PageRenderTime 25ms CodeModel.GetById 15ms app.highlight 6ms RepoModel.GetById 1ms app.codeStats 0ms

Plain Text | 367 lines | 304 code | 63 blank | 0 comment | 0 complexity | af4a0ce2ac70a603066c6ecdacf79494 MD5 | raw file
Possible License(s): GPL-2.0, LGPL-2.0, AGPL-1.0
  1Review Checklist for RCU Patches
  4This document contains a checklist for producing and reviewing patches
  5that make use of RCU.  Violating any of the rules listed below will
  6result in the same sorts of problems that leaving out a locking primitive
  7would cause.  This list is based on experiences reviewing such patches
  8over a rather long period of time, but improvements are always welcome!
 100.	Is RCU being applied to a read-mostly situation?  If the data
 11	structure is updated more than about 10% of the time, then you
 12	should strongly consider some other approach, unless detailed
 13	performance measurements show that RCU is nonetheless the right
 14	tool for the job.  Yes, RCU does reduce read-side overhead by
 15	increasing write-side overhead, which is exactly why normal uses
 16	of RCU will do much more reading than updating.
 18	Another exception is where performance is not an issue, and RCU
 19	provides a simpler implementation.  An example of this situation
 20	is the dynamic NMI code in the Linux 2.6 kernel, at least on
 21	architectures where NMIs are rare.
 23	Yet another exception is where the low real-time latency of RCU's
 24	read-side primitives is critically important.
 261.	Does the update code have proper mutual exclusion?
 28	RCU does allow -readers- to run (almost) naked, but -writers- must
 29	still use some sort of mutual exclusion, such as:
 31	a.	locking,
 32	b.	atomic operations, or
 33	c.	restricting updates to a single task.
 35	If you choose #b, be prepared to describe how you have handled
 36	memory barriers on weakly ordered machines (pretty much all of
 37	them -- even x86 allows later loads to be reordered to precede
 38	earlier stores), and be prepared to explain why this added
 39	complexity is worthwhile.  If you choose #c, be prepared to
 40	explain how this single task does not become a major bottleneck on
 41	big multiprocessor machines (for example, if the task is updating
 42	information relating to itself that other tasks can read, there
 43	by definition can be no bottleneck).
 452.	Do the RCU read-side critical sections make proper use of
 46	rcu_read_lock() and friends?  These primitives are needed
 47	to prevent grace periods from ending prematurely, which
 48	could result in data being unceremoniously freed out from
 49	under your read-side code, which can greatly increase the
 50	actuarial risk of your kernel.
 52	As a rough rule of thumb, any dereference of an RCU-protected
 53	pointer must be covered by rcu_read_lock(), rcu_read_lock_bh(),
 54	rcu_read_lock_sched(), or by the appropriate update-side lock.
 55	Disabling of preemption can serve as rcu_read_lock_sched(), but
 56	is less readable.
 583.	Does the update code tolerate concurrent accesses?
 60	The whole point of RCU is to permit readers to run without
 61	any locks or atomic operations.  This means that readers will
 62	be running while updates are in progress.  There are a number
 63	of ways to handle this concurrency, depending on the situation:
 65	a.	Use the RCU variants of the list and hlist update
 66		primitives to add, remove, and replace elements on
 67		an RCU-protected list.	Alternatively, use the other
 68		RCU-protected data structures that have been added to
 69		the Linux kernel.
 71		This is almost always the best approach.
 73	b.	Proceed as in (a) above, but also maintain per-element
 74		locks (that are acquired by both readers and writers)
 75		that guard per-element state.  Of course, fields that
 76		the readers refrain from accessing can be guarded by
 77		some other lock acquired only by updaters, if desired.
 79		This works quite well, also.
 81	c.	Make updates appear atomic to readers.  For example,
 82		pointer updates to properly aligned fields will
 83		appear atomic, as will individual atomic primitives.
 84		Sequences of perations performed under a lock will -not-
 85		appear to be atomic to RCU readers, nor will sequences
 86		of multiple atomic primitives.
 88		This can work, but is starting to get a bit tricky.
 90	d.	Carefully order the updates and the reads so that
 91		readers see valid data at all phases of the update.
 92		This is often more difficult than it sounds, especially
 93		given modern CPUs' tendency to reorder memory references.
 94		One must usually liberally sprinkle memory barriers
 95		(smp_wmb(), smp_rmb(), smp_mb()) through the code,
 96		making it difficult to understand and to test.
 98		It is usually better to group the changing data into
 99		a separate structure, so that the change may be made
100		to appear atomic by updating a pointer to reference
101		a new structure containing updated values.
1034.	Weakly ordered CPUs pose special challenges.  Almost all CPUs
104	are weakly ordered -- even x86 CPUs allow later loads to be
105	reordered to precede earlier stores.  RCU code must take all of
106	the following measures to prevent memory-corruption problems:
108	a.	Readers must maintain proper ordering of their memory
109		accesses.  The rcu_dereference() primitive ensures that
110		the CPU picks up the pointer before it picks up the data
111		that the pointer points to.  This really is necessary
112		on Alpha CPUs.	If you don't believe me, see:
116		The rcu_dereference() primitive is also an excellent
117		documentation aid, letting the person reading the code
118		know exactly which pointers are protected by RCU.
119		Please note that compilers can also reorder code, and
120		they are becoming increasingly aggressive about doing
121		just that.  The rcu_dereference() primitive therefore
122		also prevents destructive compiler optimizations.
124		The rcu_dereference() primitive is used by the
125		various "_rcu()" list-traversal primitives, such
126		as the list_for_each_entry_rcu().  Note that it is
127		perfectly legal (if redundant) for update-side code to
128		use rcu_dereference() and the "_rcu()" list-traversal
129		primitives.  This is particularly useful in code that
130		is common to readers and updaters.  However, lockdep
131		will complain if you access rcu_dereference() outside
132		of an RCU read-side critical section.  See lockdep.txt
133		to learn what to do about this.
135		Of course, neither rcu_dereference() nor the "_rcu()"
136		list-traversal primitives can substitute for a good
137		concurrency design coordinating among multiple updaters.
139	b.	If the list macros are being used, the list_add_tail_rcu()
140		and list_add_rcu() primitives must be used in order
141		to prevent weakly ordered machines from misordering
142		structure initialization and pointer planting.
143		Similarly, if the hlist macros are being used, the
144		hlist_add_head_rcu() primitive is required.
146	c.	If the list macros are being used, the list_del_rcu()
147		primitive must be used to keep list_del()'s pointer
148		poisoning from inflicting toxic effects on concurrent
149		readers.  Similarly, if the hlist macros are being used,
150		the hlist_del_rcu() primitive is required.
152		The list_replace_rcu() and hlist_replace_rcu() primitives
153		may be used to replace an old structure with a new one
154		in their respective types of RCU-protected lists.
156	d.	Rules similar to (4b) and (4c) apply to the "hlist_nulls"
157		type of RCU-protected linked lists.
159	e.	Updates must ensure that initialization of a given
160		structure happens before pointers to that structure are
161		publicized.  Use the rcu_assign_pointer() primitive
162		when publicizing a pointer to a structure that can
163		be traversed by an RCU read-side critical section.
1655.	If call_rcu(), or a related primitive such as call_rcu_bh() or
166	call_rcu_sched(), is used, the callback function must be
167	written to be called from softirq context.  In particular,
168	it cannot block.
1706.	Since synchronize_rcu() can block, it cannot be called from
171	any sort of irq context.  The same rule applies for
172	synchronize_rcu_bh(), synchronize_sched(), synchronize_srcu(),
173	synchronize_rcu_expedited(), synchronize_rcu_bh_expedited(),
174	synchronize_sched_expedite(), and synchronize_srcu_expedited().
176	The expedited forms of these primitives have the same semantics
177	as the non-expedited forms, but expediting is both expensive
178	and unfriendly to real-time workloads.	Use of the expedited
179	primitives should be restricted to rare configuration-change
180	operations that would not normally be undertaken while a real-time
181	workload is running.
1837.	If the updater uses call_rcu() or synchronize_rcu(), then the
184	corresponding readers must use rcu_read_lock() and
185	rcu_read_unlock().  If the updater uses call_rcu_bh() or
186	synchronize_rcu_bh(), then the corresponding readers must
187	use rcu_read_lock_bh() and rcu_read_unlock_bh().  If the
188	updater uses call_rcu_sched() or synchronize_sched(), then
189	the corresponding readers must disable preemption, possibly
190	by calling rcu_read_lock_sched() and rcu_read_unlock_sched().
191	If the updater uses synchronize_srcu(), the the corresponding
192	readers must use srcu_read_lock() and srcu_read_unlock(),
193	and with the same srcu_struct.	The rules for the expedited
194	primitives are the same as for their non-expedited counterparts.
195	Mixing things up will result in confusion and broken kernels.
197	One exception to this rule: rcu_read_lock() and rcu_read_unlock()
198	may be substituted for rcu_read_lock_bh() and rcu_read_unlock_bh()
199	in cases where local bottom halves are already known to be
200	disabled, for example, in irq or softirq context.  Commenting
201	such cases is a must, of course!  And the jury is still out on
202	whether the increased speed is worth it.
2048.	Although synchronize_rcu() is slower than is call_rcu(), it
205	usually results in simpler code.  So, unless update performance
206	is critically important or the updaters cannot block,
207	synchronize_rcu() should be used in preference to call_rcu().
209	An especially important property of the synchronize_rcu()
210	primitive is that it automatically self-limits: if grace periods
211	are delayed for whatever reason, then the synchronize_rcu()
212	primitive will correspondingly delay updates.  In contrast,
213	code using call_rcu() should explicitly limit update rate in
214	cases where grace periods are delayed, as failing to do so can
215	result in excessive realtime latencies or even OOM conditions.
217	Ways of gaining this self-limiting property when using call_rcu()
218	include:
220	a.	Keeping a count of the number of data-structure elements
221		used by the RCU-protected data structure, including those
222		waiting for a grace period to elapse.  Enforce a limit
223		on this number, stalling updates as needed to allow
224		previously deferred frees to complete.
226		Alternatively, limit only the number awaiting deferred
227		free rather than the total number of elements.
229	b.	Limiting update rate.  For example, if updates occur only
230		once per hour, then no explicit rate limiting is required,
231		unless your system is already badly broken.  The dcache
232		subsystem takes this approach -- updates are guarded
233		by a global lock, limiting their rate.
235	c.	Trusted update -- if updates can only be done manually by
236		superuser or some other trusted user, then it might not
237		be necessary to automatically limit them.  The theory
238		here is that superuser already has lots of ways to crash
239		the machine.
241	d.	Use call_rcu_bh() rather than call_rcu(), in order to take
242		advantage of call_rcu_bh()'s faster grace periods.
244	e.	Periodically invoke synchronize_rcu(), permitting a limited
245		number of updates per grace period.
247	The same cautions apply to call_rcu_bh() and call_rcu_sched().
2499.	All RCU list-traversal primitives, which include
250	rcu_dereference(), list_for_each_entry_rcu(),
251	list_for_each_continue_rcu(), and list_for_each_safe_rcu(),
252	must be either within an RCU read-side critical section or
253	must be protected by appropriate update-side locks.  RCU
254	read-side critical sections are delimited by rcu_read_lock()
255	and rcu_read_unlock(), or by similar primitives such as
256	rcu_read_lock_bh() and rcu_read_unlock_bh(), in which case
257	the matching rcu_dereference() primitive must be used in order
258	to keep lockdep happy, in this case, rcu_dereference_bh().
260	The reason that it is permissible to use RCU list-traversal
261	primitives when the update-side lock is held is that doing so
262	can be quite helpful in reducing code bloat when common code is
263	shared between readers and updaters.  Additional primitives
264	are provided for this case, as discussed in lockdep.txt.
26610.	Conversely, if you are in an RCU read-side critical section,
267	and you don't hold the appropriate update-side lock, you -must-
268	use the "_rcu()" variants of the list macros.  Failing to do so
269	will break Alpha, cause aggressive compilers to generate bad code,
270	and confuse people trying to read your code.
27211.	Note that synchronize_rcu() -only- guarantees to wait until
273	all currently executing rcu_read_lock()-protected RCU read-side
274	critical sections complete.  It does -not- necessarily guarantee
275	that all currently running interrupts, NMIs, preempt_disable()
276	code, or idle loops will complete.  Therefore, if you do not have
277	rcu_read_lock()-protected read-side critical sections, do -not-
278	use synchronize_rcu().
280	Similarly, disabling preemption is not an acceptable substitute
281	for rcu_read_lock().  Code that attempts to use preemption
282	disabling where it should be using rcu_read_lock() will break
283	in real-time kernel builds.
285	If you want to wait for interrupt handlers, NMI handlers, and
286	code under the influence of preempt_disable(), you instead
287	need to use synchronize_irq() or synchronize_sched().
28912.	Any lock acquired by an RCU callback must be acquired elsewhere
290	with softirq disabled, e.g., via spin_lock_irqsave(),
291	spin_lock_bh(), etc.  Failing to disable irq on a given
292	acquisition of that lock will result in deadlock as soon as
293	the RCU softirq handler happens to run your RCU callback while
294	interrupting that acquisition's critical section.
29613.	RCU callbacks can be and are executed in parallel.  In many cases,
297	the callback code simply wrappers around kfree(), so that this
298	is not an issue (or, more accurately, to the extent that it is
299	an issue, the memory-allocator locking handles it).  However,
300	if the callbacks do manipulate a shared data structure, they
301	must use whatever locking or other synchronization is required
302	to safely access and/or modify that data structure.
304	RCU callbacks are -usually- executed on the same CPU that executed
305	the corresponding call_rcu(), call_rcu_bh(), or call_rcu_sched(),
306	but are by -no- means guaranteed to be.  For example, if a given
307	CPU goes offline while having an RCU callback pending, then that
308	RCU callback will execute on some surviving CPU.  (If this was
309	not the case, a self-spawning RCU callback would prevent the
310	victim CPU from ever going offline.)
31214.	SRCU (srcu_read_lock(), srcu_read_unlock(), srcu_dereference(),
313	synchronize_srcu(), and synchronize_srcu_expedited()) may only
314	be invoked from process context.  Unlike other forms of RCU, it
315	-is- permissible to block in an SRCU read-side critical section
316	(demarked by srcu_read_lock() and srcu_read_unlock()), hence the
317	"SRCU": "sleepable RCU".  Please note that if you don't need
318	to sleep in read-side critical sections, you should be using
319	RCU rather than SRCU, because RCU is almost always faster and
320	easier to use than is SRCU.
322	Also unlike other forms of RCU, explicit initialization
323	and cleanup is required via init_srcu_struct() and
324	cleanup_srcu_struct().	These are passed a "struct srcu_struct"
325	that defines the scope of a given SRCU domain.	Once initialized,
326	the srcu_struct is passed to srcu_read_lock(), srcu_read_unlock()
327	synchronize_srcu(), and synchronize_srcu_expedited().  A given
328	synchronize_srcu() waits only for SRCU read-side critical
329	sections governed by srcu_read_lock() and srcu_read_unlock()
330	calls that have been passed the same srcu_struct.  This property
331	is what makes sleeping read-side critical sections tolerable --
332	a given subsystem delays only its own updates, not those of other
333	subsystems using SRCU.	Therefore, SRCU is less prone to OOM the
334	system than RCU would be if RCU's read-side critical sections
335	were permitted to sleep.
337	The ability to sleep in read-side critical sections does not
338	come for free.	First, corresponding srcu_read_lock() and
339	srcu_read_unlock() calls must be passed the same srcu_struct.
340	Second, grace-period-detection overhead is amortized only
341	over those updates sharing a given srcu_struct, rather than
342	being globally amortized as they are for other forms of RCU.
343	Therefore, SRCU should be used in preference to rw_semaphore
344	only in extremely read-intensive situations, or in situations
345	requiring SRCU's read-side deadlock immunity or low read-side
346	realtime latency.
348	Note that, rcu_assign_pointer() relates to SRCU just as they do
349	to other forms of RCU.
35115.	The whole point of call_rcu(), synchronize_rcu(), and friends
352	is to wait until all pre-existing readers have finished before
353	carrying out some otherwise-destructive operation.  It is
354	therefore critically important to -first- remove any path
355	that readers can follow that could be affected by the
356	destructive operation, and -only- -then- invoke call_rcu(),
357	synchronize_rcu(), or friends.
359	Because these primitives only wait for pre-existing readers, it
360	is the caller's responsibility to guarantee that any subsequent
361	readers will execute safely.
36316.	The various RCU read-side primitives do -not- necessarily contain
364	memory barriers.  You should therefore plan for the CPU
365	and the compiler to freely reorder code into and out of RCU
366	read-side critical sections.  It is the responsibility of the
367	RCU update-side primitives to deal with this.