POS47-C. Avoid using threads that can be cancelled asynchronously

In threading, pthreads have the option of being set to cancel immediately or defer until a specific cancellation point. Canceling asynchronously (immediately) is dangerous, however, because most threads are in fact not safe to cancel immediately.

The IEEE standards page states that:

only functions that are cancel-safe may be called from a thread that is asynchronously cancelable.

Canceling asynchronously would follow the same route as passing a signal in to the thread to kill it, thus posing similarities to POS44-C. Do not use signals to terminate threads, which is strongly related to SIG02-C. Avoid using signals to implement normal functionality. These expand on the dangers of canceling a thread suddenly as this can create a data race condition.

Noncompliant Code Example

In this noncompliant code example the worker thread is doing something as simple as swapping a and b repeatedly.

This code uses two locks. The init_mut mutex is used to ensure that the worker thread is initialized safely, and it notifies the main thread once it has set its cancellation policy. And the global_mut mutex is used by both threads to guarantee exclusive access to the global variables a and b. Furthermore, the worker thread uses the cleanup_worker function to clear the global_mut mutex, in case it gets cancelled while the lock is held.

volatile int a, b;

/* Lock to enable threads to access a and b safely */
pthread_mutex_t global_lock = PTHREAD_MUTEX_INITIALIZER;

/* Lock to enable worker thread to initialize safely */
pthread_mutex_t init_lock = PTHREAD_MUTEX_INITIALIZER;
pthread_cond_t init_lock_sig = PTHREAD_COND_INITIALIZER;

void cleanup_worker_thread(void* dummy) {
  int result;
  if ((result = pthread_mutex_unlock(&global_lock)) != 0) {
    /* handle error */
  }
}

void* worker_thread(void* dummy) {
  int i, c;
  int result;

  /* set the cancelability flag during mutex. */
  /* this guarantees this block calls after pthread_cond_wait() and, more importantly, before pthread_cancel() */
  if ((result = pthread_mutex_lock(&init_lock)) != 0) {
    /* handle error */
  }
  if ((result = pthread_setcanceltype(PTHREAD_CANCEL_ASYNCHRONOUS,&i)) != 0) {
    /* handle error */
  }
  pthread_cleanup_push( cleanup_worker_thread, NULL);
  if ((result = pthread_cond_signal(&init_lock_sig)) != 0) {
    /* handle error */
  }
  if ((result = pthread_mutex_unlock(&init_lock)) != 0) {
    /* handle error */
  }

  while (1) {
    if ((result = pthread_mutex_lock(&global_lock)) != 0) {
      /* handle error */
    }
    c = b;
    b = a;
    a = c;
    if ((result = pthread_mutex_unlock(&global_lock)) != 0) {
      /* handle error */
    }

    /* now we're safe to cancel, creating cancel point */
    pthread_testcancel();
  }
  pthread_cleanup_pop( 1);
  return NULL;
}


int main(void) {
  int result;
  a = 5;
  b = 10;
  pthread_t worker;

  if ((result = pthread_mutex_lock(&init_lock)) != 0) {
    /* handle error */
  }
  if ((result = pthread_create( &worker, NULL, worker_thread, NULL)) != 0) {
    /* handle error */
  }

  /* wait until canceltype is set */
  if ((result = pthread_cond_wait( &init_lock_sig, &init_lock)) != 0) {
    /* handle error */
  }
  if ((result = pthread_mutex_unlock(&init_lock)) != 0) {
    /* handle error */
  }

  /* do stuff, like data input */
  getc(stdin);
  /* since we don't know when the character is read in, the program could continue at any time */
  if ((result = pthread_cancel(worker)) != 0) {
    /* handle error */
  }

  /* pthread_join waits for the thread to finish up before continuing */
  if ((result = pthread_join(worker, 0)) != 0) {
    /* handle error */
  }

  if ((result = pthread_mutex_lock(&global_lock)) != 0) {
    /* handle error */
  }
  printf("a: %i | b: %i", a, b);
  if ((result = pthread_mutex_unlock(&global_lock)) != 0) {
    /* handle error */
  }
  /* this will always print either "a: 5 | b: 10" or "a: 10 | b: 5" */

  /* clean up */
  if ((result = pthread_mutex_destroy(&init_lock)) != 0) {
    /* handle error */
  }
  if ((result = pthread_cond_destroy(&init_lock_sig)) != 0) {
    /* handle error */
  }
  return 0;
}

This code is thread-safe in that it invokes no undefined behavior. However, this program can still create a race condition, because an asynchronous cancel can happen at any time. For instance, the worker thread could be cancelled right before the last line (a = c) and thereby lose the old value of b. Consequently the main thread might print that a and b have the same value.

Compliant Solution

From IEEE standards page:

The cancelability state and type of any newly created threads, including the thread in which main() was first invoked, shall be PTHREAD_CANCEL_ENABLE and PTHREAD_CANCEL_DEFERRED respectively.

Since the default condition according to the IEEE standards for POSIX is PTHREAD_CANCEL_DEFERRED, one would simply not set cancel type for the compliant solution.

However, since not all compilers are necessarily guaranteed to follow standards, you should also explicitly call pthread_setcanceltype() with PTHREAD_CANCEL_DEFERRED. All remaining code is identical to the noncompliant example.

void* worker_thread(void* dummy) {
  /* ... */
  if ((result = pthread_setcanceltype(PTHREAD_CANCEL_DEFERRED,&i)) != 0) {
    /* handle error */
  }
  /* ... */
}

Since this code limits cancellation of the worker thread to the end of the while loop, the worker thread can preserve the data invariant that a == b. Consequently, the program might print that a and b are both 5, or they are both 10, but they will always be revealed to have the same value when the worker thread is cancelled.

Risk Assessment

Incorrectly using threads that asynchronously cancel may result in silent corruption, resource leaks and, in the worst case, unpredictable interactions.

Automated Detection

TODO

Related Vulnerabilities

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

Other Languages

In Java, similar reasoning resulted in the deprecation of Thread.stop() and appears in the Java Secure Coding Standard as CON24-J. Do not use Thread.stop() to terminate threads .

References

[[MKS]] pthread_cancel() Man Page
[[Open Group 04]] Threads Overview