1.. _whatisrcu_doc:
2
3What is RCU?  --  "Read, Copy, Update"
4======================================
5
6Please note that the "What is RCU?" LWN series is an excellent place
7to start learning about RCU:
8
9| 1.	What is RCU, Fundamentally?  http://lwn.net/Articles/262464/
10| 2.	What is RCU? Part 2: Usage   http://lwn.net/Articles/263130/
11| 3.	RCU part 3: the RCU API      http://lwn.net/Articles/264090/
12| 4.	The RCU API, 2010 Edition    http://lwn.net/Articles/418853/
13| 	2010 Big API Table           http://lwn.net/Articles/419086/
14| 5.	The RCU API, 2014 Edition    http://lwn.net/Articles/609904/
15|	2014 Big API Table           http://lwn.net/Articles/609973/
16
17
18What is RCU?
19
20RCU is a synchronization mechanism that was added to the Linux kernel
21during the 2.5 development effort that is optimized for read-mostly
22situations.  Although RCU is actually quite simple once you understand it,
23getting there can sometimes be a challenge.  Part of the problem is that
24most of the past descriptions of RCU have been written with the mistaken
25assumption that there is "one true way" to describe RCU.  Instead,
26the experience has been that different people must take different paths
27to arrive at an understanding of RCU.  This document provides several
28different paths, as follows:
29
30:ref:`1.	RCU OVERVIEW <1_whatisRCU>`
31
32:ref:`2.	WHAT IS RCU'S CORE API? <2_whatisRCU>`
33
34:ref:`3.	WHAT ARE SOME EXAMPLE USES OF CORE RCU API? <3_whatisRCU>`
35
36:ref:`4.	WHAT IF MY UPDATING THREAD CANNOT BLOCK? <4_whatisRCU>`
37
38:ref:`5.	WHAT ARE SOME SIMPLE IMPLEMENTATIONS OF RCU? <5_whatisRCU>`
39
40:ref:`6.	ANALOGY WITH READER-WRITER LOCKING <6_whatisRCU>`
41
42:ref:`7.	FULL LIST OF RCU APIs <7_whatisRCU>`
43
44:ref:`8.	ANSWERS TO QUICK QUIZZES <8_whatisRCU>`
45
46People who prefer starting with a conceptual overview should focus on
47Section 1, though most readers will profit by reading this section at
48some point.  People who prefer to start with an API that they can then
49experiment with should focus on Section 2.  People who prefer to start
50with example uses should focus on Sections 3 and 4.  People who need to
51understand the RCU implementation should focus on Section 5, then dive
52into the kernel source code.  People who reason best by analogy should
53focus on Section 6.  Section 7 serves as an index to the docbook API
54documentation, and Section 8 is the traditional answer key.
55
56So, start with the section that makes the most sense to you and your
57preferred method of learning.  If you need to know everything about
58everything, feel free to read the whole thing -- but if you are really
59that type of person, you have perused the source code and will therefore
60never need this document anyway.  ;-)
61
62.. _1_whatisRCU:
63
641.  RCU OVERVIEW
65----------------
66
67The basic idea behind RCU is to split updates into "removal" and
68"reclamation" phases.  The removal phase removes references to data items
69within a data structure (possibly by replacing them with references to
70new versions of these data items), and can run concurrently with readers.
71The reason that it is safe to run the removal phase concurrently with
72readers is the semantics of modern CPUs guarantee that readers will see
73either the old or the new version of the data structure rather than a
74partially updated reference.  The reclamation phase does the work of reclaiming
75(e.g., freeing) the data items removed from the data structure during the
76removal phase.  Because reclaiming data items can disrupt any readers
77concurrently referencing those data items, the reclamation phase must
78not start until readers no longer hold references to those data items.
79
80Splitting the update into removal and reclamation phases permits the
81updater to perform the removal phase immediately, and to defer the
82reclamation phase until all readers active during the removal phase have
83completed, either by blocking until they finish or by registering a
84callback that is invoked after they finish.  Only readers that are active
85during the removal phase need be considered, because any reader starting
86after the removal phase will be unable to gain a reference to the removed
87data items, and therefore cannot be disrupted by the reclamation phase.
88
89So the typical RCU update sequence goes something like the following:
90
91a.	Remove pointers to a data structure, so that subsequent
92	readers cannot gain a reference to it.
93
94b.	Wait for all previous readers to complete their RCU read-side
95	critical sections.
96
97c.	At this point, there cannot be any readers who hold references
98	to the data structure, so it now may safely be reclaimed
99	(e.g., kfree()d).
100
101Step (b) above is the key idea underlying RCU's deferred destruction.
102The ability to wait until all readers are done allows RCU readers to
103use much lighter-weight synchronization, in some cases, absolutely no
104synchronization at all.  In contrast, in more conventional lock-based
105schemes, readers must use heavy-weight synchronization in order to
106prevent an updater from deleting the data structure out from under them.
107This is because lock-based updaters typically update data items in place,
108and must therefore exclude readers.  In contrast, RCU-based updaters
109typically take advantage of the fact that writes to single aligned
110pointers are atomic on modern CPUs, allowing atomic insertion, removal,
111and replacement of data items in a linked structure without disrupting
112readers.  Concurrent RCU readers can then continue accessing the old
113versions, and can dispense with the atomic operations, memory barriers,
114and communications cache misses that are so expensive on present-day
115SMP computer systems, even in absence of lock contention.
116
117In the three-step procedure shown above, the updater is performing both
118the removal and the reclamation step, but it is often helpful for an
119entirely different thread to do the reclamation, as is in fact the case
120in the Linux kernel's directory-entry cache (dcache).  Even if the same
121thread performs both the update step (step (a) above) and the reclamation
122step (step (c) above), it is often helpful to think of them separately.
123For example, RCU readers and updaters need not communicate at all,
124but RCU provides implicit low-overhead communication between readers
125and reclaimers, namely, in step (b) above.
126
127So how the heck can a reclaimer tell when a reader is done, given
128that readers are not doing any sort of synchronization operations???
129Read on to learn about how RCU's API makes this easy.
130
131.. _2_whatisRCU:
132
1332.  WHAT IS RCU'S CORE API?
134---------------------------
135
136The core RCU API is quite small:
137
138a.	rcu_read_lock()
139b.	rcu_read_unlock()
140c.	synchronize_rcu() / call_rcu()
141d.	rcu_assign_pointer()
142e.	rcu_dereference()
143
144There are many other members of the RCU API, but the rest can be
145expressed in terms of these five, though most implementations instead
146express synchronize_rcu() in terms of the call_rcu() callback API.
147
148The five core RCU APIs are described below, the other 18 will be enumerated
149later.  See the kernel docbook documentation for more info, or look directly
150at the function header comments.
151
152rcu_read_lock()
153^^^^^^^^^^^^^^^
154	void rcu_read_lock(void);
155
156	Used by a reader to inform the reclaimer that the reader is
157	entering an RCU read-side critical section.  It is illegal
158	to block while in an RCU read-side critical section, though
159	kernels built with CONFIG_PREEMPT_RCU can preempt RCU
160	read-side critical sections.  Any RCU-protected data structure
161	accessed during an RCU read-side critical section is guaranteed to
162	remain unreclaimed for the full duration of that critical section.
163	Reference counts may be used in conjunction with RCU to maintain
164	longer-term references to data structures.
165
166rcu_read_unlock()
167^^^^^^^^^^^^^^^^^
168	void rcu_read_unlock(void);
169
170	Used by a reader to inform the reclaimer that the reader is
171	exiting an RCU read-side critical section.  Note that RCU
172	read-side critical sections may be nested and/or overlapping.
173
174synchronize_rcu()
175^^^^^^^^^^^^^^^^^
176	void synchronize_rcu(void);
177
178	Marks the end of updater code and the beginning of reclaimer
179	code.  It does this by blocking until all pre-existing RCU
180	read-side critical sections on all CPUs have completed.
181	Note that synchronize_rcu() will **not** necessarily wait for
182	any subsequent RCU read-side critical sections to complete.
183	For example, consider the following sequence of events::
184
185	         CPU 0                  CPU 1                 CPU 2
186	     ----------------- ------------------------- ---------------
187	 1.  rcu_read_lock()
188	 2.                    enters synchronize_rcu()
189	 3.                                               rcu_read_lock()
190	 4.  rcu_read_unlock()
191	 5.                     exits synchronize_rcu()
192	 6.                                              rcu_read_unlock()
193
194	To reiterate, synchronize_rcu() waits only for ongoing RCU
195	read-side critical sections to complete, not necessarily for
196	any that begin after synchronize_rcu() is invoked.
197
198	Of course, synchronize_rcu() does not necessarily return
199	**immediately** after the last pre-existing RCU read-side critical
200	section completes.  For one thing, there might well be scheduling
201	delays.  For another thing, many RCU implementations process
202	requests in batches in order to improve efficiencies, which can
203	further delay synchronize_rcu().
204
205	Since synchronize_rcu() is the API that must figure out when
206	readers are done, its implementation is key to RCU.  For RCU
207	to be useful in all but the most read-intensive situations,
208	synchronize_rcu()'s overhead must also be quite small.
209
210	The call_rcu() API is a callback form of synchronize_rcu(),
211	and is described in more detail in a later section.  Instead of
212	blocking, it registers a function and argument which are invoked
213	after all ongoing RCU read-side critical sections have completed.
214	This callback variant is particularly useful in situations where
215	it is illegal to block or where update-side performance is
216	critically important.
217
218	However, the call_rcu() API should not be used lightly, as use
219	of the synchronize_rcu() API generally results in simpler code.
220	In addition, the synchronize_rcu() API has the nice property
221	of automatically limiting update rate should grace periods
222	be delayed.  This property results in system resilience in face
223	of denial-of-service attacks.  Code using call_rcu() should limit
224	update rate in order to gain this same sort of resilience.  See
225	checklist.txt for some approaches to limiting the update rate.
226
227rcu_assign_pointer()
228^^^^^^^^^^^^^^^^^^^^
229	void rcu_assign_pointer(p, typeof(p) v);
230
231	Yes, rcu_assign_pointer() **is** implemented as a macro, though it
232	would be cool to be able to declare a function in this manner.
233	(Compiler experts will no doubt disagree.)
234
235	The updater uses this function to assign a new value to an
236	RCU-protected pointer, in order to safely communicate the change
237	in value from the updater to the reader.  This macro does not
238	evaluate to an rvalue, but it does execute any memory-barrier
239	instructions required for a given CPU architecture.
240
241	Perhaps just as important, it serves to document (1) which
242	pointers are protected by RCU and (2) the point at which a
243	given structure becomes accessible to other CPUs.  That said,
244	rcu_assign_pointer() is most frequently used indirectly, via
245	the _rcu list-manipulation primitives such as list_add_rcu().
246
247rcu_dereference()
248^^^^^^^^^^^^^^^^^
249	typeof(p) rcu_dereference(p);
250
251	Like rcu_assign_pointer(), rcu_dereference() must be implemented
252	as a macro.
253
254	The reader uses rcu_dereference() to fetch an RCU-protected
255	pointer, which returns a value that may then be safely
256	dereferenced.  Note that rcu_dereference() does not actually
257	dereference the pointer, instead, it protects the pointer for
258	later dereferencing.  It also executes any needed memory-barrier
259	instructions for a given CPU architecture.  Currently, only Alpha
260	needs memory barriers within rcu_dereference() -- on other CPUs,
261	it compiles to nothing, not even a compiler directive.
262
263	Common coding practice uses rcu_dereference() to copy an
264	RCU-protected pointer to a local variable, then dereferences
265	this local variable, for example as follows::
266
267		p = rcu_dereference(head.next);
268		return p->data;
269
270	However, in this case, one could just as easily combine these
271	into one statement::
272
273		return rcu_dereference(head.next)->data;
274
275	If you are going to be fetching multiple fields from the
276	RCU-protected structure, using the local variable is of
277	course preferred.  Repeated rcu_dereference() calls look
278	ugly, do not guarantee that the same pointer will be returned
279	if an update happened while in the critical section, and incur
280	unnecessary overhead on Alpha CPUs.
281
282	Note that the value returned by rcu_dereference() is valid
283	only within the enclosing RCU read-side critical section [1]_.
284	For example, the following is **not** legal::
285
286		rcu_read_lock();
287		p = rcu_dereference(head.next);
288		rcu_read_unlock();
289		x = p->address;	/* BUG!!! */
290		rcu_read_lock();
291		y = p->data;	/* BUG!!! */
292		rcu_read_unlock();
293
294	Holding a reference from one RCU read-side critical section
295	to another is just as illegal as holding a reference from
296	one lock-based critical section to another!  Similarly,
297	using a reference outside of the critical section in which
298	it was acquired is just as illegal as doing so with normal
299	locking.
300
301	As with rcu_assign_pointer(), an important function of
302	rcu_dereference() is to document which pointers are protected by
303	RCU, in particular, flagging a pointer that is subject to changing
304	at any time, including immediately after the rcu_dereference().
305	And, again like rcu_assign_pointer(), rcu_dereference() is
306	typically used indirectly, via the _rcu list-manipulation
307	primitives, such as list_for_each_entry_rcu() [2]_.
308
309.. 	[1] The variant rcu_dereference_protected() can be used outside
310	of an RCU read-side critical section as long as the usage is
311	protected by locks acquired by the update-side code.  This variant
312	avoids the lockdep warning that would happen when using (for
313	example) rcu_dereference() without rcu_read_lock() protection.
314	Using rcu_dereference_protected() also has the advantage
315	of permitting compiler optimizations that rcu_dereference()
316	must prohibit.	The rcu_dereference_protected() variant takes
317	a lockdep expression to indicate which locks must be acquired
318	by the caller. If the indicated protection is not provided,
319	a lockdep splat is emitted.  See Documentation/RCU/Design/Requirements/Requirements.rst
320	and the API's code comments for more details and example usage.
321
322.. 	[2] If the list_for_each_entry_rcu() instance might be used by
323	update-side code as well as by RCU readers, then an additional
324	lockdep expression can be added to its list of arguments.
325	For example, given an additional "lock_is_held(&mylock)" argument,
326	the RCU lockdep code would complain only if this instance was
327	invoked outside of an RCU read-side critical section and without
328	the protection of mylock.
329
330The following diagram shows how each API communicates among the
331reader, updater, and reclaimer.
332::
333
334
335	    rcu_assign_pointer()
336	                            +--------+
337	    +---------------------->| reader |---------+
338	    |                       +--------+         |
339	    |                           |              |
340	    |                           |              | Protect:
341	    |                           |              | rcu_read_lock()
342	    |                           |              | rcu_read_unlock()
343	    |        rcu_dereference()  |              |
344	    +---------+                 |              |
345	    | updater |<----------------+              |
346	    +---------+                                V
347	    |                                    +-----------+
348	    +----------------------------------->| reclaimer |
349	                                         +-----------+
350	      Defer:
351	      synchronize_rcu() & call_rcu()
352
353
354The RCU infrastructure observes the time sequence of rcu_read_lock(),
355rcu_read_unlock(), synchronize_rcu(), and call_rcu() invocations in
356order to determine when (1) synchronize_rcu() invocations may return
357to their callers and (2) call_rcu() callbacks may be invoked.  Efficient
358implementations of the RCU infrastructure make heavy use of batching in
359order to amortize their overhead over many uses of the corresponding APIs.
360
361There are at least three flavors of RCU usage in the Linux kernel. The diagram
362above shows the most common one. On the updater side, the rcu_assign_pointer(),
363synchronize_rcu() and call_rcu() primitives used are the same for all three
364flavors. However for protection (on the reader side), the primitives used vary
365depending on the flavor:
366
367a.	rcu_read_lock() / rcu_read_unlock()
368	rcu_dereference()
369
370b.	rcu_read_lock_bh() / rcu_read_unlock_bh()
371	local_bh_disable() / local_bh_enable()
372	rcu_dereference_bh()
373
374c.	rcu_read_lock_sched() / rcu_read_unlock_sched()
375	preempt_disable() / preempt_enable()
376	local_irq_save() / local_irq_restore()
377	hardirq enter / hardirq exit
378	NMI enter / NMI exit
379	rcu_dereference_sched()
380
381These three flavors are used as follows:
382
383a.	RCU applied to normal data structures.
384
385b.	RCU applied to networking data structures that may be subjected
386	to remote denial-of-service attacks.
387
388c.	RCU applied to scheduler and interrupt/NMI-handler tasks.
389
390Again, most uses will be of (a).  The (b) and (c) cases are important
391for specialized uses, but are relatively uncommon.
392
393.. _3_whatisRCU:
394
3953.  WHAT ARE SOME EXAMPLE USES OF CORE RCU API?
396-----------------------------------------------
397
398This section shows a simple use of the core RCU API to protect a
399global pointer to a dynamically allocated structure.  More-typical
400uses of RCU may be found in :ref:`listRCU.rst <list_rcu_doc>`,
401:ref:`arrayRCU.rst <array_rcu_doc>`, and :ref:`NMI-RCU.rst <NMI_rcu_doc>`.
402::
403
404	struct foo {
405		int a;
406		char b;
407		long c;
408	};
409	DEFINE_SPINLOCK(foo_mutex);
410
411	struct foo __rcu *gbl_foo;
412
413	/*
414	 * Create a new struct foo that is the same as the one currently
415	 * pointed to by gbl_foo, except that field "a" is replaced
416	 * with "new_a".  Points gbl_foo to the new structure, and
417	 * frees up the old structure after a grace period.
418	 *
419	 * Uses rcu_assign_pointer() to ensure that concurrent readers
420	 * see the initialized version of the new structure.
421	 *
422	 * Uses synchronize_rcu() to ensure that any readers that might
423	 * have references to the old structure complete before freeing
424	 * the old structure.
425	 */
426	void foo_update_a(int new_a)
427	{
428		struct foo *new_fp;
429		struct foo *old_fp;
430
431		new_fp = kmalloc(sizeof(*new_fp), GFP_KERNEL);
432		spin_lock(&foo_mutex);
433		old_fp = rcu_dereference_protected(gbl_foo, lockdep_is_held(&foo_mutex));
434		*new_fp = *old_fp;
435		new_fp->a = new_a;
436		rcu_assign_pointer(gbl_foo, new_fp);
437		spin_unlock(&foo_mutex);
438		synchronize_rcu();
439		kfree(old_fp);
440	}
441
442	/*
443	 * Return the value of field "a" of the current gbl_foo
444	 * structure.  Use rcu_read_lock() and rcu_read_unlock()
445	 * to ensure that the structure does not get deleted out
446	 * from under us, and use rcu_dereference() to ensure that
447	 * we see the initialized version of the structure (important
448	 * for DEC Alpha and for people reading the code).
449	 */
450	int foo_get_a(void)
451	{
452		int retval;
453
454		rcu_read_lock();
455		retval = rcu_dereference(gbl_foo)->a;
456		rcu_read_unlock();
457		return retval;
458	}
459
460So, to sum up:
461
462-	Use rcu_read_lock() and rcu_read_unlock() to guard RCU
463	read-side critical sections.
464
465-	Within an RCU read-side critical section, use rcu_dereference()
466	to dereference RCU-protected pointers.
467
468-	Use some solid scheme (such as locks or semaphores) to
469	keep concurrent updates from interfering with each other.
470
471-	Use rcu_assign_pointer() to update an RCU-protected pointer.
472	This primitive protects concurrent readers from the updater,
473	**not** concurrent updates from each other!  You therefore still
474	need to use locking (or something similar) to keep concurrent
475	rcu_assign_pointer() primitives from interfering with each other.
476
477-	Use synchronize_rcu() **after** removing a data element from an
478	RCU-protected data structure, but **before** reclaiming/freeing
479	the data element, in order to wait for the completion of all
480	RCU read-side critical sections that might be referencing that
481	data item.
482
483See checklist.txt for additional rules to follow when using RCU.
484And again, more-typical uses of RCU may be found in :ref:`listRCU.rst
485<list_rcu_doc>`, :ref:`arrayRCU.rst <array_rcu_doc>`, and :ref:`NMI-RCU.rst
486<NMI_rcu_doc>`.
487
488.. _4_whatisRCU:
489
4904.  WHAT IF MY UPDATING THREAD CANNOT BLOCK?
491--------------------------------------------
492
493In the example above, foo_update_a() blocks until a grace period elapses.
494This is quite simple, but in some cases one cannot afford to wait so
495long -- there might be other high-priority work to be done.
496
497In such cases, one uses call_rcu() rather than synchronize_rcu().
498The call_rcu() API is as follows::
499
500	void call_rcu(struct rcu_head * head,
501		      void (*func)(struct rcu_head *head));
502
503This function invokes func(head) after a grace period has elapsed.
504This invocation might happen from either softirq or process context,
505so the function is not permitted to block.  The foo struct needs to
506have an rcu_head structure added, perhaps as follows::
507
508	struct foo {
509		int a;
510		char b;
511		long c;
512		struct rcu_head rcu;
513	};
514
515The foo_update_a() function might then be written as follows::
516
517	/*
518	 * Create a new struct foo that is the same as the one currently
519	 * pointed to by gbl_foo, except that field "a" is replaced
520	 * with "new_a".  Points gbl_foo to the new structure, and
521	 * frees up the old structure after a grace period.
522	 *
523	 * Uses rcu_assign_pointer() to ensure that concurrent readers
524	 * see the initialized version of the new structure.
525	 *
526	 * Uses call_rcu() to ensure that any readers that might have
527	 * references to the old structure complete before freeing the
528	 * old structure.
529	 */
530	void foo_update_a(int new_a)
531	{
532		struct foo *new_fp;
533		struct foo *old_fp;
534
535		new_fp = kmalloc(sizeof(*new_fp), GFP_KERNEL);
536		spin_lock(&foo_mutex);
537		old_fp = rcu_dereference_protected(gbl_foo, lockdep_is_held(&foo_mutex));
538		*new_fp = *old_fp;
539		new_fp->a = new_a;
540		rcu_assign_pointer(gbl_foo, new_fp);
541		spin_unlock(&foo_mutex);
542		call_rcu(&old_fp->rcu, foo_reclaim);
543	}
544
545The foo_reclaim() function might appear as follows::
546
547	void foo_reclaim(struct rcu_head *rp)
548	{
549		struct foo *fp = container_of(rp, struct foo, rcu);
550
551		foo_cleanup(fp->a);
552
553		kfree(fp);
554	}
555
556The container_of() primitive is a macro that, given a pointer into a
557struct, the type of the struct, and the pointed-to field within the
558struct, returns a pointer to the beginning of the struct.
559
560The use of call_rcu() permits the caller of foo_update_a() to
561immediately regain control, without needing to worry further about the
562old version of the newly updated element.  It also clearly shows the
563RCU distinction between updater, namely foo_update_a(), and reclaimer,
564namely foo_reclaim().
565
566The summary of advice is the same as for the previous section, except
567that we are now using call_rcu() rather than synchronize_rcu():
568
569-	Use call_rcu() **after** removing a data element from an
570	RCU-protected data structure in order to register a callback
571	function that will be invoked after the completion of all RCU
572	read-side critical sections that might be referencing that
573	data item.
574
575If the callback for call_rcu() is not doing anything more than calling
576kfree() on the structure, you can use kfree_rcu() instead of call_rcu()
577to avoid having to write your own callback::
578
579	kfree_rcu(old_fp, rcu);
580
581Again, see checklist.txt for additional rules governing the use of RCU.
582
583.. _5_whatisRCU:
584
5855.  WHAT ARE SOME SIMPLE IMPLEMENTATIONS OF RCU?
586------------------------------------------------
587
588One of the nice things about RCU is that it has extremely simple "toy"
589implementations that are a good first step towards understanding the
590production-quality implementations in the Linux kernel.  This section
591presents two such "toy" implementations of RCU, one that is implemented
592in terms of familiar locking primitives, and another that more closely
593resembles "classic" RCU.  Both are way too simple for real-world use,
594lacking both functionality and performance.  However, they are useful
595in getting a feel for how RCU works.  See kernel/rcu/update.c for a
596production-quality implementation, and see:
597
598	http://www.rdrop.com/users/paulmck/RCU
599
600for papers describing the Linux kernel RCU implementation.  The OLS'01
601and OLS'02 papers are a good introduction, and the dissertation provides
602more details on the current implementation as of early 2004.
603
604
6055A.  "TOY" IMPLEMENTATION #1: LOCKING
606^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
607This section presents a "toy" RCU implementation that is based on
608familiar locking primitives.  Its overhead makes it a non-starter for
609real-life use, as does its lack of scalability.  It is also unsuitable
610for realtime use, since it allows scheduling latency to "bleed" from
611one read-side critical section to another.  It also assumes recursive
612reader-writer locks:  If you try this with non-recursive locks, and
613you allow nested rcu_read_lock() calls, you can deadlock.
614
615However, it is probably the easiest implementation to relate to, so is
616a good starting point.
617
618It is extremely simple::
619
620	static DEFINE_RWLOCK(rcu_gp_mutex);
621
622	void rcu_read_lock(void)
623	{
624		read_lock(&rcu_gp_mutex);
625	}
626
627	void rcu_read_unlock(void)
628	{
629		read_unlock(&rcu_gp_mutex);
630	}
631
632	void synchronize_rcu(void)
633	{
634		write_lock(&rcu_gp_mutex);
635		smp_mb__after_spinlock();
636		write_unlock(&rcu_gp_mutex);
637	}
638
639[You can ignore rcu_assign_pointer() and rcu_dereference() without missing
640much.  But here are simplified versions anyway.  And whatever you do,
641don't forget about them when submitting patches making use of RCU!]::
642
643	#define rcu_assign_pointer(p, v) \
644	({ \
645		smp_store_release(&(p), (v)); \
646	})
647
648	#define rcu_dereference(p) \
649	({ \
650		typeof(p) _________p1 = READ_ONCE(p); \
651		(_________p1); \
652	})
653
654
655The rcu_read_lock() and rcu_read_unlock() primitive read-acquire
656and release a global reader-writer lock.  The synchronize_rcu()
657primitive write-acquires this same lock, then releases it.  This means
658that once synchronize_rcu() exits, all RCU read-side critical sections
659that were in progress before synchronize_rcu() was called are guaranteed
660to have completed -- there is no way that synchronize_rcu() would have
661been able to write-acquire the lock otherwise.  The smp_mb__after_spinlock()
662promotes synchronize_rcu() to a full memory barrier in compliance with
663the "Memory-Barrier Guarantees" listed in:
664
665	Documentation/RCU/Design/Requirements/Requirements.rst
666
667It is possible to nest rcu_read_lock(), since reader-writer locks may
668be recursively acquired.  Note also that rcu_read_lock() is immune
669from deadlock (an important property of RCU).  The reason for this is
670that the only thing that can block rcu_read_lock() is a synchronize_rcu().
671But synchronize_rcu() does not acquire any locks while holding rcu_gp_mutex,
672so there can be no deadlock cycle.
673
674.. _quiz_1:
675
676Quick Quiz #1:
677		Why is this argument naive?  How could a deadlock
678		occur when using this algorithm in a real-world Linux
679		kernel?  How could this deadlock be avoided?
680
681:ref:`Answers to Quick Quiz <8_whatisRCU>`
682
6835B.  "TOY" EXAMPLE #2: CLASSIC RCU
684^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
685This section presents a "toy" RCU implementation that is based on
686"classic RCU".  It is also short on performance (but only for updates) and
687on features such as hotplug CPU and the ability to run in CONFIG_PREEMPT
688kernels.  The definitions of rcu_dereference() and rcu_assign_pointer()
689are the same as those shown in the preceding section, so they are omitted.
690::
691
692	void rcu_read_lock(void) { }
693
694	void rcu_read_unlock(void) { }
695
696	void synchronize_rcu(void)
697	{
698		int cpu;
699
700		for_each_possible_cpu(cpu)
701			run_on(cpu);
702	}
703
704Note that rcu_read_lock() and rcu_read_unlock() do absolutely nothing.
705This is the great strength of classic RCU in a non-preemptive kernel:
706read-side overhead is precisely zero, at least on non-Alpha CPUs.
707And there is absolutely no way that rcu_read_lock() can possibly
708participate in a deadlock cycle!
709
710The implementation of synchronize_rcu() simply schedules itself on each
711CPU in turn.  The run_on() primitive can be implemented straightforwardly
712in terms of the sched_setaffinity() primitive.  Of course, a somewhat less
713"toy" implementation would restore the affinity upon completion rather
714than just leaving all tasks running on the last CPU, but when I said
715"toy", I meant **toy**!
716
717So how the heck is this supposed to work???
718
719Remember that it is illegal to block while in an RCU read-side critical
720section.  Therefore, if a given CPU executes a context switch, we know
721that it must have completed all preceding RCU read-side critical sections.
722Once **all** CPUs have executed a context switch, then **all** preceding
723RCU read-side critical sections will have completed.
724
725So, suppose that we remove a data item from its structure and then invoke
726synchronize_rcu().  Once synchronize_rcu() returns, we are guaranteed
727that there are no RCU read-side critical sections holding a reference
728to that data item, so we can safely reclaim it.
729
730.. _quiz_2:
731
732Quick Quiz #2:
733		Give an example where Classic RCU's read-side
734		overhead is **negative**.
735
736:ref:`Answers to Quick Quiz <8_whatisRCU>`
737
738.. _quiz_3:
739
740Quick Quiz #3:
741		If it is illegal to block in an RCU read-side
742		critical section, what the heck do you do in
743		PREEMPT_RT, where normal spinlocks can block???
744
745:ref:`Answers to Quick Quiz <8_whatisRCU>`
746
747.. _6_whatisRCU:
748
7496.  ANALOGY WITH READER-WRITER LOCKING
750--------------------------------------
751
752Although RCU can be used in many different ways, a very common use of
753RCU is analogous to reader-writer locking.  The following unified
754diff shows how closely related RCU and reader-writer locking can be.
755::
756
757	@@ -5,5 +5,5 @@ struct el {
758	 	int data;
759	 	/* Other data fields */
760	 };
761	-rwlock_t listmutex;
762	+spinlock_t listmutex;
763	 struct el head;
764
765	@@ -13,15 +14,15 @@
766		struct list_head *lp;
767		struct el *p;
768
769	-	read_lock(&listmutex);
770	-	list_for_each_entry(p, head, lp) {
771	+	rcu_read_lock();
772	+	list_for_each_entry_rcu(p, head, lp) {
773			if (p->key == key) {
774				*result = p->data;
775	-			read_unlock(&listmutex);
776	+			rcu_read_unlock();
777				return 1;
778			}
779		}
780	-	read_unlock(&listmutex);
781	+	rcu_read_unlock();
782		return 0;
783	 }
784
785	@@ -29,15 +30,16 @@
786	 {
787		struct el *p;
788
789	-	write_lock(&listmutex);
790	+	spin_lock(&listmutex);
791		list_for_each_entry(p, head, lp) {
792			if (p->key == key) {
793	-			list_del(&p->list);
794	-			write_unlock(&listmutex);
795	+			list_del_rcu(&p->list);
796	+			spin_unlock(&listmutex);
797	+			synchronize_rcu();
798				kfree(p);
799				return 1;
800			}
801		}
802	-	write_unlock(&listmutex);
803	+	spin_unlock(&listmutex);
804		return 0;
805	 }
806
807Or, for those who prefer a side-by-side listing::
808
809 1 struct el {                          1 struct el {
810 2   struct list_head list;             2   struct list_head list;
811 3   long key;                          3   long key;
812 4   spinlock_t mutex;                  4   spinlock_t mutex;
813 5   int data;                          5   int data;
814 6   /* Other data fields */            6   /* Other data fields */
815 7 };                                   7 };
816 8 rwlock_t listmutex;                  8 spinlock_t listmutex;
817 9 struct el head;                      9 struct el head;
818
819::
820
821  1 int search(long key, int *result)    1 int search(long key, int *result)
822  2 {                                    2 {
823  3   struct list_head *lp;              3   struct list_head *lp;
824  4   struct el *p;                      4   struct el *p;
825  5                                      5
826  6   read_lock(&listmutex);             6   rcu_read_lock();
827  7   list_for_each_entry(p, head, lp) { 7   list_for_each_entry_rcu(p, head, lp) {
828  8     if (p->key == key) {             8     if (p->key == key) {
829  9       *result = p->data;             9       *result = p->data;
830 10       read_unlock(&listmutex);      10       rcu_read_unlock();
831 11       return 1;                     11       return 1;
832 12     }                               12     }
833 13   }                                 13   }
834 14   read_unlock(&listmutex);          14   rcu_read_unlock();
835 15   return 0;                         15   return 0;
836 16 }                                   16 }
837
838::
839
840  1 int delete(long key)                 1 int delete(long key)
841  2 {                                    2 {
842  3   struct el *p;                      3   struct el *p;
843  4                                      4
844  5   write_lock(&listmutex);            5   spin_lock(&listmutex);
845  6   list_for_each_entry(p, head, lp) { 6   list_for_each_entry(p, head, lp) {
846  7     if (p->key == key) {             7     if (p->key == key) {
847  8       list_del(&p->list);            8       list_del_rcu(&p->list);
848  9       write_unlock(&listmutex);      9       spin_unlock(&listmutex);
849                                        10       synchronize_rcu();
850 10       kfree(p);                     11       kfree(p);
851 11       return 1;                     12       return 1;
852 12     }                               13     }
853 13   }                                 14   }
854 14   write_unlock(&listmutex);         15   spin_unlock(&listmutex);
855 15   return 0;                         16   return 0;
856 16 }                                   17 }
857
858Either way, the differences are quite small.  Read-side locking moves
859to rcu_read_lock() and rcu_read_unlock, update-side locking moves from
860a reader-writer lock to a simple spinlock, and a synchronize_rcu()
861precedes the kfree().
862
863However, there is one potential catch: the read-side and update-side
864critical sections can now run concurrently.  In many cases, this will
865not be a problem, but it is necessary to check carefully regardless.
866For example, if multiple independent list updates must be seen as
867a single atomic update, converting to RCU will require special care.
868
869Also, the presence of synchronize_rcu() means that the RCU version of
870delete() can now block.  If this is a problem, there is a callback-based
871mechanism that never blocks, namely call_rcu() or kfree_rcu(), that can
872be used in place of synchronize_rcu().
873
874.. _7_whatisRCU:
875
8767.  FULL LIST OF RCU APIs
877-------------------------
878
879The RCU APIs are documented in docbook-format header comments in the
880Linux-kernel source code, but it helps to have a full list of the
881APIs, since there does not appear to be a way to categorize them
882in docbook.  Here is the list, by category.
883
884RCU list traversal::
885
886	list_entry_rcu
887	list_entry_lockless
888	list_first_entry_rcu
889	list_next_rcu
890	list_for_each_entry_rcu
891	list_for_each_entry_continue_rcu
892	list_for_each_entry_from_rcu
893	list_first_or_null_rcu
894	list_next_or_null_rcu
895	hlist_first_rcu
896	hlist_next_rcu
897	hlist_pprev_rcu
898	hlist_for_each_entry_rcu
899	hlist_for_each_entry_rcu_bh
900	hlist_for_each_entry_from_rcu
901	hlist_for_each_entry_continue_rcu
902	hlist_for_each_entry_continue_rcu_bh
903	hlist_nulls_first_rcu
904	hlist_nulls_for_each_entry_rcu
905	hlist_bl_first_rcu
906	hlist_bl_for_each_entry_rcu
907
908RCU pointer/list update::
909
910	rcu_assign_pointer
911	list_add_rcu
912	list_add_tail_rcu
913	list_del_rcu
914	list_replace_rcu
915	hlist_add_behind_rcu
916	hlist_add_before_rcu
917	hlist_add_head_rcu
918	hlist_add_tail_rcu
919	hlist_del_rcu
920	hlist_del_init_rcu
921	hlist_replace_rcu
922	list_splice_init_rcu
923	list_splice_tail_init_rcu
924	hlist_nulls_del_init_rcu
925	hlist_nulls_del_rcu
926	hlist_nulls_add_head_rcu
927	hlist_bl_add_head_rcu
928	hlist_bl_del_init_rcu
929	hlist_bl_del_rcu
930	hlist_bl_set_first_rcu
931
932RCU::
933
934	Critical sections	Grace period		Barrier
935
936	rcu_read_lock		synchronize_net		rcu_barrier
937	rcu_read_unlock		synchronize_rcu
938	rcu_dereference		synchronize_rcu_expedited
939	rcu_read_lock_held	call_rcu
940	rcu_dereference_check	kfree_rcu
941	rcu_dereference_protected
942
943bh::
944
945	Critical sections	Grace period		Barrier
946
947	rcu_read_lock_bh	call_rcu		rcu_barrier
948	rcu_read_unlock_bh	synchronize_rcu
949	[local_bh_disable]	synchronize_rcu_expedited
950	[and friends]
951	rcu_dereference_bh
952	rcu_dereference_bh_check
953	rcu_dereference_bh_protected
954	rcu_read_lock_bh_held
955
956sched::
957
958	Critical sections	Grace period		Barrier
959
960	rcu_read_lock_sched	call_rcu		rcu_barrier
961	rcu_read_unlock_sched	synchronize_rcu
962	[preempt_disable]	synchronize_rcu_expedited
963	[and friends]
964	rcu_read_lock_sched_notrace
965	rcu_read_unlock_sched_notrace
966	rcu_dereference_sched
967	rcu_dereference_sched_check
968	rcu_dereference_sched_protected
969	rcu_read_lock_sched_held
970
971
972SRCU::
973
974	Critical sections	Grace period		Barrier
975
976	srcu_read_lock		call_srcu		srcu_barrier
977	srcu_read_unlock	synchronize_srcu
978	srcu_dereference	synchronize_srcu_expedited
979	srcu_dereference_check
980	srcu_read_lock_held
981
982SRCU: Initialization/cleanup::
983
984	DEFINE_SRCU
985	DEFINE_STATIC_SRCU
986	init_srcu_struct
987	cleanup_srcu_struct
988
989All: lockdep-checked RCU-protected pointer access::
990
991	rcu_access_pointer
992	rcu_dereference_raw
993	RCU_LOCKDEP_WARN
994	rcu_sleep_check
995	RCU_NONIDLE
996
997See the comment headers in the source code (or the docbook generated
998from them) for more information.
999
1000However, given that there are no fewer than four families of RCU APIs
1001in the Linux kernel, how do you choose which one to use?  The following
1002list can be helpful:
1003
1004a.	Will readers need to block?  If so, you need SRCU.
1005
1006b.	What about the -rt patchset?  If readers would need to block
1007	in an non-rt kernel, you need SRCU.  If readers would block
1008	in a -rt kernel, but not in a non-rt kernel, SRCU is not
1009	necessary.  (The -rt patchset turns spinlocks into sleeplocks,
1010	hence this distinction.)
1011
1012c.	Do you need to treat NMI handlers, hardirq handlers,
1013	and code segments with preemption disabled (whether
1014	via preempt_disable(), local_irq_save(), local_bh_disable(),
1015	or some other mechanism) as if they were explicit RCU readers?
1016	If so, RCU-sched is the only choice that will work for you.
1017
1018d.	Do you need RCU grace periods to complete even in the face
1019	of softirq monopolization of one or more of the CPUs?  For
1020	example, is your code subject to network-based denial-of-service
1021	attacks?  If so, you should disable softirq across your readers,
1022	for example, by using rcu_read_lock_bh().
1023
1024e.	Is your workload too update-intensive for normal use of
1025	RCU, but inappropriate for other synchronization mechanisms?
1026	If so, consider SLAB_TYPESAFE_BY_RCU (which was originally
1027	named SLAB_DESTROY_BY_RCU).  But please be careful!
1028
1029f.	Do you need read-side critical sections that are respected
1030	even though they are in the middle of the idle loop, during
1031	user-mode execution, or on an offlined CPU?  If so, SRCU is the
1032	only choice that will work for you.
1033
1034g.	Otherwise, use RCU.
1035
1036Of course, this all assumes that you have determined that RCU is in fact
1037the right tool for your job.
1038
1039.. _8_whatisRCU:
1040
10418.  ANSWERS TO QUICK QUIZZES
1042----------------------------
1043
1044Quick Quiz #1:
1045		Why is this argument naive?  How could a deadlock
1046		occur when using this algorithm in a real-world Linux
1047		kernel?  [Referring to the lock-based "toy" RCU
1048		algorithm.]
1049
1050Answer:
1051		Consider the following sequence of events:
1052
1053		1.	CPU 0 acquires some unrelated lock, call it
1054			"problematic_lock", disabling irq via
1055			spin_lock_irqsave().
1056
1057		2.	CPU 1 enters synchronize_rcu(), write-acquiring
1058			rcu_gp_mutex.
1059
1060		3.	CPU 0 enters rcu_read_lock(), but must wait
1061			because CPU 1 holds rcu_gp_mutex.
1062
1063		4.	CPU 1 is interrupted, and the irq handler
1064			attempts to acquire problematic_lock.
1065
1066		The system is now deadlocked.
1067
1068		One way to avoid this deadlock is to use an approach like
1069		that of CONFIG_PREEMPT_RT, where all normal spinlocks
1070		become blocking locks, and all irq handlers execute in
1071		the context of special tasks.  In this case, in step 4
1072		above, the irq handler would block, allowing CPU 1 to
1073		release rcu_gp_mutex, avoiding the deadlock.
1074
1075		Even in the absence of deadlock, this RCU implementation
1076		allows latency to "bleed" from readers to other
1077		readers through synchronize_rcu().  To see this,
1078		consider task A in an RCU read-side critical section
1079		(thus read-holding rcu_gp_mutex), task B blocked
1080		attempting to write-acquire rcu_gp_mutex, and
1081		task C blocked in rcu_read_lock() attempting to
1082		read_acquire rcu_gp_mutex.  Task A's RCU read-side
1083		latency is holding up task C, albeit indirectly via
1084		task B.
1085
1086		Realtime RCU implementations therefore use a counter-based
1087		approach where tasks in RCU read-side critical sections
1088		cannot be blocked by tasks executing synchronize_rcu().
1089
1090:ref:`Back to Quick Quiz #1 <quiz_1>`
1091
1092Quick Quiz #2:
1093		Give an example where Classic RCU's read-side
1094		overhead is **negative**.
1095
1096Answer:
1097		Imagine a single-CPU system with a non-CONFIG_PREEMPT
1098		kernel where a routing table is used by process-context
1099		code, but can be updated by irq-context code (for example,
1100		by an "ICMP REDIRECT" packet).	The usual way of handling
1101		this would be to have the process-context code disable
1102		interrupts while searching the routing table.  Use of
1103		RCU allows such interrupt-disabling to be dispensed with.
1104		Thus, without RCU, you pay the cost of disabling interrupts,
1105		and with RCU you don't.
1106
1107		One can argue that the overhead of RCU in this
1108		case is negative with respect to the single-CPU
1109		interrupt-disabling approach.  Others might argue that
1110		the overhead of RCU is merely zero, and that replacing
1111		the positive overhead of the interrupt-disabling scheme
1112		with the zero-overhead RCU scheme does not constitute
1113		negative overhead.
1114
1115		In real life, of course, things are more complex.  But
1116		even the theoretical possibility of negative overhead for
1117		a synchronization primitive is a bit unexpected.  ;-)
1118
1119:ref:`Back to Quick Quiz #2 <quiz_2>`
1120
1121Quick Quiz #3:
1122		If it is illegal to block in an RCU read-side
1123		critical section, what the heck do you do in
1124		PREEMPT_RT, where normal spinlocks can block???
1125
1126Answer:
1127		Just as PREEMPT_RT permits preemption of spinlock
1128		critical sections, it permits preemption of RCU
1129		read-side critical sections.  It also permits
1130		spinlocks blocking while in RCU read-side critical
1131		sections.
1132
1133		Why the apparent inconsistency?  Because it is
1134		possible to use priority boosting to keep the RCU
1135		grace periods short if need be (for example, if running
1136		short of memory).  In contrast, if blocking waiting
1137		for (say) network reception, there is no way to know
1138		what should be boosted.  Especially given that the
1139		process we need to boost might well be a human being
1140		who just went out for a pizza or something.  And although
1141		a computer-operated cattle prod might arouse serious
1142		interest, it might also provoke serious objections.
1143		Besides, how does the computer know what pizza parlor
1144		the human being went to???
1145
1146:ref:`Back to Quick Quiz #3 <quiz_3>`
1147
1148ACKNOWLEDGEMENTS
1149
1150My thanks to the people who helped make this human-readable, including
1151Jon Walpole, Josh Triplett, Serge Hallyn, Suzanne Wood, and Alan Stern.
1152
1153
1154For more information, see http://www.rdrop.com/users/paulmck/RCU.
1155