Kernel-2.6.32-573.12.1.el6_rcuref

Reference-count design for elements of lists/arrays protected by RCU.

Reference counting on elements of lists which are protected by traditional
reader/writer spinlocks or semaphores are straightforward:

  1.         2.
    

    add() search_and_reference()
    { {
    alloc_object read_lock(&list_lock);
    … search_for_element
    atomic_set(&el->rc, 1); atomic_inc(&el->rc);
    write_lock(&list_lock); …
    add_element read_unlock(&list_lock);
    … …
    write_unlock(&list_lock); }
    }

  2.             4.
    

    release_referenced() delete()
    { {
    … write_lock(&list_lock);
    atomic_dec(&el->rc, relfunc) …
    … delete_element
    } write_unlock(&list_lock);

                      ...
                     if (atomic_dec_and_test(&el->rc))
                         kfree(el);
                     ...
                 }
    

If this list/array is made lock free using RCU as in changing the
write_lock() in add() and delete() to spin_lock() and changing read_lock()
in search_and_reference() to rcu_read_lock(), the atomic_inc() in
search_and_reference() could potentially hold reference to an element which
has already been deleted from the list/array. Use atomic_inc_not_zero()
in this scenario as follows:

  1.             2.
    
    add() search_and_reference()
    { {
    alloc_object rcu_read_lock();
    … search_for_element
    atomic_set(&el->rc, 1); if (!atomic_inc_not_zero(&el->rc)) {
    spin_lock(&list_lock); rcu_read_unlock();
                         return FAIL;
    
    add_element }
    … …
    spin_unlock(&list_lock); rcu_read_unlock();
    } }
  2.             4.
    
    release_referenced() delete()
    { {
    … spin_lock(&list_lock);
    if (atomic_dec_and_test(&el->rc)) …
     call_rcu(&el->head, el_free);       delete_element
    
    … spin_unlock(&list_lock);
    } …
                     if (atomic_dec_and_test(&el->rc))
                         call_rcu(&el->head, el_free);
                     ...
                 }
    

Sometimes, a reference to the element needs to be obtained in the
update (write) stream. In such cases, atomic_inc_not_zero() might be
overkill, since we hold the update-side spinlock. One might instead
use atomic_inc() in such cases.