Any subsequent use of the ta array is insecure. For example, using the destroy() method to destroy the thread group and its sub-groups does not work as expected. The precondition to calling destroy() is that the thread group is empty with no executing threads and the code tries to ensure this. However, when the destroy() method is called, the thread group is not empty which raises a java.lang.IllegalThreadStateException.

Compliant Solution

This compliant solution uses a fixed thread pool to group its three tasks, instead of a ThreadGroup. It also uses a java.lang.util.concurrent.ExecutorService to manage the pool. The thread pool provided does not provide an API to find the number of actively executing threads or enumerate through them. However, the logical grouping can help control the behavior of the group as a whole. For instance, all threads belonging to a particular thread pool can be terminated at will.

public final class NetworkHandler {
  private final ExecutorService executor;
  private final int poolSize;

  NetworkHandler(int poolSize) {
    this.poolSize = poolSize;
    this.executor = Executors.newFixedThreadPool(poolSize);
  }
      
  public void startThreads() {
    executor.execute(new Runnable() {
      public void run() {
        for (int i = 0; i < poolSize; i++) {
          executor.submit(new HandleRequest());
        }
      }
    });
  }
  
  public void shutdownPool() {
    executor.shutdown();
  }

  public static void main(String[] args)  {
    NetworkHandler nh = new NetworkHandler(3);
    nh.startThreads();
    nh.shutdownPool();
  }
}

Before Java 5.0, the ThreadGroup class had to be extended because there was no other direct way to catch an uncaught exception in a separate thread. If the application had an installed exception handler UncaughtExceptionHandler, the only way to control it was to subclass ThreadGroup. In recent versions, the UncaughtExceptionHandler was maintained on a per-thread basis using an interface enclosed by the Thread class, leaving little to no functionality for the ThreadGroup class. [[Goetz 06]], [[Bloch 08]]

Refer to [CON37-J. Ensure that tasks executing in a thread pool do not fail silently] for more information on using uncaught exception handlers in thread pools.

Risk Assessment

Using the ThreadGroup APIs may result in race conditions, memory leaks and inconsistent object state.

Automated Detection

TODO

Related Vulnerabilities

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

References

[[API 06]] Methods activeCount and enumerate, Classes ThreadGroup and Thread
[[JavaThreads 04]] 13.1 ThreadGroups
[[Bloch 01]] Item 53: Avoid thread groups
[[Bloch 08]] Item 73: Avoid thread groups
[[Goetz 06]] 7.3.1. Uncaught Exception Handlers
[[JPL 06]] 23.3.3. Shutdown Strategies
[[SDN 06]] Bug ID: 4089701 and 4229558

[!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!]