Documentation/RCU/whatisRCU.rst

140 c.	synchronize_rcu() / call_rcu()
146 express synchronize_rcu() in terms of the call_rcu() callback API.
174 synchronize_rcu()
176 	void synchronize_rcu(void);
181 	Note that synchronize_rcu() will **not** necessarily wait for
188 	 2.                    enters synchronize_rcu()
191 	 5.                     exits synchronize_rcu()
194 	To reiterate, synchronize_rcu() waits only for ongoing RCU
196 	any that begin after synchronize_rcu() is invoked.
198 	Of course, synchronize_rcu() does not necessarily return
203 	further delay synchronize_rcu().
205 	Since synchronize_rcu() is the API that must figure out when
208 	synchronize_rcu()'s overhead must also be quite small.
210 	The call_rcu() API is a callback form of synchronize_rcu(),
219 	of the synchronize_rcu() API generally results in simpler code.
220 	In addition, the synchronize_rcu() API has the nice property
351 	      synchronize_rcu() & call_rcu()
355 rcu_read_unlock(), synchronize_rcu(), and call_rcu() invocations in
356 order to determine when (1) synchronize_rcu() invocations may return
363 synchronize_rcu() and call_rcu() primitives used are the same for all three
422 	 * Uses synchronize_rcu() to ensure that any readers that might
438 		synchronize_rcu();
477 -	Use synchronize_rcu() **after** removing a data element from an
497 In such cases, one uses call_rcu() rather than synchronize_rcu().
567 that we are now using call_rcu() rather than synchronize_rcu():
632 	void synchronize_rcu(void)
656 and release a global reader-writer lock.  The synchronize_rcu()
658 that once synchronize_rcu() exits, all RCU read-side critical sections
659 that were in progress before synchronize_rcu() was called are guaranteed
660 to have completed -- there is no way that synchronize_rcu() would have
662 promotes synchronize_rcu() to a full memory barrier in compliance with
670 that the only thing that can block rcu_read_lock() is a synchronize_rcu().
671 But synchronize_rcu() does not acquire any locks while holding rcu_gp_mutex,
696 	void synchronize_rcu(void)
710 The implementation of synchronize_rcu() simply schedules itself on each
726 synchronize_rcu().  Once synchronize_rcu() returns, we are guaranteed
797 	+			synchronize_rcu();
849                                         10       synchronize_rcu();
860 a reader-writer lock to a simple spinlock, and a synchronize_rcu()
869 Also, the presence of synchronize_rcu() means that the RCU version of
872 be used in place of synchronize_rcu().
937 	rcu_read_unlock		synchronize_rcu
948 	rcu_read_unlock_bh	synchronize_rcu
961 	rcu_read_unlock_sched	synchronize_rcu
1057 		2.	CPU 1 enters synchronize_rcu(), write-acquiring
1077 		readers through synchronize_rcu().  To see this,
1088 		cannot be blocked by tasks executing synchronize_rcu().