xref: /linux/kernel/rcu/Kconfig (revision 1a72f4bb6f3eaa5af674cb10802f7064bf71d10a)
1# SPDX-License-Identifier: GPL-2.0-only
2#
3# RCU-related configuration options
4#
5
6menu "RCU Subsystem"
7
8config TREE_RCU
9	bool
10	default y if SMP
11	# Dynticks-idle tracking
12	select CONTEXT_TRACKING_IDLE
13	help
14	  This option selects the RCU implementation that is
15	  designed for very large SMP system with hundreds or
16	  thousands of CPUs.  It also scales down nicely to
17	  smaller systems.
18
19config PREEMPT_RCU
20	bool
21	default y if (PREEMPT || PREEMPT_RT || PREEMPT_DYNAMIC)
22	select TREE_RCU
23	help
24	  This option selects the RCU implementation that is
25	  designed for very large SMP systems with hundreds or
26	  thousands of CPUs, but for which real-time response
27	  is also required.  It also scales down nicely to
28	  smaller systems.
29
30	  Select this option if you are unsure.
31
32config TINY_RCU
33	bool
34	default y if !PREEMPT_RCU && !SMP
35	help
36	  This option selects the RCU implementation that is
37	  designed for UP systems from which real-time response
38	  is not required.  This option greatly reduces the
39	  memory footprint of RCU.
40
41config RCU_EXPERT
42	bool "Make expert-level adjustments to RCU configuration"
43	default n
44	help
45	  This option needs to be enabled if you wish to make
46	  expert-level adjustments to RCU configuration.  By default,
47	  no such adjustments can be made, which has the often-beneficial
48	  side-effect of preventing "make oldconfig" from asking you all
49	  sorts of detailed questions about how you would like numerous
50	  obscure RCU options to be set up.
51
52	  Say Y if you need to make expert-level adjustments to RCU.
53
54	  Say N if you are unsure.
55
56config TINY_SRCU
57	bool
58	default y if TINY_RCU
59	help
60	  This option selects the single-CPU non-preemptible version of SRCU.
61
62config TREE_SRCU
63	bool
64	default y if !TINY_RCU
65	help
66	  This option selects the full-fledged version of SRCU.
67
68config FORCE_NEED_SRCU_NMI_SAFE
69	bool "Force selection of NEED_SRCU_NMI_SAFE"
70	depends on !TINY_SRCU
71	depends on RCU_EXPERT
72	depends on ARCH_HAS_NMI_SAFE_THIS_CPU_OPS
73	select NEED_SRCU_NMI_SAFE
74	default n
75	help
76	  This option forces selection of the NEED_SRCU_NMI_SAFE
77	  Kconfig option, allowing testing of srcu_read_lock_nmisafe()
78	  and srcu_read_unlock_nmisafe() on architectures (like x86)
79	  that select the ARCH_HAS_NMI_SAFE_THIS_CPU_OPS Kconfig option.
80
81config NEED_SRCU_NMI_SAFE
82	def_bool HAVE_NMI && !ARCH_HAS_NMI_SAFE_THIS_CPU_OPS && !TINY_SRCU
83
84config TASKS_RCU_GENERIC
85	def_bool TASKS_RCU || TASKS_RUDE_RCU
86	help
87	  This option enables generic infrastructure code supporting
88	  task-based RCU implementations.  Not for manual selection.
89
90config FORCE_TASKS_RCU
91	bool "Force selection of TASKS_RCU"
92	depends on RCU_EXPERT
93	select TASKS_RCU
94	default n
95	help
96	  This option force-enables a task-based RCU implementation
97	  that uses only voluntary context switch (not preemption!),
98	  idle, and user-mode execution as quiescent states.  Not for
99	  manual selection in most cases.
100
101config NEED_TASKS_RCU
102	bool
103	default n
104
105config TASKS_RCU
106	bool
107	default NEED_TASKS_RCU && PREEMPTION
108	select IRQ_WORK
109
110config FORCE_TASKS_RUDE_RCU
111	bool "Force selection of Tasks Rude RCU"
112	depends on RCU_EXPERT
113	select TASKS_RUDE_RCU
114	default n
115	help
116	  This option force-enables a task-based RCU implementation
117	  that uses only context switch (including preemption) and
118	  user-mode execution as quiescent states.  It forces IPIs and
119	  context switches on all online CPUs, including idle ones,
120	  so use with caution.	Not for manual selection in most cases.
121
122config TASKS_RUDE_RCU
123	bool
124	default n
125	select IRQ_WORK
126
127config FORCE_TASKS_TRACE_RCU
128	bool "Force selection of Tasks Trace RCU"
129	depends on RCU_EXPERT
130	select TASKS_TRACE_RCU
131	default n
132	help
133	  This option enables a task-based RCU implementation that uses
134	  explicit rcu_read_lock_trace() read-side markers, and allows
135	  these readers to appear in the idle loop as well as on the
136	  CPU hotplug code paths.  It can force IPIs on online CPUs,
137	  including idle ones, so use with caution.  Not for manual
138	  selection in most cases.
139
140config TASKS_TRACE_RCU
141	bool
142	default n
143	select IRQ_WORK
144
145config TASKS_TRACE_RCU_NO_MB
146	bool "Override RCU Tasks Trace inclusion of read-side memory barriers"
147	depends on RCU_EXPERT && TASKS_TRACE_RCU
148	default ARCH_WANTS_NO_INSTR
149	help
150	  This option prevents the use of read-side memory barriers in
151	  rcu_read_lock_tasks_trace() and rcu_read_unlock_tasks_trace()
152	  even in kernels built with CONFIG_ARCH_WANTS_NO_INSTR=n, that is,
153	  in kernels that do not have noinstr set up in entry/exit code.
154	  By setting this option, you are promising to carefully review
155	  use of ftrace, BPF, and friends to ensure that no tracing
156	  operation is attached to a function that runs in that portion
157	  of the entry/exit code that RCU does not watch, that is,
158	  where rcu_is_watching() returns false.  Alternatively, you
159	  might choose to never remove traces except by rebooting.
160
161	  Those wishing to disable read-side memory barriers for an entire
162	  architecture can select this Kconfig option, hence the polarity.
163
164	  Say Y here if you need speed and will review use of tracing.
165	  Say N here for certain esoteric testing of RCU itself.
166	  Take the default if you are unsure.
167
168config RCU_STALL_COMMON
169	def_bool TREE_RCU
170	help
171	  This option enables RCU CPU stall code that is common between
172	  the TINY and TREE variants of RCU.  The purpose is to allow
173	  the tiny variants to disable RCU CPU stall warnings, while
174	  making these warnings mandatory for the tree variants.
175
176config RCU_NEED_SEGCBLIST
177	def_bool ( TREE_RCU || TREE_SRCU || TASKS_RCU_GENERIC )
178
179config RCU_FANOUT
180	int "Tree-based hierarchical RCU fanout value"
181	range 2 64 if 64BIT
182	range 2 32 if !64BIT
183	depends on TREE_RCU && RCU_EXPERT
184	default 64 if 64BIT
185	default 32 if !64BIT
186	help
187	  This option controls the fanout of hierarchical implementations
188	  of RCU, allowing RCU to work efficiently on machines with
189	  large numbers of CPUs.  This value must be at least the fourth
190	  root of NR_CPUS, which allows NR_CPUS to be insanely large.
191	  The default value of RCU_FANOUT should be used for production
192	  systems, but if you are stress-testing the RCU implementation
193	  itself, small RCU_FANOUT values allow you to test large-system
194	  code paths on small(er) systems.
195
196	  Select a specific number if testing RCU itself.
197	  Take the default if unsure.
198
199config RCU_FANOUT_LEAF
200	int "Tree-based hierarchical RCU leaf-level fanout value"
201	range 2 64 if 64BIT && !RCU_STRICT_GRACE_PERIOD
202	range 2 32 if !64BIT && !RCU_STRICT_GRACE_PERIOD
203	range 2 3 if RCU_STRICT_GRACE_PERIOD
204	depends on TREE_RCU && RCU_EXPERT
205	default 16 if !RCU_STRICT_GRACE_PERIOD
206	default 2 if RCU_STRICT_GRACE_PERIOD
207	help
208	  This option controls the leaf-level fanout of hierarchical
209	  implementations of RCU, and allows trading off cache misses
210	  against lock contention.  Systems that synchronize their
211	  scheduling-clock interrupts for energy-efficiency reasons will
212	  want the default because the smaller leaf-level fanout keeps
213	  lock contention levels acceptably low.  Very large systems
214	  (hundreds or thousands of CPUs) will instead want to set this
215	  value to the maximum value possible in order to reduce the
216	  number of cache misses incurred during RCU's grace-period
217	  initialization.  These systems tend to run CPU-bound, and thus
218	  are not helped by synchronized interrupts, and thus tend to
219	  skew them, which reduces lock contention enough that large
220	  leaf-level fanouts work well.  That said, setting leaf-level
221	  fanout to a large number will likely cause problematic
222	  lock contention on the leaf-level rcu_node structures unless
223	  you boot with the skew_tick kernel parameter.
224
225	  Select a specific number if testing RCU itself.
226
227	  Select the maximum permissible value for large systems, but
228	  please understand that you may also need to set the skew_tick
229	  kernel boot parameter to avoid contention on the rcu_node
230	  structure's locks.
231
232	  Take the default if unsure.
233
234config RCU_BOOST
235	bool "Enable RCU priority boosting"
236	depends on (RT_MUTEXES && PREEMPT_RCU && RCU_EXPERT) || PREEMPT_RT
237	default y if PREEMPT_RT
238	help
239	  This option boosts the priority of preempted RCU readers that
240	  block the current preemptible RCU grace period for too long.
241	  This option also prevents heavy loads from blocking RCU
242	  callback invocation.
243
244	  Say Y here if you are working with real-time apps or heavy loads
245	  Say N here if you are unsure.
246
247config RCU_BOOST_DELAY
248	int "Milliseconds to delay boosting after RCU grace-period start"
249	range 0 3000
250	depends on RCU_BOOST
251	default 500
252	help
253	  This option specifies the time to wait after the beginning of
254	  a given grace period before priority-boosting preempted RCU
255	  readers blocking that grace period.  Note that any RCU reader
256	  blocking an expedited RCU grace period is boosted immediately.
257
258	  Accept the default if unsure.
259
260config RCU_EXP_KTHREAD
261	bool "Perform RCU expedited work in a real-time kthread"
262	depends on RCU_BOOST && RCU_EXPERT
263	default !PREEMPT_RT && NR_CPUS <= 32
264	help
265	  Use this option to further reduce the latencies of expedited
266	  grace periods at the expense of being more disruptive.
267
268	  This option is disabled by default on PREEMPT_RT=y kernels which
269	  disable expedited grace periods after boot by unconditionally
270	  setting rcupdate.rcu_normal_after_boot=1.
271
272	  Accept the default if unsure.
273
274config RCU_NOCB_CPU
275	bool "Offload RCU callback processing from boot-selected CPUs"
276	depends on TREE_RCU
277	depends on RCU_EXPERT || NO_HZ_FULL
278	default n
279	help
280	  Use this option to reduce OS jitter for aggressive HPC or
281	  real-time workloads.	It can also be used to offload RCU
282	  callback invocation to energy-efficient CPUs in battery-powered
283	  asymmetric multiprocessors.  The price of this reduced jitter
284	  is that the overhead of call_rcu() increases and that some
285	  workloads will incur significant increases in context-switch
286	  rates.
287
288	  This option offloads callback invocation from the set of
289	  CPUs specified at boot time by the rcu_nocbs parameter.
290	  For each such CPU, a kthread ("rcuox/N") will be created to
291	  invoke callbacks, where the "N" is the CPU being offloaded,
292	  and where the "x" is "p" for RCU-preempt (PREEMPTION kernels)
293	  and "s" for RCU-sched (!PREEMPTION kernels).	This option
294	  also creates another kthread for each sqrt(nr_cpu_ids) CPUs
295	  ("rcuog/N", where N is the first CPU in that group to come
296	  online), which handles grace periods for its group.  Nothing
297	  prevents these kthreads from running on the specified CPUs,
298	  but (1) the kthreads may be preempted between each callback,
299	  and (2) affinity or cgroups can be used to force the kthreads
300	  to run on whatever set of CPUs is desired.
301
302	  The sqrt(nr_cpu_ids) grouping may be overridden using the
303	  rcutree.rcu_nocb_gp_stride kernel boot parameter.  This can
304	  be especially helpful for smaller numbers of CPUs, where
305	  sqrt(nr_cpu_ids) can be a bit of a blunt instrument.
306
307	  Say Y here if you need reduced OS jitter, despite added overhead.
308	  Say N here if you are unsure.
309
310config RCU_NOCB_CPU_DEFAULT_ALL
311	bool "Offload RCU callback processing from all CPUs by default"
312	depends on RCU_NOCB_CPU
313	default n
314	help
315	  Use this option to offload callback processing from all CPUs
316	  by default, in the absence of the rcu_nocbs or nohz_full boot
317	  parameter. This also avoids the need to use any boot parameters
318	  to achieve the effect of offloading all CPUs on boot.
319
320	  Say Y here if you want offload all CPUs by default on boot.
321	  Say N here if you are unsure.
322
323config RCU_NOCB_CPU_CB_BOOST
324	bool "Offload RCU callback from real-time kthread"
325	depends on RCU_NOCB_CPU && RCU_BOOST
326	default y if PREEMPT_RT
327	help
328	  Use this option to invoke offloaded callbacks as SCHED_FIFO
329	  to avoid starvation by heavy SCHED_OTHER background load.
330	  Of course, running as SCHED_FIFO during callback floods will
331	  cause the rcuo[ps] kthreads to monopolize the CPU for hundreds
332	  of milliseconds or more.  Therefore, when enabling this option,
333	  it is your responsibility to ensure that latency-sensitive
334	  tasks either run with higher priority or run on some other CPU.
335
336	  Say Y here if you want to set RT priority for offloading kthreads.
337	  Say N here if you are building a !PREEMPT_RT kernel and are unsure.
338
339config RCU_LAZY
340	bool "RCU callback lazy invocation functionality"
341	depends on RCU_NOCB_CPU
342	default n
343	help
344	  To save power, batch RCU callbacks and delay starting the
345	  corresponding grace  period for multiple seconds.  The grace
346	  period will be started after this delay, in case of memory
347	  pressure, or if the corresponding CPU's callback list grows
348	  too large.
349
350	  These delays happen only on rcu_nocbs CPUs, that is, CPUs
351	  whose callbacks have been offloaded.
352
353	  Use the rcutree.enable_rcu_lazy=0 kernel-boot parameter to
354	  globally disable these delays.
355
356config RCU_LAZY_DEFAULT_OFF
357	bool "Turn RCU lazy invocation off by default"
358	depends on RCU_LAZY
359	default n
360	help
361	  Build the kernel with CONFIG_RCU_LAZY=y, but cause the kernel
362	  to boot with these energy-efficiency delays disabled.  Use the
363	  rcutree.enable_rcu_lazy=0 kernel-boot parameter to override
364	  the this option at boot time, thus re-enabling these delays.
365
366config RCU_DOUBLE_CHECK_CB_TIME
367	bool "RCU callback-batch backup time check"
368	depends on RCU_EXPERT
369	default n
370	help
371	  Use this option to provide more precise enforcement of the
372	  rcutree.rcu_resched_ns module parameter in situations where
373	  a single RCU callback might run for hundreds of microseconds,
374	  thus defeating the 32-callback batching used to amortize the
375	  cost of the fine-grained but expensive local_clock() function.
376
377	  This option rounds rcutree.rcu_resched_ns up to the next
378	  jiffy, and overrides the 32-callback batching if this limit
379	  is exceeded.
380
381	  Say Y here if you need tighter callback-limit enforcement.
382	  Say N here if you are unsure.
383
384endmenu # "RCU Subsystem"
385