linux-hardened/Documentation/RCU/rcuref.txt
Paul E. McKenney d19720a909 [PATCH] RCU documentation fixes (January 2006 update)
Updates to in-tree RCU documentation based on comments over the past few
months.

Signed-off-by: "Paul E. McKenney" <paulmck@us.ibm.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-02-01 08:53:25 -08:00

66 lines
2.2 KiB
Text

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); }
}
3. 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_get 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)) {
write_lock(&list_lock); rcu_read_unlock();
return FAIL;
add_element }
... ...
write_unlock(&list_lock); rcu_read_unlock();
} }
3. 4.
release_referenced() delete()
{ {
... write_lock(&list_lock);
if (atomic_dec_and_test(&el->rc)) ...
call_rcu(&el->head, el_free); delete_element
... write_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.