1 .. SPDX-License-Identifier: GPL-2.0 <<
2 .. include:: <isonum.txt> <<
3 <<
4 .. |struct cpuidle_state| replace:: :c:type:`s 1 .. |struct cpuidle_state| replace:: :c:type:`struct cpuidle_state <cpuidle_state>`
5 .. |cpufreq| replace:: :doc:`CPU Performance S 2 .. |cpufreq| replace:: :doc:`CPU Performance Scaling <cpufreq>`
6 3
7 ======================== 4 ========================
8 CPU Idle Time Management 5 CPU Idle Time Management
9 ======================== 6 ========================
10 7
11 :Copyright: |copy| 2018 Intel Corporation !! 8 ::
12 <<
13 :Author: Rafael J. Wysocki <rafael.j.wysocki@in <<
14 9
>> 10 Copyright (c) 2018 Intel Corp., Rafael J. Wysocki <rafael.j.wysocki@intel.com>
15 11
16 Concepts 12 Concepts
17 ======== 13 ========
18 14
19 Modern processors are generally able to enter 15 Modern processors are generally able to enter states in which the execution of
20 a program is suspended and instructions belong 16 a program is suspended and instructions belonging to it are not fetched from
21 memory or executed. Those states are the *idl 17 memory or executed. Those states are the *idle* states of the processor.
22 18
23 Since part of the processor hardware is not us 19 Since part of the processor hardware is not used in idle states, entering them
24 generally allows power drawn by the processor 20 generally allows power drawn by the processor to be reduced and, in consequence,
25 it is an opportunity to save energy. 21 it is an opportunity to save energy.
26 22
27 CPU idle time management is an energy-efficien 23 CPU idle time management is an energy-efficiency feature concerned about using
28 the idle states of processors for this purpose 24 the idle states of processors for this purpose.
29 25
30 Logical CPUs 26 Logical CPUs
31 ------------ 27 ------------
32 28
33 CPU idle time management operates on CPUs as s 29 CPU idle time management operates on CPUs as seen by the *CPU scheduler* (that
34 is the part of the kernel responsible for the 30 is the part of the kernel responsible for the distribution of computational
35 work in the system). In its view, CPUs are *l 31 work in the system). In its view, CPUs are *logical* units. That is, they need
36 not be separate physical entities and may just 32 not be separate physical entities and may just be interfaces appearing to
37 software as individual single-core processors. 33 software as individual single-core processors. In other words, a CPU is an
38 entity which appears to be fetching instructio 34 entity which appears to be fetching instructions that belong to one sequence
39 (program) from memory and executing them, but 35 (program) from memory and executing them, but it need not work this way
40 physically. Generally, three different cases 36 physically. Generally, three different cases can be consider here.
41 37
42 First, if the whole processor can only follow 38 First, if the whole processor can only follow one sequence of instructions (one
43 program) at a time, it is a CPU. In that case 39 program) at a time, it is a CPU. In that case, if the hardware is asked to
44 enter an idle state, that applies to the proce 40 enter an idle state, that applies to the processor as a whole.
45 41
46 Second, if the processor is multi-core, each c 42 Second, if the processor is multi-core, each core in it is able to follow at
47 least one program at a time. The cores need n 43 least one program at a time. The cores need not be entirely independent of each
48 other (for example, they may share caches), bu 44 other (for example, they may share caches), but still most of the time they
49 work physically in parallel with each other, s 45 work physically in parallel with each other, so if each of them executes only
50 one program, those programs run mostly indepen 46 one program, those programs run mostly independently of each other at the same
51 time. The entire cores are CPUs in that case 47 time. The entire cores are CPUs in that case and if the hardware is asked to
52 enter an idle state, that applies to the core 48 enter an idle state, that applies to the core that asked for it in the first
53 place, but it also may apply to a larger unit 49 place, but it also may apply to a larger unit (say a "package" or a "cluster")
54 that the core belongs to (in fact, it may appl 50 that the core belongs to (in fact, it may apply to an entire hierarchy of larger
55 units containing the core). Namely, if all of 51 units containing the core). Namely, if all of the cores in the larger unit
56 except for one have been put into idle states 52 except for one have been put into idle states at the "core level" and the
57 remaining core asks the processor to enter an 53 remaining core asks the processor to enter an idle state, that may trigger it
58 to put the whole larger unit into an idle stat 54 to put the whole larger unit into an idle state which also will affect the
59 other cores in that unit. 55 other cores in that unit.
60 56
61 Finally, each core in a multi-core processor m 57 Finally, each core in a multi-core processor may be able to follow more than one
62 program in the same time frame (that is, each 58 program in the same time frame (that is, each core may be able to fetch
63 instructions from multiple locations in memory 59 instructions from multiple locations in memory and execute them in the same time
64 frame, but not necessarily entirely in paralle 60 frame, but not necessarily entirely in parallel with each other). In that case
65 the cores present themselves to software as "b 61 the cores present themselves to software as "bundles" each consisting of
66 multiple individual single-core "processors", 62 multiple individual single-core "processors", referred to as *hardware threads*
67 (or hyper-threads specifically on Intel hardwa 63 (or hyper-threads specifically on Intel hardware), that each can follow one
68 sequence of instructions. Then, the hardware 64 sequence of instructions. Then, the hardware threads are CPUs from the CPU idle
69 time management perspective and if the process 65 time management perspective and if the processor is asked to enter an idle state
70 by one of them, the hardware thread (or CPU) t 66 by one of them, the hardware thread (or CPU) that asked for it is stopped, but
71 nothing more happens, unless all of the other 67 nothing more happens, unless all of the other hardware threads within the same
72 core also have asked the processor to enter an 68 core also have asked the processor to enter an idle state. In that situation,
73 the core may be put into an idle state individ 69 the core may be put into an idle state individually or a larger unit containing
74 it may be put into an idle state as a whole (i 70 it may be put into an idle state as a whole (if the other cores within the
75 larger unit are in idle states already). 71 larger unit are in idle states already).
76 72
77 Idle CPUs 73 Idle CPUs
78 --------- 74 ---------
79 75
80 Logical CPUs, simply referred to as "CPUs" in 76 Logical CPUs, simply referred to as "CPUs" in what follows, are regarded as
81 *idle* by the Linux kernel when there are no t 77 *idle* by the Linux kernel when there are no tasks to run on them except for the
82 special "idle" task. 78 special "idle" task.
83 79
84 Tasks are the CPU scheduler's representation o 80 Tasks are the CPU scheduler's representation of work. Each task consists of a
85 sequence of instructions to execute, or code, 81 sequence of instructions to execute, or code, data to be manipulated while
86 running that code, and some context informatio 82 running that code, and some context information that needs to be loaded into the
87 processor every time the task's code is run by 83 processor every time the task's code is run by a CPU. The CPU scheduler
88 distributes work by assigning tasks to run to 84 distributes work by assigning tasks to run to the CPUs present in the system.
89 85
90 Tasks can be in various states. In particular 86 Tasks can be in various states. In particular, they are *runnable* if there are
91 no specific conditions preventing their code f 87 no specific conditions preventing their code from being run by a CPU as long as
92 there is a CPU available for that (for example 88 there is a CPU available for that (for example, they are not waiting for any
93 events to occur or similar). When a task beco 89 events to occur or similar). When a task becomes runnable, the CPU scheduler
94 assigns it to one of the available CPUs to run 90 assigns it to one of the available CPUs to run and if there are no more runnable
95 tasks assigned to it, the CPU will load the gi 91 tasks assigned to it, the CPU will load the given task's context and run its
96 code (from the instruction following the last 92 code (from the instruction following the last one executed so far, possibly by
97 another CPU). [If there are multiple runnable 93 another CPU). [If there are multiple runnable tasks assigned to one CPU
98 simultaneously, they will be subject to priori 94 simultaneously, they will be subject to prioritization and time sharing in order
99 to allow them to make some progress over time. 95 to allow them to make some progress over time.]
100 96
101 The special "idle" task becomes runnable if th 97 The special "idle" task becomes runnable if there are no other runnable tasks
102 assigned to the given CPU and the CPU is then 98 assigned to the given CPU and the CPU is then regarded as idle. In other words,
103 in Linux idle CPUs run the code of the "idle" 99 in Linux idle CPUs run the code of the "idle" task called *the idle loop*. That
104 code may cause the processor to be put into on 100 code may cause the processor to be put into one of its idle states, if they are
105 supported, in order to save energy, but if the 101 supported, in order to save energy, but if the processor does not support any
106 idle states, or there is not enough time to sp 102 idle states, or there is not enough time to spend in an idle state before the
107 next wakeup event, or there are strict latency 103 next wakeup event, or there are strict latency constraints preventing any of the
108 available idle states from being used, the CPU 104 available idle states from being used, the CPU will simply execute more or less
109 useless instructions in a loop until it is ass 105 useless instructions in a loop until it is assigned a new task to run.
110 106
111 107
112 .. _idle-loop: 108 .. _idle-loop:
113 109
114 The Idle Loop 110 The Idle Loop
115 ============= 111 =============
116 112
117 The idle loop code takes two major steps in ev 113 The idle loop code takes two major steps in every iteration of it. First, it
118 calls into a code module referred to as the *g 114 calls into a code module referred to as the *governor* that belongs to the CPU
119 idle time management subsystem called ``CPUIdl 115 idle time management subsystem called ``CPUIdle`` to select an idle state for
120 the CPU to ask the hardware to enter. Second, 116 the CPU to ask the hardware to enter. Second, it invokes another code module
121 from the ``CPUIdle`` subsystem, called the *dr 117 from the ``CPUIdle`` subsystem, called the *driver*, to actually ask the
122 processor hardware to enter the idle state sel 118 processor hardware to enter the idle state selected by the governor.
123 119
124 The role of the governor is to find an idle st 120 The role of the governor is to find an idle state most suitable for the
125 conditions at hand. For this purpose, idle st 121 conditions at hand. For this purpose, idle states that the hardware can be
126 asked to enter by logical CPUs are represented 122 asked to enter by logical CPUs are represented in an abstract way independent of
127 the platform or the processor architecture and 123 the platform or the processor architecture and organized in a one-dimensional
128 (linear) array. That array has to be prepared 124 (linear) array. That array has to be prepared and supplied by the ``CPUIdle``
129 driver matching the platform the kernel is run 125 driver matching the platform the kernel is running on at the initialization
130 time. This allows ``CPUIdle`` governors to be 126 time. This allows ``CPUIdle`` governors to be independent of the underlying
131 hardware and to work with any platforms that t 127 hardware and to work with any platforms that the Linux kernel can run on.
132 128
133 Each idle state present in that array is chara 129 Each idle state present in that array is characterized by two parameters to be
134 taken into account by the governor, the *targe 130 taken into account by the governor, the *target residency* and the (worst-case)
135 *exit latency*. The target residency is the m 131 *exit latency*. The target residency is the minimum time the hardware must
136 spend in the given state, including the time n 132 spend in the given state, including the time needed to enter it (which may be
137 substantial), in order to save more energy tha 133 substantial), in order to save more energy than it would save by entering one of
138 the shallower idle states instead. [The "dept 134 the shallower idle states instead. [The "depth" of an idle state roughly
139 corresponds to the power drawn by the processo 135 corresponds to the power drawn by the processor in that state.] The exit
140 latency, in turn, is the maximum time it will 136 latency, in turn, is the maximum time it will take a CPU asking the processor
141 hardware to enter an idle state to start execu 137 hardware to enter an idle state to start executing the first instruction after a
142 wakeup from that state. Note that in general 138 wakeup from that state. Note that in general the exit latency also must cover
143 the time needed to enter the given state in ca 139 the time needed to enter the given state in case the wakeup occurs when the
144 hardware is entering it and it must be entered 140 hardware is entering it and it must be entered completely to be exited in an
145 ordered manner. 141 ordered manner.
146 142
147 There are two types of information that can in 143 There are two types of information that can influence the governor's decisions.
148 First of all, the governor knows the time unti 144 First of all, the governor knows the time until the closest timer event. That
149 time is known exactly, because the kernel prog 145 time is known exactly, because the kernel programs timers and it knows exactly
150 when they will trigger, and it is the maximum 146 when they will trigger, and it is the maximum time the hardware that the given
151 CPU depends on can spend in an idle state, inc 147 CPU depends on can spend in an idle state, including the time necessary to enter
152 and exit it. However, the CPU may be woken up 148 and exit it. However, the CPU may be woken up by a non-timer event at any time
153 (in particular, before the closest timer trigg 149 (in particular, before the closest timer triggers) and it generally is not known
154 when that may happen. The governor can only s 150 when that may happen. The governor can only see how much time the CPU actually
155 was idle after it has been woken up (that time 151 was idle after it has been woken up (that time will be referred to as the *idle
156 duration* from now on) and it can use that inf 152 duration* from now on) and it can use that information somehow along with the
157 time until the closest timer to estimate the i 153 time until the closest timer to estimate the idle duration in future. How the
158 governor uses that information depends on what 154 governor uses that information depends on what algorithm is implemented by it
159 and that is the primary reason for having more 155 and that is the primary reason for having more than one governor in the
160 ``CPUIdle`` subsystem. 156 ``CPUIdle`` subsystem.
161 157
162 There are four ``CPUIdle`` governors available !! 158 There are three ``CPUIdle`` governors available, ``menu``, `TEO <teo-gov_>`_
163 ``ladder`` and ``haltpoll``. Which of them is !! 159 and ``ladder``. Which of them is used by default depends on the configuration
164 configuration of the kernel and in particular !! 160 of the kernel and in particular on whether or not the scheduler tick can be
165 tick can be `stopped by the idle loop <idle-cp !! 161 `stopped by the idle loop <idle-cpus-and-tick_>`_. It is possible to change the
166 governors can be read from the :file:`availabl !! 162 governor at run time if the ``cpuidle_sysfs_switch`` command line parameter has
167 can be changed at runtime. The name of the `` !! 163 been passed to the kernel, but that is not safe in general, so it should not be
168 used by the kernel can be read from the :file: !! 164 done on production systems (that may change in the future, though). The name of
169 :file:`current_governor` file under :file:`/sy !! 165 the ``CPUIdle`` governor currently used by the kernel can be read from the
170 in ``sysfs``. !! 166 :file:`current_governor_ro` (or :file:`current_governor` if
>> 167 ``cpuidle_sysfs_switch`` is present in the kernel command line) file under
>> 168 :file:`/sys/devices/system/cpu/cpuidle/` in ``sysfs``.
171 169
172 Which ``CPUIdle`` driver is used, on the other 170 Which ``CPUIdle`` driver is used, on the other hand, usually depends on the
173 platform the kernel is running on, but there a 171 platform the kernel is running on, but there are platforms with more than one
174 matching driver. For example, there are two d 172 matching driver. For example, there are two drivers that can work with the
175 majority of Intel platforms, ``intel_idle`` an 173 majority of Intel platforms, ``intel_idle`` and ``acpi_idle``, one with
176 hardcoded idle states information and the othe 174 hardcoded idle states information and the other able to read that information
177 from the system's ACPI tables, respectively. 175 from the system's ACPI tables, respectively. Still, even in those cases, the
178 driver chosen at the system initialization tim 176 driver chosen at the system initialization time cannot be replaced later, so the
179 decision on which one of them to use has to be 177 decision on which one of them to use has to be made early (on Intel platforms
180 the ``acpi_idle`` driver will be used if ``int 178 the ``acpi_idle`` driver will be used if ``intel_idle`` is disabled for some
181 reason or if it does not recognize the process 179 reason or if it does not recognize the processor). The name of the ``CPUIdle``
182 driver currently used by the kernel can be rea 180 driver currently used by the kernel can be read from the :file:`current_driver`
183 file under :file:`/sys/devices/system/cpu/cpui 181 file under :file:`/sys/devices/system/cpu/cpuidle/` in ``sysfs``.
184 182
185 183
186 .. _idle-cpus-and-tick: 184 .. _idle-cpus-and-tick:
187 185
188 Idle CPUs and The Scheduler Tick 186 Idle CPUs and The Scheduler Tick
189 ================================ 187 ================================
190 188
191 The scheduler tick is a timer that triggers pe 189 The scheduler tick is a timer that triggers periodically in order to implement
192 the time sharing strategy of the CPU scheduler 190 the time sharing strategy of the CPU scheduler. Of course, if there are
193 multiple runnable tasks assigned to one CPU at 191 multiple runnable tasks assigned to one CPU at the same time, the only way to
194 allow them to make reasonable progress in a gi 192 allow them to make reasonable progress in a given time frame is to make them
195 share the available CPU time. Namely, in roug 193 share the available CPU time. Namely, in rough approximation, each task is
196 given a slice of the CPU time to run its code, 194 given a slice of the CPU time to run its code, subject to the scheduling class,
197 prioritization and so on and when that time sl 195 prioritization and so on and when that time slice is used up, the CPU should be
198 switched over to running (the code of) another 196 switched over to running (the code of) another task. The currently running task
199 may not want to give the CPU away voluntarily, 197 may not want to give the CPU away voluntarily, however, and the scheduler tick
200 is there to make the switch happen regardless. 198 is there to make the switch happen regardless. That is not the only role of the
201 tick, but it is the primary reason for using i 199 tick, but it is the primary reason for using it.
202 200
203 The scheduler tick is problematic from the CPU 201 The scheduler tick is problematic from the CPU idle time management perspective,
204 because it triggers periodically and relativel 202 because it triggers periodically and relatively often (depending on the kernel
205 configuration, the length of the tick period i 203 configuration, the length of the tick period is between 1 ms and 10 ms).
206 Thus, if the tick is allowed to trigger on idl 204 Thus, if the tick is allowed to trigger on idle CPUs, it will not make sense
207 for them to ask the hardware to enter idle sta 205 for them to ask the hardware to enter idle states with target residencies above
208 the tick period length. Moreover, in that cas 206 the tick period length. Moreover, in that case the idle duration of any CPU
209 will never exceed the tick period length and t 207 will never exceed the tick period length and the energy used for entering and
210 exiting idle states due to the tick wakeups on 208 exiting idle states due to the tick wakeups on idle CPUs will be wasted.
211 209
212 Fortunately, it is not really necessary to all 210 Fortunately, it is not really necessary to allow the tick to trigger on idle
213 CPUs, because (by definition) they have no tas 211 CPUs, because (by definition) they have no tasks to run except for the special
214 "idle" one. In other words, from the CPU sche 212 "idle" one. In other words, from the CPU scheduler perspective, the only user
215 of the CPU time on them is the idle loop. Sin 213 of the CPU time on them is the idle loop. Since the time of an idle CPU need
216 not be shared between multiple runnable tasks, 214 not be shared between multiple runnable tasks, the primary reason for using the
217 tick goes away if the given CPU is idle. Cons 215 tick goes away if the given CPU is idle. Consequently, it is possible to stop
218 the scheduler tick entirely on idle CPUs in pr 216 the scheduler tick entirely on idle CPUs in principle, even though that may not
219 always be worth the effort. 217 always be worth the effort.
220 218
221 Whether or not it makes sense to stop the sche 219 Whether or not it makes sense to stop the scheduler tick in the idle loop
222 depends on what is expected by the governor. 220 depends on what is expected by the governor. First, if there is another
223 (non-tick) timer due to trigger within the tic 221 (non-tick) timer due to trigger within the tick range, stopping the tick clearly
224 would be a waste of time, even though the time 222 would be a waste of time, even though the timer hardware may not need to be
225 reprogrammed in that case. Second, if the gov 223 reprogrammed in that case. Second, if the governor is expecting a non-timer
226 wakeup within the tick range, stopping the tic 224 wakeup within the tick range, stopping the tick is not necessary and it may even
227 be harmful. Namely, in that case the governor 225 be harmful. Namely, in that case the governor will select an idle state with
228 the target residency within the time until the 226 the target residency within the time until the expected wakeup, so that state is
229 going to be relatively shallow. The governor 227 going to be relatively shallow. The governor really cannot select a deep idle
230 state then, as that would contradict its own e 228 state then, as that would contradict its own expectation of a wakeup in short
231 order. Now, if the wakeup really occurs short 229 order. Now, if the wakeup really occurs shortly, stopping the tick would be a
232 waste of time and in this case the timer hardw 230 waste of time and in this case the timer hardware would need to be reprogrammed,
233 which is expensive. On the other hand, if the 231 which is expensive. On the other hand, if the tick is stopped and the wakeup
234 does not occur any time soon, the hardware may 232 does not occur any time soon, the hardware may spend indefinite amount of time
235 in the shallow idle state selected by the gove 233 in the shallow idle state selected by the governor, which will be a waste of
236 energy. Hence, if the governor is expecting a 234 energy. Hence, if the governor is expecting a wakeup of any kind within the
237 tick range, it is better to allow the tick tri 235 tick range, it is better to allow the tick trigger. Otherwise, however, the
238 governor will select a relatively deep idle st 236 governor will select a relatively deep idle state, so the tick should be stopped
239 so that it does not wake up the CPU too early. 237 so that it does not wake up the CPU too early.
240 238
241 In any case, the governor knows what it is exp 239 In any case, the governor knows what it is expecting and the decision on whether
242 or not to stop the scheduler tick belongs to i 240 or not to stop the scheduler tick belongs to it. Still, if the tick has been
243 stopped already (in one of the previous iterat 241 stopped already (in one of the previous iterations of the loop), it is better
244 to leave it as is and the governor needs to ta 242 to leave it as is and the governor needs to take that into account.
245 243
246 The kernel can be configured to disable stoppi 244 The kernel can be configured to disable stopping the scheduler tick in the idle
247 loop altogether. That can be done through the 245 loop altogether. That can be done through the build-time configuration of it
248 (by unsetting the ``CONFIG_NO_HZ_IDLE`` config 246 (by unsetting the ``CONFIG_NO_HZ_IDLE`` configuration option) or by passing
249 ``nohz=off`` to it in the command line. In bo 247 ``nohz=off`` to it in the command line. In both cases, as the stopping of the
250 scheduler tick is disabled, the governor's dec 248 scheduler tick is disabled, the governor's decisions regarding it are simply
251 ignored by the idle loop code and the tick is 249 ignored by the idle loop code and the tick is never stopped.
252 250
253 The systems that run kernels configured to all 251 The systems that run kernels configured to allow the scheduler tick to be
254 stopped on idle CPUs are referred to as *tickl 252 stopped on idle CPUs are referred to as *tickless* systems and they are
255 generally regarded as more energy-efficient th 253 generally regarded as more energy-efficient than the systems running kernels in
256 which the tick cannot be stopped. If the give 254 which the tick cannot be stopped. If the given system is tickless, it will use
257 the ``menu`` governor by default and if it is 255 the ``menu`` governor by default and if it is not tickless, the default
258 ``CPUIdle`` governor on it will be ``ladder``. 256 ``CPUIdle`` governor on it will be ``ladder``.
259 257
260 258
261 .. _menu-gov: 259 .. _menu-gov:
262 260
263 The ``menu`` Governor 261 The ``menu`` Governor
264 ===================== 262 =====================
265 263
266 The ``menu`` governor is the default ``CPUIdle 264 The ``menu`` governor is the default ``CPUIdle`` governor for tickless systems.
267 It is quite complex, but the basic principle o 265 It is quite complex, but the basic principle of its design is straightforward.
268 Namely, when invoked to select an idle state f 266 Namely, when invoked to select an idle state for a CPU (i.e. an idle state that
269 the CPU will ask the processor hardware to ent 267 the CPU will ask the processor hardware to enter), it attempts to predict the
270 idle duration and uses the predicted value for 268 idle duration and uses the predicted value for idle state selection.
271 269
272 It first obtains the time until the closest ti 270 It first obtains the time until the closest timer event with the assumption
273 that the scheduler tick will be stopped. That 271 that the scheduler tick will be stopped. That time, referred to as the *sleep
274 length* in what follows, is the upper bound on 272 length* in what follows, is the upper bound on the time before the next CPU
275 wakeup. It is used to determine the sleep len 273 wakeup. It is used to determine the sleep length range, which in turn is needed
276 to get the sleep length correction factor. 274 to get the sleep length correction factor.
277 275
278 The ``menu`` governor maintains two arrays of 276 The ``menu`` governor maintains two arrays of sleep length correction factors.
279 One of them is used when tasks previously runn 277 One of them is used when tasks previously running on the given CPU are waiting
280 for some I/O operations to complete and the ot 278 for some I/O operations to complete and the other one is used when that is not
281 the case. Each array contains several correct 279 the case. Each array contains several correction factor values that correspond
282 to different sleep length ranges organized so 280 to different sleep length ranges organized so that each range represented in the
283 array is approximately 10 times wider than the 281 array is approximately 10 times wider than the previous one.
284 282
285 The correction factor for the given sleep leng 283 The correction factor for the given sleep length range (determined before
286 selecting the idle state for the CPU) is updat 284 selecting the idle state for the CPU) is updated after the CPU has been woken
287 up and the closer the sleep length is to the o 285 up and the closer the sleep length is to the observed idle duration, the closer
288 to 1 the correction factor becomes (it must fa 286 to 1 the correction factor becomes (it must fall between 0 and 1 inclusive).
289 The sleep length is multiplied by the correcti 287 The sleep length is multiplied by the correction factor for the range that it
290 falls into to obtain the first approximation o 288 falls into to obtain the first approximation of the predicted idle duration.
291 289
292 Next, the governor uses a simple pattern recog 290 Next, the governor uses a simple pattern recognition algorithm to refine its
293 idle duration prediction. Namely, it saves th 291 idle duration prediction. Namely, it saves the last 8 observed idle duration
294 values and, when predicting the idle duration 292 values and, when predicting the idle duration next time, it computes the average
295 and variance of them. If the variance is smal 293 and variance of them. If the variance is small (smaller than 400 square
296 milliseconds) or it is small relative to the a 294 milliseconds) or it is small relative to the average (the average is greater
297 that 6 times the standard deviation), the aver 295 that 6 times the standard deviation), the average is regarded as the "typical
298 interval" value. Otherwise, the longest of th 296 interval" value. Otherwise, the longest of the saved observed idle duration
299 values is discarded and the computation is rep 297 values is discarded and the computation is repeated for the remaining ones.
300 Again, if the variance of them is small (in th 298 Again, if the variance of them is small (in the above sense), the average is
301 taken as the "typical interval" value and so o 299 taken as the "typical interval" value and so on, until either the "typical
302 interval" is determined or too many data point 300 interval" is determined or too many data points are disregarded, in which case
303 the "typical interval" is assumed to equal "in 301 the "typical interval" is assumed to equal "infinity" (the maximum unsigned
304 integer value). The "typical interval" comput 302 integer value). The "typical interval" computed this way is compared with the
305 sleep length multiplied by the correction fact 303 sleep length multiplied by the correction factor and the minimum of the two is
306 taken as the predicted idle duration. 304 taken as the predicted idle duration.
307 305
308 Then, the governor computes an extra latency l 306 Then, the governor computes an extra latency limit to help "interactive"
309 workloads. It uses the observation that if th 307 workloads. It uses the observation that if the exit latency of the selected
310 idle state is comparable with the predicted id 308 idle state is comparable with the predicted idle duration, the total time spent
311 in that state probably will be very short and 309 in that state probably will be very short and the amount of energy to save by
312 entering it will be relatively small, so likel 310 entering it will be relatively small, so likely it is better to avoid the
313 overhead related to entering that state and ex 311 overhead related to entering that state and exiting it. Thus selecting a
314 shallower state is likely to be a better optio 312 shallower state is likely to be a better option then. The first approximation
315 of the extra latency limit is the predicted id 313 of the extra latency limit is the predicted idle duration itself which
316 additionally is divided by a value depending o 314 additionally is divided by a value depending on the number of tasks that
317 previously ran on the given CPU and now they a 315 previously ran on the given CPU and now they are waiting for I/O operations to
318 complete. The result of that division is comp 316 complete. The result of that division is compared with the latency limit coming
319 from the power management quality of service, 317 from the power management quality of service, or `PM QoS <cpu-pm-qos_>`_,
320 framework and the minimum of the two is taken 318 framework and the minimum of the two is taken as the limit for the idle states'
321 exit latency. 319 exit latency.
322 320
323 Now, the governor is ready to walk the list of 321 Now, the governor is ready to walk the list of idle states and choose one of
324 them. For this purpose, it compares the targe 322 them. For this purpose, it compares the target residency of each state with
325 the predicted idle duration and the exit laten 323 the predicted idle duration and the exit latency of it with the computed latency
326 limit. It selects the state with the target r 324 limit. It selects the state with the target residency closest to the predicted
327 idle duration, but still below it, and exit la 325 idle duration, but still below it, and exit latency that does not exceed the
328 limit. 326 limit.
329 327
330 In the final step the governor may still need 328 In the final step the governor may still need to refine the idle state selection
331 if it has not decided to `stop the scheduler t 329 if it has not decided to `stop the scheduler tick <idle-cpus-and-tick_>`_. That
332 happens if the idle duration predicted by it i 330 happens if the idle duration predicted by it is less than the tick period and
333 the tick has not been stopped already (in a pr 331 the tick has not been stopped already (in a previous iteration of the idle
334 loop). Then, the sleep length used in the pre 332 loop). Then, the sleep length used in the previous computations may not reflect
335 the real time until the closest timer event an 333 the real time until the closest timer event and if it really is greater than
336 that time, the governor may need to select a s 334 that time, the governor may need to select a shallower state with a suitable
337 target residency. 335 target residency.
338 336
339 337
340 .. _teo-gov: 338 .. _teo-gov:
341 339
342 The Timer Events Oriented (TEO) Governor 340 The Timer Events Oriented (TEO) Governor
343 ======================================== 341 ========================================
344 342
345 The timer events oriented (TEO) governor is an 343 The timer events oriented (TEO) governor is an alternative ``CPUIdle`` governor
346 for tickless systems. It follows the same bas 344 for tickless systems. It follows the same basic strategy as the ``menu`` `one
347 <menu-gov_>`_: it always tries to find the dee 345 <menu-gov_>`_: it always tries to find the deepest idle state suitable for the
348 given conditions. However, it applies a diffe 346 given conditions. However, it applies a different approach to that problem.
349 347
350 .. kernel-doc:: drivers/cpuidle/governors/teo. !! 348 First, it does not use sleep length correction factors, but instead it attempts
351 :doc: teo-description !! 349 to correlate the observed idle duration values with the available idle states
>> 350 and use that information to pick up the idle state that is most likely to
>> 351 "match" the upcoming CPU idle interval. Second, it does not take the tasks
>> 352 that were running on the given CPU in the past and are waiting on some I/O
>> 353 operations to complete now at all (there is no guarantee that they will run on
>> 354 the same CPU when they become runnable again) and the pattern detection code in
>> 355 it avoids taking timer wakeups into account. It also only uses idle duration
>> 356 values less than the current time till the closest timer (with the scheduler
>> 357 tick excluded) for that purpose.
>> 358
>> 359 Like in the ``menu`` governor `case <menu-gov_>`_, the first step is to obtain
>> 360 the *sleep length*, which is the time until the closest timer event with the
>> 361 assumption that the scheduler tick will be stopped (that also is the upper bound
>> 362 on the time until the next CPU wakeup). That value is then used to preselect an
>> 363 idle state on the basis of three metrics maintained for each idle state provided
>> 364 by the ``CPUIdle`` driver: ``hits``, ``misses`` and ``early_hits``.
>> 365
>> 366 The ``hits`` and ``misses`` metrics measure the likelihood that a given idle
>> 367 state will "match" the observed (post-wakeup) idle duration if it "matches" the
>> 368 sleep length. They both are subject to decay (after a CPU wakeup) every time
>> 369 the target residency of the idle state corresponding to them is less than or
>> 370 equal to the sleep length and the target residency of the next idle state is
>> 371 greater than the sleep length (that is, when the idle state corresponding to
>> 372 them "matches" the sleep length). The ``hits`` metric is increased if the
>> 373 former condition is satisfied and the target residency of the given idle state
>> 374 is less than or equal to the observed idle duration and the target residency of
>> 375 the next idle state is greater than the observed idle duration at the same time
>> 376 (that is, it is increased when the given idle state "matches" both the sleep
>> 377 length and the observed idle duration). In turn, the ``misses`` metric is
>> 378 increased when the given idle state "matches" the sleep length only and the
>> 379 observed idle duration is too short for its target residency.
>> 380
>> 381 The ``early_hits`` metric measures the likelihood that a given idle state will
>> 382 "match" the observed (post-wakeup) idle duration if it does not "match" the
>> 383 sleep length. It is subject to decay on every CPU wakeup and it is increased
>> 384 when the idle state corresponding to it "matches" the observed (post-wakeup)
>> 385 idle duration and the target residency of the next idle state is less than or
>> 386 equal to the sleep length (i.e. the idle state "matching" the sleep length is
>> 387 deeper than the given one).
>> 388
>> 389 The governor walks the list of idle states provided by the ``CPUIdle`` driver
>> 390 and finds the last (deepest) one with the target residency less than or equal
>> 391 to the sleep length. Then, the ``hits`` and ``misses`` metrics of that idle
>> 392 state are compared with each other and it is preselected if the ``hits`` one is
>> 393 greater (which means that that idle state is likely to "match" the observed idle
>> 394 duration after CPU wakeup). If the ``misses`` one is greater, the governor
>> 395 preselects the shallower idle state with the maximum ``early_hits`` metric
>> 396 (or if there are multiple shallower idle states with equal ``early_hits``
>> 397 metric which also is the maximum, the shallowest of them will be preselected).
>> 398 [If there is a wakeup latency constraint coming from the `PM QoS framework
>> 399 <cpu-pm-qos_>`_ which is hit before reaching the deepest idle state with the
>> 400 target residency within the sleep length, the deepest idle state with the exit
>> 401 latency within the constraint is preselected without consulting the ``hits``,
>> 402 ``misses`` and ``early_hits`` metrics.]
>> 403
>> 404 Next, the governor takes several idle duration values observed most recently
>> 405 into consideration and if at least a half of them are greater than or equal to
>> 406 the target residency of the preselected idle state, that idle state becomes the
>> 407 final candidate to ask for. Otherwise, the average of the most recent idle
>> 408 duration values below the target residency of the preselected idle state is
>> 409 computed and the governor walks the idle states shallower than the preselected
>> 410 one and finds the deepest of them with the target residency within that average.
>> 411 That idle state is then taken as the final candidate to ask for.
>> 412
>> 413 Still, at this point the governor may need to refine the idle state selection if
>> 414 it has not decided to `stop the scheduler tick <idle-cpus-and-tick_>`_. That
>> 415 generally happens if the target residency of the idle state selected so far is
>> 416 less than the tick period and the tick has not been stopped already (in a
>> 417 previous iteration of the idle loop). Then, like in the ``menu`` governor
>> 418 `case <menu-gov_>`_, the sleep length used in the previous computations may not
>> 419 reflect the real time until the closest timer event and if it really is greater
>> 420 than that time, a shallower state with a suitable target residency may need to
>> 421 be selected.
>> 422
352 423
353 .. _idle-states-representation: 424 .. _idle-states-representation:
354 425
355 Representation of Idle States 426 Representation of Idle States
356 ============================= 427 =============================
357 428
358 For the CPU idle time management purposes all 429 For the CPU idle time management purposes all of the physical idle states
359 supported by the processor have to be represen 430 supported by the processor have to be represented as a one-dimensional array of
360 |struct cpuidle_state| objects each allowing a 431 |struct cpuidle_state| objects each allowing an individual (logical) CPU to ask
361 the processor hardware to enter an idle state 432 the processor hardware to enter an idle state of certain properties. If there
362 is a hierarchy of units in the processor, one 433 is a hierarchy of units in the processor, one |struct cpuidle_state| object can
363 cover a combination of idle states supported b 434 cover a combination of idle states supported by the units at different levels of
364 the hierarchy. In that case, the `target resi 435 the hierarchy. In that case, the `target residency and exit latency parameters
365 of it <idle-loop_>`_, must reflect the propert 436 of it <idle-loop_>`_, must reflect the properties of the idle state at the
366 deepest level (i.e. the idle state of the unit 437 deepest level (i.e. the idle state of the unit containing all of the other
367 units). 438 units).
368 439
369 For example, take a processor with two cores i 440 For example, take a processor with two cores in a larger unit referred to as
370 a "module" and suppose that asking the hardwar 441 a "module" and suppose that asking the hardware to enter a specific idle state
371 (say "X") at the "core" level by one core will 442 (say "X") at the "core" level by one core will trigger the module to try to
372 enter a specific idle state of its own (say "M 443 enter a specific idle state of its own (say "MX") if the other core is in idle
373 state "X" already. In other words, asking for 444 state "X" already. In other words, asking for idle state "X" at the "core"
374 level gives the hardware a license to go as de 445 level gives the hardware a license to go as deep as to idle state "MX" at the
375 "module" level, but there is no guarantee that 446 "module" level, but there is no guarantee that this is going to happen (the core
376 asking for idle state "X" may just end up in t 447 asking for idle state "X" may just end up in that state by itself instead).
377 Then, the target residency of the |struct cpui 448 Then, the target residency of the |struct cpuidle_state| object representing
378 idle state "X" must reflect the minimum time t 449 idle state "X" must reflect the minimum time to spend in idle state "MX" of
379 the module (including the time needed to enter 450 the module (including the time needed to enter it), because that is the minimum
380 time the CPU needs to be idle to save any ener 451 time the CPU needs to be idle to save any energy in case the hardware enters
381 that state. Analogously, the exit latency par 452 that state. Analogously, the exit latency parameter of that object must cover
382 the exit time of idle state "MX" of the module 453 the exit time of idle state "MX" of the module (and usually its entry time too),
383 because that is the maximum delay between a wa 454 because that is the maximum delay between a wakeup signal and the time the CPU
384 will start to execute the first new instructio 455 will start to execute the first new instruction (assuming that both cores in the
385 module will always be ready to execute instruc 456 module will always be ready to execute instructions as soon as the module
386 becomes operational as a whole). 457 becomes operational as a whole).
387 458
388 There are processors without direct coordinati 459 There are processors without direct coordination between different levels of the
389 hierarchy of units inside them, however. In t 460 hierarchy of units inside them, however. In those cases asking for an idle
390 state at the "core" level does not automatical 461 state at the "core" level does not automatically affect the "module" level, for
391 example, in any way and the ``CPUIdle`` driver 462 example, in any way and the ``CPUIdle`` driver is responsible for the entire
392 handling of the hierarchy. Then, the definiti 463 handling of the hierarchy. Then, the definition of the idle state objects is
393 entirely up to the driver, but still the physi 464 entirely up to the driver, but still the physical properties of the idle state
394 that the processor hardware finally goes into 465 that the processor hardware finally goes into must always follow the parameters
395 used by the governor for idle state selection 466 used by the governor for idle state selection (for instance, the actual exit
396 latency of that idle state must not exceed the 467 latency of that idle state must not exceed the exit latency parameter of the
397 idle state object selected by the governor). 468 idle state object selected by the governor).
398 469
399 In addition to the target residency and exit l 470 In addition to the target residency and exit latency idle state parameters
400 discussed above, the objects representing idle 471 discussed above, the objects representing idle states each contain a few other
401 parameters describing the idle state and a poi 472 parameters describing the idle state and a pointer to the function to run in
402 order to ask the hardware to enter that state. 473 order to ask the hardware to enter that state. Also, for each
403 |struct cpuidle_state| object, there is a corr 474 |struct cpuidle_state| object, there is a corresponding
404 :c:type:`struct cpuidle_state_usage <cpuidle_s 475 :c:type:`struct cpuidle_state_usage <cpuidle_state_usage>` one containing usage
405 statistics of the given idle state. That info 476 statistics of the given idle state. That information is exposed by the kernel
406 via ``sysfs``. 477 via ``sysfs``.
407 478
408 For each CPU in the system, there is a :file:` !! 479 For each CPU in the system, there is a :file:`/sys/devices/system/cpu<N>/cpuidle/`
409 directory in ``sysfs``, where the number ``<N> 480 directory in ``sysfs``, where the number ``<N>`` is assigned to the given
410 CPU at the initialization time. That director 481 CPU at the initialization time. That directory contains a set of subdirectories
411 called :file:`state0`, :file:`state1` and so o 482 called :file:`state0`, :file:`state1` and so on, up to the number of idle state
412 objects defined for the given CPU minus one. 483 objects defined for the given CPU minus one. Each of these directories
413 corresponds to one idle state object and the l 484 corresponds to one idle state object and the larger the number in its name, the
414 deeper the (effective) idle state represented 485 deeper the (effective) idle state represented by it. Each of them contains
415 a number of files (attributes) representing th 486 a number of files (attributes) representing the properties of the idle state
416 object corresponding to it, as follows: 487 object corresponding to it, as follows:
417 488
418 ``above`` 489 ``above``
419 Total number of times this idle state 490 Total number of times this idle state had been asked for, but the
420 observed idle duration was certainly t 491 observed idle duration was certainly too short to match its target
421 residency. 492 residency.
422 493
423 ``below`` 494 ``below``
424 Total number of times this idle state !! 495 Total number of times this idle state had been asked for, but cerainly
425 a deeper idle state would have been a 496 a deeper idle state would have been a better match for the observed idle
426 duration. 497 duration.
427 498
428 ``desc`` 499 ``desc``
429 Description of the idle state. 500 Description of the idle state.
430 501
431 ``disable`` 502 ``disable``
432 Whether or not this idle state is disa 503 Whether or not this idle state is disabled.
433 504
434 ``default_status`` <<
435 The default status of this state, "ena <<
436 <<
437 ``latency`` 505 ``latency``
438 Exit latency of the idle state in micr 506 Exit latency of the idle state in microseconds.
439 507
440 ``name`` 508 ``name``
441 Name of the idle state. 509 Name of the idle state.
442 510
443 ``power`` 511 ``power``
444 Power drawn by hardware in this idle s 512 Power drawn by hardware in this idle state in milliwatts (if specified,
445 0 otherwise). 513 0 otherwise).
446 514
447 ``residency`` 515 ``residency``
448 Target residency of the idle state in 516 Target residency of the idle state in microseconds.
449 517
450 ``time`` 518 ``time``
451 Total time spent in this idle state by 519 Total time spent in this idle state by the given CPU (as measured by the
452 kernel) in microseconds. 520 kernel) in microseconds.
453 521
454 ``usage`` 522 ``usage``
455 Total number of times the hardware has 523 Total number of times the hardware has been asked by the given CPU to
456 enter this idle state. 524 enter this idle state.
457 525
458 ``rejected`` <<
459 Total number of times a request to ent <<
460 CPU was rejected. <<
461 <<
462 The :file:`desc` and :file:`name` files both c 526 The :file:`desc` and :file:`name` files both contain strings. The difference
463 between them is that the name is expected to b 527 between them is that the name is expected to be more concise, while the
464 description may be longer and it may contain w 528 description may be longer and it may contain white space or special characters.
465 The other files listed above contain integer n 529 The other files listed above contain integer numbers.
466 530
467 The :file:`disable` attribute is the only writ 531 The :file:`disable` attribute is the only writeable one. If it contains 1, the
468 given idle state is disabled for this particul 532 given idle state is disabled for this particular CPU, which means that the
469 governor will never select it for this particu 533 governor will never select it for this particular CPU and the ``CPUIdle``
470 driver will never ask the hardware to enter it 534 driver will never ask the hardware to enter it for that CPU as a result.
471 However, disabling an idle state for one CPU d 535 However, disabling an idle state for one CPU does not prevent it from being
472 asked for by the other CPUs, so it must be dis 536 asked for by the other CPUs, so it must be disabled for all of them in order to
473 never be asked for by any of them. [Note that 537 never be asked for by any of them. [Note that, due to the way the ``ladder``
474 governor is implemented, disabling an idle sta 538 governor is implemented, disabling an idle state prevents that governor from
475 selecting any idle states deeper than the disa 539 selecting any idle states deeper than the disabled one too.]
476 540
477 If the :file:`disable` attribute contains 0, t 541 If the :file:`disable` attribute contains 0, the given idle state is enabled for
478 this particular CPU, but it still may be disab 542 this particular CPU, but it still may be disabled for some or all of the other
479 CPUs in the system at the same time. Writing 543 CPUs in the system at the same time. Writing 1 to it causes the idle state to
480 be disabled for this particular CPU and writin 544 be disabled for this particular CPU and writing 0 to it allows the governor to
481 take it into consideration for the given CPU a 545 take it into consideration for the given CPU and the driver to ask for it,
482 unless that state was disabled globally in the 546 unless that state was disabled globally in the driver (in which case it cannot
483 be used at all). 547 be used at all).
484 548
485 The :file:`power` attribute is not defined ver 549 The :file:`power` attribute is not defined very well, especially for idle state
486 objects representing combinations of idle stat 550 objects representing combinations of idle states at different levels of the
487 hierarchy of units in the processor, and it ge 551 hierarchy of units in the processor, and it generally is hard to obtain idle
488 state power numbers for complex hardware, so : 552 state power numbers for complex hardware, so :file:`power` often contains 0 (not
489 available) and if it contains a nonzero number 553 available) and if it contains a nonzero number, that number may not be very
490 accurate and it should not be relied on for an 554 accurate and it should not be relied on for anything meaningful.
491 555
492 The number in the :file:`time` file generally 556 The number in the :file:`time` file generally may be greater than the total time
493 really spent by the given CPU in the given idl 557 really spent by the given CPU in the given idle state, because it is measured by
494 the kernel and it may not cover the cases in w 558 the kernel and it may not cover the cases in which the hardware refused to enter
495 this idle state and entered a shallower one in 559 this idle state and entered a shallower one instead of it (or even it did not
496 enter any idle state at all). The kernel can 560 enter any idle state at all). The kernel can only measure the time span between
497 asking the hardware to enter an idle state and 561 asking the hardware to enter an idle state and the subsequent wakeup of the CPU
498 and it cannot say what really happened in the 562 and it cannot say what really happened in the meantime at the hardware level.
499 Moreover, if the idle state object in question 563 Moreover, if the idle state object in question represents a combination of idle
500 states at different levels of the hierarchy of 564 states at different levels of the hierarchy of units in the processor,
501 the kernel can never say how deep the hardware 565 the kernel can never say how deep the hardware went down the hierarchy in any
502 particular case. For these reasons, the only 566 particular case. For these reasons, the only reliable way to find out how
503 much time has been spent by the hardware in di 567 much time has been spent by the hardware in different idle states supported by
504 it is to use idle state residency counters in 568 it is to use idle state residency counters in the hardware, if available.
505 569
506 Generally, an interrupt received when trying t <<
507 idle state entry request to be rejected, in wh <<
508 may return an error code to indicate that this <<
509 and :file:`rejected` files report the number o <<
510 was entered successfully or rejected, respecti <<
511 570
512 .. _cpu-pm-qos: 571 .. _cpu-pm-qos:
513 572
514 Power Management Quality of Service for CPUs 573 Power Management Quality of Service for CPUs
515 ============================================ 574 ============================================
516 575
517 The power management quality of service (PM Qo 576 The power management quality of service (PM QoS) framework in the Linux kernel
518 allows kernel code and user space processes to 577 allows kernel code and user space processes to set constraints on various
519 energy-efficiency features of the kernel to pr 578 energy-efficiency features of the kernel to prevent performance from dropping
520 below a required level. !! 579 below a required level. The PM QoS constraints can be set globally, in
>> 580 predefined categories referred to as PM QoS classes, or against individual
>> 581 devices.
521 582
522 CPU idle time management can be affected by PM 583 CPU idle time management can be affected by PM QoS in two ways, through the
523 global CPU latency limit and through the resum !! 584 global constraint in the ``PM_QOS_CPU_DMA_LATENCY`` class and through the
524 individual CPUs. Kernel code (e.g. device dri !! 585 resume latency constraints for individual CPUs. Kernel code (e.g. device
525 the help of special internal interfaces provid !! 586 drivers) can set both of them with the help of special internal interfaces
526 space can modify the former by opening the :fi !! 587 provided by the PM QoS framework. User space can modify the former by opening
527 device file under :file:`/dev/` and writing a !! 588 the :file:`cpu_dma_latency` special device file under :file:`/dev/` and writing
528 signed 32-bit integer) to it. In turn, the re !! 589 a binary value (interpreted as a signed 32-bit integer) to it. In turn, the
529 can be modified from user space by writing a s !! 590 resume latency constraint for a CPU can be modified by user space by writing a
530 32-bit integer) to the :file:`power/pm_qos_res !! 591 string (representing a signed 32-bit integer) to the
>> 592 :file:`power/pm_qos_resume_latency_us` file under
531 :file:`/sys/devices/system/cpu/cpu<N>/` in ``s 593 :file:`/sys/devices/system/cpu/cpu<N>/` in ``sysfs``, where the CPU number
532 ``<N>`` is allocated at the system initializat 594 ``<N>`` is allocated at the system initialization time. Negative values
533 will be rejected in both cases and, also in bo 595 will be rejected in both cases and, also in both cases, the written integer
534 number will be interpreted as a requested PM Q 596 number will be interpreted as a requested PM QoS constraint in microseconds.
535 597
536 The requested value is not automatically appli 598 The requested value is not automatically applied as a new constraint, however,
537 as it may be less restrictive (greater in this 599 as it may be less restrictive (greater in this particular case) than another
538 constraint previously requested by someone els 600 constraint previously requested by someone else. For this reason, the PM QoS
539 framework maintains a list of requests that ha !! 601 framework maintains a list of requests that have been made so far in each
540 global CPU latency limit and for each individu !! 602 global class and for each device, aggregates them and applies the effective
541 applies the effective (minimum in this particu !! 603 (minimum in this particular case) value as the new constraint.
542 constraint. <<
543 604
544 In fact, opening the :file:`cpu_dma_latency` s 605 In fact, opening the :file:`cpu_dma_latency` special device file causes a new
545 PM QoS request to be created and added to a gl !! 606 PM QoS request to be created and added to the priority list of requests in the
546 limit requests and the file descriptor coming !! 607 ``PM_QOS_CPU_DMA_LATENCY`` class and the file descriptor coming from the
547 represents that request. If that file descrip !! 608 "open" operation represents that request. If that file descriptor is then
548 number written to it will be associated with t !! 609 used for writing, the number written to it will be associated with the PM QoS
549 it as a new requested limit value. Next, the !! 610 request represented by it as a new requested constraint value. Next, the
550 used to determine the new effective value of t !! 611 priority list mechanism will be used to determine the new effective value of
551 that effective value will be set as a new CPU !! 612 the entire list of requests and that effective value will be set as a new
552 new limit value will only change the real limi !! 613 constraint. Thus setting a new requested constraint value will only change the
553 affected by it, which is the case if it is the !! 614 real constraint if the effective "list" value is affected by it. In particular,
554 in the list. !! 615 for the ``PM_QOS_CPU_DMA_LATENCY`` class it only affects the real constraint if
555 !! 616 it is the minimum of the requested constraints in the list. The process holding
556 The process holding a file descriptor obtained !! 617 a file descriptor obtained by opening the :file:`cpu_dma_latency` special device
557 :file:`cpu_dma_latency` special device file co !! 618 file controls the PM QoS request associated with that file descriptor, but it
558 associated with that file descriptor, but it c !! 619 controls this particular PM QoS request only.
559 request only. <<
560 620
561 Closing the :file:`cpu_dma_latency` special de 621 Closing the :file:`cpu_dma_latency` special device file or, more precisely, the
562 file descriptor obtained while opening it, cau 622 file descriptor obtained while opening it, causes the PM QoS request associated
563 with that file descriptor to be removed from t !! 623 with that file descriptor to be removed from the ``PM_QOS_CPU_DMA_LATENCY``
564 latency limit requests and destroyed. If that !! 624 class priority list and destroyed. If that happens, the priority list mechanism
565 mechanism will be used again, to determine the !! 625 will be used, again, to determine the new effective value for the whole list
566 list and that value will become the new limit. !! 626 and that value will become the new real constraint.
567 627
568 In turn, for each CPU there is one resume late !! 628 In turn, for each CPU there is only one resume latency PM QoS request
569 the :file:`power/pm_qos_resume_latency_us` fil !! 629 associated with the :file:`power/pm_qos_resume_latency_us` file under
570 :file:`/sys/devices/system/cpu/cpu<N>/` in ``s 630 :file:`/sys/devices/system/cpu/cpu<N>/` in ``sysfs`` and writing to it causes
571 this single PM QoS request to be updated regar 631 this single PM QoS request to be updated regardless of which user space
572 process does that. In other words, this PM Qo 632 process does that. In other words, this PM QoS request is shared by the entire
573 user space, so access to the file associated w 633 user space, so access to the file associated with it needs to be arbitrated
574 to avoid confusion. [Arguably, the only legit 634 to avoid confusion. [Arguably, the only legitimate use of this mechanism in
575 practice is to pin a process to the CPU in que 635 practice is to pin a process to the CPU in question and let it use the
576 ``sysfs`` interface to control the resume late !! 636 ``sysfs`` interface to control the resume latency constraint for it.] It
577 still only a request, however. It is an entry !! 637 still only is a request, however. It is a member of a priority list used to
578 determine the effective value to be set as the 638 determine the effective value to be set as the resume latency constraint for the
579 CPU in question every time the list of request 639 CPU in question every time the list of requests is updated this way or another
580 (there may be other requests coming from kerne 640 (there may be other requests coming from kernel code in that list).
581 641
582 CPU idle time governors are expected to regard 642 CPU idle time governors are expected to regard the minimum of the global
583 (effective) CPU latency limit and the effectiv !! 643 effective ``PM_QOS_CPU_DMA_LATENCY`` class constraint and the effective
584 the given CPU as the upper limit for the exit !! 644 resume latency constraint for the given CPU as the upper limit for the exit
585 they are allowed to select for that CPU. They !! 645 latency of the idle states they can select for that CPU. They should never
586 states with exit latency beyond that limit. !! 646 select any idle states with exit latency beyond that limit.
587 647
588 648
589 Idle States Control Via Kernel Command Line 649 Idle States Control Via Kernel Command Line
590 =========================================== 650 ===========================================
591 651
592 In addition to the ``sysfs`` interface allowin 652 In addition to the ``sysfs`` interface allowing individual idle states to be
593 `disabled for individual CPUs <idle-states-rep 653 `disabled for individual CPUs <idle-states-representation_>`_, there are kernel
594 command line parameters affecting CPU idle tim 654 command line parameters affecting CPU idle time management.
595 655
596 The ``cpuidle.off=1`` kernel command line opti 656 The ``cpuidle.off=1`` kernel command line option can be used to disable the
597 CPU idle time management entirely. It does no 657 CPU idle time management entirely. It does not prevent the idle loop from
598 running on idle CPUs, but it prevents the CPU 658 running on idle CPUs, but it prevents the CPU idle time governors and drivers
599 from being invoked. If it is added to the ker 659 from being invoked. If it is added to the kernel command line, the idle loop
600 will ask the hardware to enter idle states on 660 will ask the hardware to enter idle states on idle CPUs via the CPU architecture
601 support code that is expected to provide a def 661 support code that is expected to provide a default mechanism for this purpose.
602 That default mechanism usually is the least co 662 That default mechanism usually is the least common denominator for all of the
603 processors implementing the architecture (i.e. 663 processors implementing the architecture (i.e. CPU instruction set) in question,
604 however, so it is rather crude and not very en 664 however, so it is rather crude and not very energy-efficient. For this reason,
605 it is not recommended for production use. 665 it is not recommended for production use.
606 666
607 The ``cpuidle.governor=`` kernel command line 667 The ``cpuidle.governor=`` kernel command line switch allows the ``CPUIdle``
608 governor to use to be specified. It has to be 668 governor to use to be specified. It has to be appended with a string matching
609 the name of an available governor (e.g. ``cpui 669 the name of an available governor (e.g. ``cpuidle.governor=menu``) and that
610 governor will be used instead of the default o 670 governor will be used instead of the default one. It is possible to force
611 the ``menu`` governor to be used on the system 671 the ``menu`` governor to be used on the systems that use the ``ladder`` governor
612 by default this way, for example. 672 by default this way, for example.
613 673
614 The other kernel command line parameters contr 674 The other kernel command line parameters controlling CPU idle time management
615 described below are only relevant for the *x86 !! 675 described below are only relevant for the *x86* architecture and some of
616 to ``intel_idle`` affect Intel processors only !! 676 them affect Intel processors only.
617 677
618 The *x86* architecture support code recognizes 678 The *x86* architecture support code recognizes three kernel command line
619 options related to CPU idle time management: ` 679 options related to CPU idle time management: ``idle=poll``, ``idle=halt``,
620 and ``idle=nomwait``. The first two of them d 680 and ``idle=nomwait``. The first two of them disable the ``acpi_idle`` and
621 ``intel_idle`` drivers altogether, which effec 681 ``intel_idle`` drivers altogether, which effectively causes the entire
622 ``CPUIdle`` subsystem to be disabled and makes 682 ``CPUIdle`` subsystem to be disabled and makes the idle loop invoke the
623 architecture support code to deal with idle CP 683 architecture support code to deal with idle CPUs. How it does that depends on
624 which of the two parameters is added to the ke 684 which of the two parameters is added to the kernel command line. In the
625 ``idle=halt`` case, the architecture support c 685 ``idle=halt`` case, the architecture support code will use the ``HLT``
626 instruction of the CPUs (which, as a rule, sus 686 instruction of the CPUs (which, as a rule, suspends the execution of the program
627 and causes the hardware to attempt to enter th 687 and causes the hardware to attempt to enter the shallowest available idle state)
628 for this purpose, and if ``idle=poll`` is used 688 for this purpose, and if ``idle=poll`` is used, idle CPUs will execute a
629 more or less "lightweight" sequence of instruc !! 689 more or less ``lightweight'' sequence of instructions in a tight loop. [Note
630 that using ``idle=poll`` is somewhat drastic i 690 that using ``idle=poll`` is somewhat drastic in many cases, as preventing idle
631 CPUs from saving almost any energy at all may 691 CPUs from saving almost any energy at all may not be the only effect of it.
632 For example, on Intel hardware it effectively 692 For example, on Intel hardware it effectively prevents CPUs from using
633 P-states (see |cpufreq|) that require any numb 693 P-states (see |cpufreq|) that require any number of CPUs in a package to be
634 idle, so it very well may hurt single-thread c 694 idle, so it very well may hurt single-thread computations performance as well as
635 energy-efficiency. Thus using it for performa 695 energy-efficiency. Thus using it for performance reasons may not be a good idea
636 at all.] 696 at all.]
637 697
638 The ``idle=nomwait`` option prevents the use o !! 698 The ``idle=nomwait`` option disables the ``intel_idle`` driver and causes
639 the CPU to enter idle states. When this option !! 699 ``acpi_idle`` to be used (as long as all of the information needed by it is
640 driver will use the ``HLT`` instruction instea !! 700 there in the system's ACPI tables), but it is not allowed to use the
641 running Intel processors, this option disables !! 701 ``MWAIT`` instruction of the CPUs to ask the hardware to enter idle states.
642 and forces the use of the ``acpi_idle`` driver <<
643 case, ``acpi_idle`` driver will function only <<
644 by it is in the system's ACPI tables. <<
645 702
646 In addition to the architecture-level kernel c 703 In addition to the architecture-level kernel command line options affecting CPU
647 idle time management, there are parameters aff 704 idle time management, there are parameters affecting individual ``CPUIdle``
648 drivers that can be passed to them via the ker 705 drivers that can be passed to them via the kernel command line. Specifically,
649 the ``intel_idle.max_cstate=<n>`` and ``proces 706 the ``intel_idle.max_cstate=<n>`` and ``processor.max_cstate=<n>`` parameters,
650 where ``<n>`` is an idle state index also used 707 where ``<n>`` is an idle state index also used in the name of the given
651 state's directory in ``sysfs`` (see 708 state's directory in ``sysfs`` (see
652 `Representation of Idle States <idle-states-re 709 `Representation of Idle States <idle-states-representation_>`_), causes the
653 ``intel_idle`` and ``acpi_idle`` drivers, resp 710 ``intel_idle`` and ``acpi_idle`` drivers, respectively, to discard all of the
654 idle states deeper than idle state ``<n>``. I 711 idle states deeper than idle state ``<n>``. In that case, they will never ask
655 for any of those idle states or expose them to 712 for any of those idle states or expose them to the governor. [The behavior of
656 the two drivers is different for ``<n>`` equal 713 the two drivers is different for ``<n>`` equal to ``0``. Adding
657 ``intel_idle.max_cstate=0`` to the kernel comm 714 ``intel_idle.max_cstate=0`` to the kernel command line disables the
658 ``intel_idle`` driver and allows ``acpi_idle`` 715 ``intel_idle`` driver and allows ``acpi_idle`` to be used, whereas
659 ``processor.max_cstate=0`` is equivalent to `` 716 ``processor.max_cstate=0`` is equivalent to ``processor.max_cstate=1``.
660 Also, the ``acpi_idle`` driver is part of the 717 Also, the ``acpi_idle`` driver is part of the ``processor`` kernel module that
661 can be loaded separately and ``max_cstate=<n>` 718 can be loaded separately and ``max_cstate=<n>`` can be passed to it as a module
662 parameter when it is loaded.] 719 parameter when it is loaded.]
Linux® is a registered trademark of Linus Torvalds in the United States and other countries.
TOMOYO® is a registered trademark of NTT DATA CORPORATION.