According to Goetz and colleagues [Goetz 2006]:

Client-side locking entails guarding client code that uses some object X with the lock X uses to guard its own state. In order to use client-side locking, you must know what lock X uses.

While client-side locking is acceptable when the thread-safe class commits to and clearly documents its locking strategy, Goetz and colleagues caution against its misuse [Goetz 2006]:

If extending a class to add another atomic operation is fragile because it distributes the locking code for a class over multiple classes in an object hierarchy, client-side locking is even more fragile because it entails putting locking code for class C into classes that are totally unrelated to C. Exercise care when using client-side locking on classes that do not commit to their locking strategy.

The documentation of a class that supports client-side locking should explicitly state its applicability. For example, the class java.util.concurrent.ConcurrentHashMap<K,V> should not be used for client-side locking because its documentation [API 2014] states that

even though all operations are thread-safe, retrieval operations do not entail locking, and there is not any support for locking the entire table in a way that prevents all access. This class is fully interoperable with Hashtable in programs that rely on its thread safety but not on its synchronization details.

Use of client-side locking is permitted only when the documentation of the class recommends it. For example, the documentation of the synchronizedList() wrapper method of java.util.Collections class [API 2014] states:

In order to guarantee serial access, it is critical that all access to the backing list is accomplished through the returned list. It is imperative that the user manually synchronize on the returned list when iterating over it. Failure to follow this advice may result in non-deterministic behavior.

When the backing list is inaccessible to an untrusted client, this advice is consistent with LCK04-J. Do not synchronize on a collection view if the backing collection is accessible.

Noncompliant Code Example (Intrinsic Lock)

This noncompliant code example uses a thread-safe Book class that cannot be refactored. Refactoring might be impossible, for example, when the source code is unavailable for review or when the class is part of a general library that cannot be extended.

final class Book {
  // Could change its locking policy in the future
  // to use private final locks
  private final String title;
  private Calendar dateIssued;
  private Calendar dateDue;

  Book(String title) {
    this.title = title;
  }

  public synchronized void issue(int days) {
    dateIssued = Calendar.getInstance();
    dateDue = Calendar.getInstance();
    dateDue.add(dateIssued.DATE, days);
  }

  public synchronized Calendar getDueDate() {
    return dateDue;
  }
}

This class fails to commit to its locking strategy (that is, it reserves the right to change its locking strategy without notice). Furthermore, it fails to document that callers can safely use client-side locking. The BookWrapper client class uses client-side locking in the renew() method by synchronizing on a Book instance.

// Client
public class BookWrapper {
  private final Book book;

  BookWrapper(Book book) {
    this.book = book;
  }

  public void issue(int days) {
    book.issue(days);
  }

  public Calendar getDueDate() {
    return book.getDueDate();
  }

  public void renew() {
    synchronized(book) {
      if (book.getDueDate().before(Calendar.getInstance())) {
        throw new IllegalStateException("Book overdue");
      } else {
        book.issue(14); // Issue book for 14 days
      }
    }
  }
}

If the Book class were to change its synchronization policy in the future, the BookWrapper class's locking strategy might silently break. For instance, the BookWrapper class's locking strategy would break if Book were modified to use a private final lock object, as recommended by LCK00-J. Use private final lock objects to synchronize classes that may interact with untrusted code. This is because threads that call BookWrapper.getDueDate() would perform operations on the thread-safe Book using its new locking policy. However, threads that call the renew() method would always synchronize on the intrinsic lock of the Book instance. Consequently, the implementation would use two different locks.

Compliant Solution (Private Final Lock Object)

This compliant solution uses a private final lock object and synchronizes the methods of the BookWrapper class using this lock:

public final class BookWrapper {
  private final Book book;
  private final Object lock = new Object();

  BookWrapper(Book book) {
    this.book = book;
  }

  public void issue(int days) {
    synchronized(lock) {
      book.issue(days);
    }
  }

  public Calendar getDueDate() {
    synchronized(lock) {
      return book.getDueDate();
    }
  }

  public void renew() {
    synchronized(lock) {
      if (book.getDueDate().before(Calendar.getInstance())) {
        throw new IllegalStateException("Book overdue");
      } else {
        book.issue(14); // Issue book for 14 days
      }
    }
  }
}

The BookWrapper class's locking strategy is now independent of the locking policy of the Book instance.

Noncompliant Code Example (Class Extension and Accessible Member Lock)

Goetz and colleagues describe the fragility of class extension for adding functionality to thread-safe classes [Goetz 2006]:

Extension is more fragile than adding code directly to a class, because the implementation of the synchronization policy is now distributed over multiple, separately maintained source files. If the underlying class were to change its synchronization policy by choosing a different lock to guard its state variables, the subclass would subtly and silently break because it no longer used the right lock to control concurrent access to the base class's state.

In this noncompliant code example, the PrintableIPAddressList class extends the thread-safe IPAddressList class. PrintableIPAddressList locks on IPAddressList.ips in the addAndPrintIPAddresses() method. This is another example of client-side locking because a subclass is using an object owned and locked by its superclass.

// This class could change its locking policy in the future,
// for example, if new non-atomic methods are added
class IPAddressList {
  private final List<InetAddress> ips = 
      Collections.synchronizedList(new ArrayList<InetAddress>());

  public List<InetAddress> getList() {
    return ips; // No defensive copies required
                // as visibility is package-private
  }

  public void addIPAddress(InetAddress address) {
    ips.add(address);
  }
}

class PrintableIPAddressList extends IPAddressList {
  public void addAndPrintIPAddresses(InetAddress address) {
    synchronized (getList()) {
      addIPAddress(address);
      InetAddress[] ia =
          (InetAddress[]) getList().toArray(new InetAddress[0]);
      // ...
    }
  }
}

If the IPAddressList class were modified to use block synchronization on a private final lock object, as recommended by LCK00-J. Use private final lock objects to synchronize classes that may interact with untrusted code, the PrintableIPAddressList subclass would silently break. Moreover, if a wrapper such as Collections.synchronizedList() were used, it would be difficult for a client to determine the type of the class being wrapped to extend it [Goetz 2006].

Compliant Solution (Composition)

This compliant solution wraps an object of the IPAddressList class and provides synchronized accessors to manipulate the state of the object. Composition offers encapsulation benefits, usually with minimal overhead (refer to OBJ02-J. Preserve dependencies in subclasses when changing superclasses for more information on composition).

// Class IPAddressList remains unchanged
class PrintableIPAddressList {
  private final IPAddressList ips;

  public PrintableIPAddressList(IPAddressList list) {
    this.ips = list;
  }

  public synchronized void addIPAddress(InetAddress address) {
    ips.addIPAddress(address);
  }

  public synchronized void addAndPrintIPAddresses(InetAddress address) {
    addIPAddress(address);
    InetAddress[] ia =
        (InetAddress[]) ips.getList().toArray(new InetAddress[0]);
    // ...
  }
}

In this case, composition allows the PrintableIPAddressList class to use its own intrinsic lock independent of the underlying list class's lock. The underlying collection lacks a requirement for thread-safety because the PrintableIPAddressList wrapper prevents direct access to its methods by publishing its own synchronized equivalents. This approach provides consistent locking even when the underlying class changes its locking policy in the future [Goetz 2006].

Risk Assessment

Using client-side locking when the thread-safe class fails to commit to its locking strategy can cause data inconsistencies and deadlock.

Rule

Severity

Likelihood

Remediation Cost

Priority

Level

LCK11-J

Low

Probable

Medium

P4

L3

Bibliography

[API 2014]

Class Collections
Class ConcurrentHashMap<K,V>

Class Vector<E>
Class WeakReference<T>

[JavaThreads 2004]

Section 8.2, "Synchronization and Collection Classes"

[Goetz 2006]

Section 4.4.1, "Client-side Locking"
Section 4.4.2, "Composition"
Section 5.2.1, "ConcurrentHashMap"

[Lee 2009]

"Map & Compound Operation"

 


4 Comments

  1. In order to guarantee serial access, it is critical that all access to the backing list is accomplished through the returned list. It is imperative that the user manually synchronize on the returned list when iterating over it. Failure to follow this advice may result in non-deterministic behavior.

    It sounds like this quote (from Goetz) advises to manually synchronize on the view returned by Collections.synchronizedList(), which is prohibited by LCK04-J. Do not synchronize on a collection view if the backing collection is accessible.

    Maybe I'm misinterpreting the quote, but if so, I won't be the last.

    1. As discussed this is not a violation because the wrapped class is not accessible.

      EDIT: But it is still possible to violate it; so I've added a disclaimer below that quote.

  2. For case:

    final class Outter {
        public synchronized void test() {
        }
    
        private class Inner {
            public void fun() {
                synchronized(Outter.this) {
                    test();
                }
            }
        }
    }

    Does synchronized(Outter.this) belong to client-side locking?

    1. I depends on whether Outter commits to its locking strategy, which would include publicly documenting its locking strategy.  It also depends on whether instances of Inner can escape from Outter.