2 # RCU-related configuration options
9 default y if !PREEMPT && SMP
11 This option selects the RCU implementation that is
12 designed for very large SMP system with hundreds or
13 thousands of CPUs. It also scales down nicely to
20 This option selects the RCU implementation that is
21 designed for very large SMP systems with hundreds or
22 thousands of CPUs, but for which real-time response
23 is also required. It also scales down nicely to
26 Select this option if you are unsure.
30 default y if !PREEMPT && !SMP
32 This option selects the RCU implementation that is
33 designed for UP systems from which real-time response
34 is not required. This option greatly reduces the
35 memory footprint of RCU.
38 bool "Make expert-level adjustments to RCU configuration"
41 This option needs to be enabled if you wish to make
42 expert-level adjustments to RCU configuration. By default,
43 no such adjustments can be made, which has the often-beneficial
44 side-effect of preventing "make oldconfig" from asking you all
45 sorts of detailed questions about how you would like numerous
46 obscure RCU options to be set up.
48 Say Y if you need to make expert-level adjustments to RCU.
50 Say N if you are unsure.
55 This option selects the sleepable version of RCU. This version
56 permits arbitrary sleeping or blocking within RCU read-side critical
61 default y if SRCU && TINY_RCU
63 This option selects the single-CPU non-preemptible version of SRCU.
67 default y if SRCU && !TINY_RCU
69 This option selects the full-fledged version of SRCU.
75 This option enables a task-based RCU implementation that uses
76 only voluntary context switch (not preemption!), idle, and
77 user-mode execution as quiescent states.
79 config RCU_STALL_COMMON
80 def_bool ( TREE_RCU || PREEMPT_RCU )
82 This option enables RCU CPU stall code that is common between
83 the TINY and TREE variants of RCU. The purpose is to allow
84 the tiny variants to disable RCU CPU stall warnings, while
85 making these warnings mandatory for the tree variants.
87 config RCU_NEED_SEGCBLIST
88 def_bool ( TREE_RCU || PREEMPT_RCU || TREE_SRCU )
90 config CONTEXT_TRACKING
93 config CONTEXT_TRACKING_FORCE
94 bool "Force context tracking"
95 depends on CONTEXT_TRACKING
96 default y if !NO_HZ_FULL
98 The major pre-requirement for full dynticks to work is to
99 support the context tracking subsystem. But there are also
100 other dependencies to provide in order to make the full
103 This option stands for testing when an arch implements the
104 context tracking backend but doesn't yet fullfill all the
105 requirements to make the full dynticks feature working.
106 Without the full dynticks, there is no way to test the support
107 for context tracking and the subsystems that rely on it: RCU
108 userspace extended quiescent state and tickless cputime
109 accounting. This option copes with the absence of the full
110 dynticks subsystem by forcing the context tracking on all
113 Say Y only if you're working on the development of an
114 architecture backend for the context tracking.
116 Say N otherwise, this option brings an overhead that you
117 don't want in production.
121 int "Tree-based hierarchical RCU fanout value"
124 depends on (TREE_RCU || PREEMPT_RCU) && RCU_EXPERT
128 This option controls the fanout of hierarchical implementations
129 of RCU, allowing RCU to work efficiently on machines with
130 large numbers of CPUs. This value must be at least the fourth
131 root of NR_CPUS, which allows NR_CPUS to be insanely large.
132 The default value of RCU_FANOUT should be used for production
133 systems, but if you are stress-testing the RCU implementation
134 itself, small RCU_FANOUT values allow you to test large-system
135 code paths on small(er) systems.
137 Select a specific number if testing RCU itself.
138 Take the default if unsure.
140 config RCU_FANOUT_LEAF
141 int "Tree-based hierarchical RCU leaf-level fanout value"
144 depends on (TREE_RCU || PREEMPT_RCU) && RCU_EXPERT
147 This option controls the leaf-level fanout of hierarchical
148 implementations of RCU, and allows trading off cache misses
149 against lock contention. Systems that synchronize their
150 scheduling-clock interrupts for energy-efficiency reasons will
151 want the default because the smaller leaf-level fanout keeps
152 lock contention levels acceptably low. Very large systems
153 (hundreds or thousands of CPUs) will instead want to set this
154 value to the maximum value possible in order to reduce the
155 number of cache misses incurred during RCU's grace-period
156 initialization. These systems tend to run CPU-bound, and thus
157 are not helped by synchronized interrupts, and thus tend to
158 skew them, which reduces lock contention enough that large
159 leaf-level fanouts work well. That said, setting leaf-level
160 fanout to a large number will likely cause problematic
161 lock contention on the leaf-level rcu_node structures unless
162 you boot with the skew_tick kernel parameter.
164 Select a specific number if testing RCU itself.
166 Select the maximum permissible value for large systems, but
167 please understand that you may also need to set the skew_tick
168 kernel boot parameter to avoid contention on the rcu_node
171 Take the default if unsure.
173 config RCU_FAST_NO_HZ
174 bool "Accelerate last non-dyntick-idle CPU's grace periods"
175 depends on NO_HZ_COMMON && SMP && RCU_EXPERT
178 This option permits CPUs to enter dynticks-idle state even if
179 they have RCU callbacks queued, and prevents RCU from waking
180 these CPUs up more than roughly once every four jiffies (by
181 default, you can adjust this using the rcutree.rcu_idle_gp_delay
182 parameter), thus improving energy efficiency. On the other
183 hand, this option increases the duration of RCU grace periods,
184 for example, slowing down synchronize_rcu().
186 Say Y if energy efficiency is critically important, and you
187 don't care about increased grace-period durations.
189 Say N if you are unsure.
192 bool "Enable RCU priority boosting"
193 depends on RT_MUTEXES && PREEMPT_RCU && RCU_EXPERT
196 This option boosts the priority of preempted RCU readers that
197 block the current preemptible RCU grace period for too long.
198 This option also prevents heavy loads from blocking RCU
199 callback invocation for all flavors of RCU.
201 Say Y here if you are working with real-time apps or heavy loads
202 Say N here if you are unsure.
204 config RCU_BOOST_DELAY
205 int "Milliseconds to delay boosting after RCU grace-period start"
210 This option specifies the time to wait after the beginning of
211 a given grace period before priority-boosting preempted RCU
212 readers blocking that grace period. Note that any RCU reader
213 blocking an expedited RCU grace period is boosted immediately.
215 Accept the default if unsure.
218 bool "Offload RCU callback processing from boot-selected CPUs"
219 depends on TREE_RCU || PREEMPT_RCU
220 depends on RCU_EXPERT || NO_HZ_FULL
223 Use this option to reduce OS jitter for aggressive HPC or
224 real-time workloads. It can also be used to offload RCU
225 callback invocation to energy-efficient CPUs in battery-powered
226 asymmetric multiprocessors.
228 This option offloads callback invocation from the set of
229 CPUs specified at boot time by the rcu_nocbs parameter.
230 For each such CPU, a kthread ("rcuox/N") will be created to
231 invoke callbacks, where the "N" is the CPU being offloaded,
232 and where the "x" is "b" for RCU-bh, "p" for RCU-preempt, and
233 "s" for RCU-sched. Nothing prevents this kthread from running
234 on the specified CPUs, but (1) the kthreads may be preempted
235 between each callback, and (2) affinity or cgroups can be used
236 to force the kthreads to run on whatever set of CPUs is desired.
238 Say Y here if you want to help to debug reduced OS jitter.
239 Say N here if you are unsure.
241 endmenu # "RCU Subsystem"