2020-04-21 20:04:07 +03:00
.. SPDX-License-Identifier: GPL-2.0
====================================================================
Reference-count design for elements of lists/arrays protected by RCU
====================================================================
2005-09-10 00:04:09 +04:00
2014-06-23 22:52:59 +04:00
Please note that the percpu-ref feature is likely your first
stop if you need to combine reference counts and RCU. Please see
include/linux/percpu-refcount.h for more information. However, in
those unusual cases where percpu-ref would consume too much memory,
please read on.
------------------------------------------------------------------------
2006-02-01 14:06:42 +03:00
Reference counting on elements of lists which are protected by traditional
reader/writer spinlocks or semaphores are straightforward:
2005-09-10 00:04:09 +04:00
2020-04-21 20:04:07 +03:00
CODE LISTING A::
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);
if(atomic_dec_and_test(&el->rc)) ...
kfree(el);
... remove_element
} write_unlock(&list_lock);
...
if (atomic_dec_and_test(&el->rc))
kfree(el);
...
}
2005-09-10 00:04:09 +04:00
2006-02-01 14:06:42 +03:00
If this list/array is made lock free using RCU as in changing the
2008-09-10 07:01:07 +04:00
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
2006-02-01 14:06:42 +03:00
has already been deleted from the list/array. Use atomic_inc_not_zero()
in this scenario as follows:
2005-09-10 00:04:09 +04:00
2020-04-21 20:04:07 +03:00
CODE LISTING B::
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();
} }
3. 4.
release_referenced() delete()
{ {
... spin_lock(&list_lock);
if (atomic_dec_and_test(&el->rc)) ...
call_rcu(&el->head, el_free); remove_element
... spin_unlock(&list_lock);
} ...
if (atomic_dec_and_test(&el->rc))
call_rcu(&el->head, el_free);
...
}
2005-09-10 00:04:09 +04:00
2006-02-01 14:06:42 +03:00
Sometimes, a reference to the element needs to be obtained in the
2020-04-21 20:04:07 +03:00
update (write) stream. In such cases, atomic_inc_not_zero() might be
2006-02-01 14:06:42 +03:00
overkill, since we hold the update-side spinlock. One might instead
use atomic_inc() in such cases.
2012-10-28 03:34:51 +04:00
It is not always convenient to deal with "FAIL" in the
search_and_reference() code path. In such cases, the
atomic_dec_and_test() may be moved from delete() to el_free()
as follows:
2020-04-21 20:04:07 +03:00
CODE LISTING C::
1. 2.
add() search_and_reference()
{ {
alloc_object rcu_read_lock();
... search_for_element
atomic_set(&el->rc, 1); atomic_inc(&el->rc);
spin_lock(&list_lock); ...
add_element rcu_read_unlock();
... }
spin_unlock(&list_lock); 4.
} delete()
3. {
release_referenced() spin_lock(&list_lock);
{ ...
... remove_element
if (atomic_dec_and_test(&el->rc)) spin_unlock(&list_lock);
kfree(el); ...
... call_rcu(&el->head, el_free);
} ...
5. }
void el_free(struct rcu_head *rhp)
{
release_referenced();
}
2012-10-28 03:34:51 +04:00
The key point is that the initial reference added by add() is not removed
until after a grace period has elapsed following removal. This means that
search_and_reference() cannot find this element, which means that the value
of el->rc cannot increase. Thus, once it reaches zero, there are no
2020-04-21 20:04:07 +03:00
readers that can or ever will be able to reference the element. The
element can therefore safely be freed. This in turn guarantees that if
2012-10-28 03:34:51 +04:00
any reader finds the element, that reader may safely acquire a reference
without checking the value of the reference counter.
2019-03-29 17:05:55 +03:00
A clear advantage of the RCU-based pattern in listing C over the one
in listing B is that any call to search_and_reference() that locates
a given object will succeed in obtaining a reference to that object,
even given a concurrent invocation of delete() for that same object.
Similarly, a clear advantage of both listings B and C over listing A is
that a call to delete() is not delayed even if there are an arbitrarily
large number of calls to search_and_reference() searching for the same
object that delete() was invoked on. Instead, all that is delayed is
the eventual invocation of kfree(), which is usually not a problem on
modern computer systems, even the small ones.
2012-10-28 03:34:51 +04:00
In cases where delete() can sleep, synchronize_rcu() can be called from
2020-04-21 20:04:07 +03:00
delete(), so that el_free() can be subsumed into delete as follows::
4.
delete()
{
spin_lock(&list_lock);
...
remove_element
spin_unlock(&list_lock);
...
synchronize_rcu();
if (atomic_dec_and_test(&el->rc))
kfree(el);
...
}
2019-03-29 17:05:55 +03:00
As additional examples in the kernel, the pattern in listing C is used by
reference counting of struct pid, while the pattern in listing B is used by
struct posix_acl.