1 ===============================================
2 ``intel_pstate`` CPU Performance Scaling Driver
3 ===============================================
7 Copyright (c) 2017 Intel Corp., Rafael J. Wysocki <rafael.j.wysocki@intel.com>
13 ``intel_pstate`` is a part of the
14 :doc:`CPU performance scaling subsystem <cpufreq>` in the Linux kernel
15 (``CPUFreq``). It is a scaling driver for the Sandy Bridge and later
16 generations of Intel processors. Note, however, that some of those processors
17 may not be supported. [To understand ``intel_pstate`` it is necessary to know
18 how ``CPUFreq`` works in general, so this is the time to read :doc:`cpufreq` if
19 you have not done that yet.]
21 For the processors supported by ``intel_pstate``, the P-state concept is broader
22 than just an operating frequency or an operating performance point (see the
23 `LinuxCon Europe 2015 presentation by Kristen Accardi <LCEU2015_>`_ for more
24 information about that). For this reason, the representation of P-states used
25 by ``intel_pstate`` internally follows the hardware specification (for details
26 refer to `Intel® 64 and IA-32 Architectures Software Developer’s Manual
27 Volume 3: System Programming Guide <SDM_>`_). However, the ``CPUFreq`` core
28 uses frequencies for identifying operating performance points of CPUs and
29 frequencies are involved in the user space interface exposed by it, so
30 ``intel_pstate`` maps its internal representation of P-states to frequencies too
31 (fortunately, that mapping is unambiguous). At the same time, it would not be
32 practical for ``intel_pstate`` to supply the ``CPUFreq`` core with a table of
33 available frequencies due to the possible size of it, so the driver does not do
34 that. Some functionality of the core is limited by that.
36 Since the hardware P-state selection interface used by ``intel_pstate`` is
37 available at the logical CPU level, the driver always works with individual
38 CPUs. Consequently, if ``intel_pstate`` is in use, every ``CPUFreq`` policy
39 object corresponds to one logical CPU and ``CPUFreq`` policies are effectively
40 equivalent to CPUs. In particular, this means that they become "inactive" every
41 time the corresponding CPU is taken offline and need to be re-initialized when
44 ``intel_pstate`` is not modular, so it cannot be unloaded, which means that the
45 only way to pass early-configuration-time parameters to it is via the kernel
46 command line. However, its configuration can be adjusted via ``sysfs`` to a
47 great extent. In some configurations it even is possible to unregister it via
48 ``sysfs`` which allows another ``CPUFreq`` scaling driver to be loaded and
49 registered (see `below <status_attr_>`_).
55 ``intel_pstate`` can operate in three different modes: in the active mode with
56 or without hardware-managed P-states support and in the passive mode. Which of
57 them will be in effect depends on what kernel command line options are used and
58 on the capabilities of the processor.
63 This is the default operation mode of ``intel_pstate``. If it works in this
64 mode, the ``scaling_driver`` policy attribute in ``sysfs`` for all ``CPUFreq``
65 policies contains the string "intel_pstate".
67 In this mode the driver bypasses the scaling governors layer of ``CPUFreq`` and
68 provides its own scaling algorithms for P-state selection. Those algorithms
69 can be applied to ``CPUFreq`` policies in the same way as generic scaling
70 governors (that is, through the ``scaling_governor`` policy attribute in
71 ``sysfs``). [Note that different P-state selection algorithms may be chosen for
72 different policies, but that is not recommended.]
74 They are not generic scaling governors, but their names are the same as the
75 names of some of those governors. Moreover, confusingly enough, they generally
76 do not work in the same way as the generic governors they share the names with.
77 For example, the ``powersave`` P-state selection algorithm provided by
78 ``intel_pstate`` is not a counterpart of the generic ``powersave`` governor
79 (roughly, it corresponds to the ``schedutil`` and ``ondemand`` governors).
81 There are two P-state selection algorithms provided by ``intel_pstate`` in the
82 active mode: ``powersave`` and ``performance``. The way they both operate
83 depends on whether or not the hardware-managed P-states (HWP) feature has been
84 enabled in the processor and possibly on the processor model.
86 Which of the P-state selection algorithms is used by default depends on the
87 :c:macro:`CONFIG_CPU_FREQ_DEFAULT_GOV_PERFORMANCE` kernel configuration option.
88 Namely, if that option is set, the ``performance`` algorithm will be used by
89 default, and the other one will be used by default if it is not set.
94 If the processor supports the HWP feature, it will be enabled during the
95 processor initialization and cannot be disabled after that. It is possible
96 to avoid enabling it by passing the ``intel_pstate=no_hwp`` argument to the
97 kernel in the command line.
99 If the HWP feature has been enabled, ``intel_pstate`` relies on the processor to
100 select P-states by itself, but still it can give hints to the processor's
101 internal P-state selection logic. What those hints are depends on which P-state
102 selection algorithm has been applied to the given policy (or to the CPU it
105 Even though the P-state selection is carried out by the processor automatically,
106 ``intel_pstate`` registers utilization update callbacks with the CPU scheduler
107 in this mode. However, they are not used for running a P-state selection
108 algorithm, but for periodic updates of the current CPU frequency information to
109 be made available from the ``scaling_cur_freq`` policy attribute in ``sysfs``.
111 HWP + ``performance``
112 .....................
114 In this configuration ``intel_pstate`` will write 0 to the processor's
115 Energy-Performance Preference (EPP) knob (if supported) or its
116 Energy-Performance Bias (EPB) knob (otherwise), which means that the processor's
117 internal P-state selection logic is expected to focus entirely on performance.
119 This will override the EPP/EPB setting coming from the ``sysfs`` interface
120 (see `Energy vs Performance Hints`_ below).
122 Also, in this configuration the range of P-states available to the processor's
123 internal P-state selection logic is always restricted to the upper boundary
124 (that is, the maximum P-state that the driver is allowed to use).
129 In this configuration ``intel_pstate`` will set the processor's
130 Energy-Performance Preference (EPP) knob (if supported) or its
131 Energy-Performance Bias (EPB) knob (otherwise) to whatever value it was
132 previously set to via ``sysfs`` (or whatever default value it was
133 set to by the platform firmware). This usually causes the processor's
134 internal P-state selection logic to be less performance-focused.
136 Active Mode Without HWP
137 ~~~~~~~~~~~~~~~~~~~~~~~
139 This is the default operation mode for processors that do not support the HWP
140 feature. It also is used by default with the ``intel_pstate=no_hwp`` argument
141 in the kernel command line. However, in this mode ``intel_pstate`` may refuse
142 to work with the given processor if it does not recognize it. [Note that
143 ``intel_pstate`` will never refuse to work with any processor with the HWP
146 In this mode ``intel_pstate`` registers utilization update callbacks with the
147 CPU scheduler in order to run a P-state selection algorithm, either
148 ``powersave`` or ``performance``, depending on the ``scaling_cur_freq`` policy
149 setting in ``sysfs``. The current CPU frequency information to be made
150 available from the ``scaling_cur_freq`` policy attribute in ``sysfs`` is
151 periodically updated by those utilization update callbacks too.
156 Without HWP, this P-state selection algorithm is always the same regardless of
157 the processor model and platform configuration.
159 It selects the maximum P-state it is allowed to use, subject to limits set via
160 ``sysfs``, every time the driver configuration for the given CPU is updated
161 (e.g. via ``sysfs``).
163 This is the default P-state selection algorithm if the
164 :c:macro:`CONFIG_CPU_FREQ_DEFAULT_GOV_PERFORMANCE` kernel configuration option
170 Without HWP, this P-state selection algorithm generally depends on the
171 processor model and/or the system profile setting in the ACPI tables and there
172 are two variants of it.
174 One of them is used with processors from the Atom line and (regardless of the
175 processor model) on platforms with the system profile in the ACPI tables set to
176 "mobile" (laptops mostly), "tablet", "appliance PC", "desktop", or
177 "workstation". It is also used with processors supporting the HWP feature if
178 that feature has not been enabled (that is, with the ``intel_pstate=no_hwp``
179 argument in the kernel command line). It is similar to the algorithm
180 implemented by the generic ``schedutil`` scaling governor except that the
181 utilization metric used by it is based on numbers coming from feedback
182 registers of the CPU. It generally selects P-states proportional to the
183 current CPU utilization, so it is referred to as the "proportional" algorithm.
185 The second variant of the ``powersave`` P-state selection algorithm, used in all
186 of the other cases (generally, on processors from the Core line, so it is
187 referred to as the "Core" algorithm), is based on the values read from the APERF
188 and MPERF feedback registers and the previously requested target P-state.
189 It does not really take CPU utilization into account explicitly, but as a rule
190 it causes the CPU P-state to ramp up very quickly in response to increased
191 utilization which is generally desirable in server environments.
193 Regardless of the variant, this algorithm is run by the driver's utilization
194 update callback for the given CPU when it is invoked by the CPU scheduler, but
195 not more often than every 10 ms (that can be tweaked via ``debugfs`` in `this
196 particular case <Tuning Interface in debugfs_>`_). Like in the ``performance``
197 case, the hardware configuration is not touched if the new P-state turns out to
198 be the same as the current one.
200 This is the default P-state selection algorithm if the
201 :c:macro:`CONFIG_CPU_FREQ_DEFAULT_GOV_PERFORMANCE` kernel configuration option
207 This mode is used if the ``intel_pstate=passive`` argument is passed to the
208 kernel in the command line (it implies the ``intel_pstate=no_hwp`` setting too).
209 Like in the active mode without HWP support, in this mode ``intel_pstate`` may
210 refuse to work with the given processor if it does not recognize it.
212 If the driver works in this mode, the ``scaling_driver`` policy attribute in
213 ``sysfs`` for all ``CPUFreq`` policies contains the string "intel_cpufreq".
214 Then, the driver behaves like a regular ``CPUFreq`` scaling driver. That is,
215 it is invoked by generic scaling governors when necessary to talk to the
216 hardware in order to change the P-state of a CPU (in particular, the
217 ``schedutil`` governor can invoke it directly from scheduler context).
219 While in this mode, ``intel_pstate`` can be used with all of the (generic)
220 scaling governors listed by the ``scaling_available_governors`` policy attribute
221 in ``sysfs`` (and the P-state selection algorithms described above are not
222 used). Then, it is responsible for the configuration of policy objects
223 corresponding to CPUs and provides the ``CPUFreq`` core (and the scaling
224 governors attached to the policy objects) with accurate information on the
225 maximum and minimum operating frequencies supported by the hardware (including
226 the so-called "turbo" frequency ranges). In other words, in the passive mode
227 the entire range of available P-states is exposed by ``intel_pstate`` to the
228 ``CPUFreq`` core. However, in this mode the driver does not register
229 utilization update callbacks with the CPU scheduler and the ``scaling_cur_freq``
230 information comes from the ``CPUFreq`` core (and is the last frequency selected
231 by the current scaling governor for the given policy).
236 Turbo P-states Support
237 ======================
239 In the majority of cases, the entire range of P-states available to
240 ``intel_pstate`` can be divided into two sub-ranges that correspond to
241 different types of processor behavior, above and below a boundary that
242 will be referred to as the "turbo threshold" in what follows.
244 The P-states above the turbo threshold are referred to as "turbo P-states" and
245 the whole sub-range of P-states they belong to is referred to as the "turbo
246 range". These names are related to the Turbo Boost technology allowing a
247 multicore processor to opportunistically increase the P-state of one or more
248 cores if there is enough power to do that and if that is not going to cause the
249 thermal envelope of the processor package to be exceeded.
251 Specifically, if software sets the P-state of a CPU core within the turbo range
252 (that is, above the turbo threshold), the processor is permitted to take over
253 performance scaling control for that core and put it into turbo P-states of its
254 choice going forward. However, that permission is interpreted differently by
255 different processor generations. Namely, the Sandy Bridge generation of
256 processors will never use any P-states above the last one set by software for
257 the given core, even if it is within the turbo range, whereas all of the later
258 processor generations will take it as a license to use any P-states from the
259 turbo range, even above the one set by software. In other words, on those
260 processors setting any P-state from the turbo range will enable the processor
261 to put the given core into all turbo P-states up to and including the maximum
262 supported one as it sees fit.
264 One important property of turbo P-states is that they are not sustainable. More
265 precisely, there is no guarantee that any CPUs will be able to stay in any of
266 those states indefinitely, because the power distribution within the processor
267 package may change over time or the thermal envelope it was designed for might
268 be exceeded if a turbo P-state was used for too long.
270 In turn, the P-states below the turbo threshold generally are sustainable. In
271 fact, if one of them is set by software, the processor is not expected to change
272 it to a lower one unless in a thermal stress or a power limit violation
273 situation (a higher P-state may still be used if it is set for another CPU in
274 the same package at the same time, for example).
276 Some processors allow multiple cores to be in turbo P-states at the same time,
277 but the maximum P-state that can be set for them generally depends on the number
278 of cores running concurrently. The maximum turbo P-state that can be set for 3
279 cores at the same time usually is lower than the analogous maximum P-state for
280 2 cores, which in turn usually is lower than the maximum turbo P-state that can
281 be set for 1 core. The one-core maximum turbo P-state is thus the maximum
282 supported one overall.
284 The maximum supported turbo P-state, the turbo threshold (the maximum supported
285 non-turbo P-state) and the minimum supported P-state are specific to the
286 processor model and can be determined by reading the processor's model-specific
287 registers (MSRs). Moreover, some processors support the Configurable TDP
288 (Thermal Design Power) feature and, when that feature is enabled, the turbo
289 threshold effectively becomes a configurable value that can be set by the
292 Unlike ``_PSS`` objects in the ACPI tables, ``intel_pstate`` always exposes
293 the entire range of available P-states, including the whole turbo range, to the
294 ``CPUFreq`` core and (in the passive mode) to generic scaling governors. This
295 generally causes turbo P-states to be set more often when ``intel_pstate`` is
296 used relative to ACPI-based CPU performance scaling (see `below <acpi-cpufreq_>`_
297 for more information).
299 Moreover, since ``intel_pstate`` always knows what the real turbo threshold is
300 (even if the Configurable TDP feature is enabled in the processor), its
301 ``no_turbo`` attribute in ``sysfs`` (described `below <no_turbo_attr_>`_) should
302 work as expected in all cases (that is, if set to disable turbo P-states, it
303 always should prevent ``intel_pstate`` from using them).
309 To handle a given processor ``intel_pstate`` requires a number of different
310 pieces of information on it to be known, including:
312 * The minimum supported P-state.
314 * The maximum supported `non-turbo P-state <turbo_>`_.
316 * Whether or not turbo P-states are supported at all.
318 * The maximum supported `one-core turbo P-state <turbo_>`_ (if turbo P-states
321 * The scaling formula to translate the driver's internal representation
322 of P-states into frequencies and the other way around.
324 Generally, ways to obtain that information are specific to the processor model
325 or family. Although it often is possible to obtain all of it from the processor
326 itself (using model-specific registers), there are cases in which hardware
327 manuals need to be consulted to get to it too.
329 For this reason, there is a list of supported processors in ``intel_pstate`` and
330 the driver initialization will fail if the detected processor is not in that
331 list, unless it supports the `HWP feature <Active Mode_>`_. [The interface to
332 obtain all of the information listed above is the same for all of the processors
333 supporting the HWP feature, which is why they all are supported by
337 User Space Interface in ``sysfs``
338 =================================
343 ``intel_pstate`` exposes several global attributes (files) in ``sysfs`` to
344 control its functionality at the system level. They are located in the
345 ``/sys/devices/system/cpu/cpufreq/intel_pstate/`` directory and affect all
348 Some of them are not present if the ``intel_pstate=per_cpu_perf_limits``
349 argument is passed to the kernel in the command line.
352 Maximum P-state the driver is allowed to set in percent of the
353 maximum supported performance level (the highest supported `turbo
356 This attribute will not be exposed if the
357 ``intel_pstate=per_cpu_perf_limits`` argument is present in the kernel
361 Minimum P-state the driver is allowed to set in percent of the
362 maximum supported performance level (the highest supported `turbo
365 This attribute will not be exposed if the
366 ``intel_pstate=per_cpu_perf_limits`` argument is present in the kernel
370 Number of P-states supported by the processor (between 0 and 255
371 inclusive) including both turbo and non-turbo P-states (see
372 `Turbo P-states Support`_).
374 The value of this attribute is not affected by the ``no_turbo``
375 setting described `below <no_turbo_attr_>`_.
377 This attribute is read-only.
380 Ratio of the `turbo range <turbo_>`_ size to the size of the entire
381 range of supported P-states, in percent.
383 This attribute is read-only.
388 If set (equal to 1), the driver is not allowed to set any turbo P-states
389 (see `Turbo P-states Support`_). If unset (equalt to 0, which is the
390 default), turbo P-states can be set by the driver.
391 [Note that ``intel_pstate`` does not support the general ``boost``
392 attribute (supported by some other scaling drivers) which is replaced
395 This attrubute does not affect the maximum supported frequency value
396 supplied to the ``CPUFreq`` core and exposed via the policy interface,
397 but it affects the maximum possible value of per-policy P-state limits
398 (see `Interpretation of Policy Attributes`_ below for details).
403 Operation mode of the driver: "active", "passive" or "off".
406 The driver is functional and in the `active mode
410 The driver is functional and in the `passive mode
414 The driver is not functional (it is not registered as a scaling
415 driver with the ``CPUFreq`` core).
417 This attribute can be written to in order to change the driver's
418 operation mode or to unregister it. The string written to it must be
419 one of the possible values of it and, if successful, the write will
420 cause the driver to switch over to the operation mode represented by
421 that string - or to be unregistered in the "off" case. [Actually,
422 switching over from the active mode to the passive mode or the other
423 way around causes the driver to be unregistered and registered again
424 with a different set of callbacks, so all of its settings (the global
425 as well as the per-policy ones) are then reset to their default
426 values, possibly depending on the target operation mode.]
428 That only is supported in some configurations, though (for example, if
429 the `HWP feature is enabled in the processor <Active Mode With HWP_>`_,
430 the operation mode of the driver cannot be changed), and if it is not
431 supported in the current configuration, writes to this attribute with
432 fail with an appropriate error.
434 Interpretation of Policy Attributes
435 -----------------------------------
437 The interpretation of some ``CPUFreq`` policy attributes described in
438 :doc:`cpufreq` is special with ``intel_pstate`` as the current scaling driver
439 and it generally depends on the driver's `operation mode <Operation Modes_>`_.
441 First of all, the values of the ``cpuinfo_max_freq``, ``cpuinfo_min_freq`` and
442 ``scaling_cur_freq`` attributes are produced by applying a processor-specific
443 multiplier to the internal P-state representation used by ``intel_pstate``.
444 Also, the values of the ``scaling_max_freq`` and ``scaling_min_freq``
445 attributes are capped by the frequency corresponding to the maximum P-state that
446 the driver is allowed to set.
448 If the ``no_turbo`` `global attribute <no_turbo_attr_>`_ is set, the driver is
449 not allowed to use turbo P-states, so the maximum value of ``scaling_max_freq``
450 and ``scaling_min_freq`` is limited to the maximum non-turbo P-state frequency.
451 Accordingly, setting ``no_turbo`` causes ``scaling_max_freq`` and
452 ``scaling_min_freq`` to go down to that value if they were above it before.
453 However, the old values of ``scaling_max_freq`` and ``scaling_min_freq`` will be
454 restored after unsetting ``no_turbo``, unless these attributes have been written
455 to after ``no_turbo`` was set.
457 If ``no_turbo`` is not set, the maximum possible value of ``scaling_max_freq``
458 and ``scaling_min_freq`` corresponds to the maximum supported turbo P-state,
459 which also is the value of ``cpuinfo_max_freq`` in either case.
461 Next, the following policy attributes have special meaning if
462 ``intel_pstate`` works in the `active mode <Active Mode_>`_:
464 ``scaling_available_governors``
465 List of P-state selection algorithms provided by ``intel_pstate``.
468 P-state selection algorithm provided by ``intel_pstate`` currently in
469 use with the given policy.
472 Frequency of the average P-state of the CPU represented by the given
473 policy for the time interval between the last two invocations of the
474 driver's utilization update callback by the CPU scheduler for that CPU.
476 The meaning of these attributes in the `passive mode <Passive Mode_>`_ is the
477 same as for other scaling drivers.
479 Additionally, the value of the ``scaling_driver`` attribute for ``intel_pstate``
480 depends on the operation mode of the driver. Namely, it is either
481 "intel_pstate" (in the `active mode <Active Mode_>`_) or "intel_cpufreq" (in the
482 `passive mode <Passive Mode_>`_).
484 Coordination of P-State Limits
485 ------------------------------
487 ``intel_pstate`` allows P-state limits to be set in two ways: with the help of
488 the ``max_perf_pct`` and ``min_perf_pct`` `global attributes
489 <Global Attributes_>`_ or via the ``scaling_max_freq`` and ``scaling_min_freq``
490 ``CPUFreq`` policy attributes. The coordination between those limits is based
491 on the following rules, regardless of the current operation mode of the driver:
493 1. All CPUs are affected by the global limits (that is, none of them can be
494 requested to run faster than the global maximum and none of them can be
495 requested to run slower than the global minimum).
497 2. Each individual CPU is affected by its own per-policy limits (that is, it
498 cannot be requested to run faster than its own per-policy maximum and it
499 cannot be requested to run slower than its own per-policy minimum).
501 3. The global and per-policy limits can be set independently.
503 If the `HWP feature is enabled in the processor <Active Mode With HWP_>`_, the
504 resulting effective values are written into its registers whenever the limits
505 change in order to request its internal P-state selection logic to always set
506 P-states within these limits. Otherwise, the limits are taken into account by
507 scaling governors (in the `passive mode <Passive Mode_>`_) and by the driver
508 every time before setting a new P-state for a CPU.
510 Additionally, if the ``intel_pstate=per_cpu_perf_limits`` command line argument
511 is passed to the kernel, ``max_perf_pct`` and ``min_perf_pct`` are not exposed
512 at all and the only way to set the limits is by using the policy attributes.
515 Energy vs Performance Hints
516 ---------------------------
518 If ``intel_pstate`` works in the `active mode with the HWP feature enabled
519 <Active Mode With HWP_>`_ in the processor, additional attributes are present
520 in every ``CPUFreq`` policy directory in ``sysfs``. They are intended to allow
521 user space to help ``intel_pstate`` to adjust the processor's internal P-state
522 selection logic by focusing it on performance or on energy-efficiency, or
523 somewhere between the two extremes:
525 ``energy_performance_preference``
526 Current value of the energy vs performance hint for the given policy
527 (or the CPU represented by it).
529 The hint can be changed by writing to this attribute.
531 ``energy_performance_available_preferences``
532 List of strings that can be written to the
533 ``energy_performance_preference`` attribute.
535 They represent different energy vs performance hints and should be
536 self-explanatory, except that ``default`` represents whatever hint
537 value was set by the platform firmware.
539 Strings written to the ``energy_performance_preference`` attribute are
540 internally translated to integer values written to the processor's
541 Energy-Performance Preference (EPP) knob (if supported) or its
542 Energy-Performance Bias (EPB) knob.
544 [Note that tasks may by migrated from one CPU to another by the scheduler's
545 load-balancing algorithm and if different energy vs performance hints are
546 set for those CPUs, that may lead to undesirable outcomes. To avoid such
547 issues it is better to set the same energy vs performance hint for all CPUs
548 or to pin every task potentially sensitive to them to a specific CPU.]
552 ``intel_pstate`` vs ``acpi-cpufreq``
553 ====================================
555 On the majority of systems supported by ``intel_pstate``, the ACPI tables
556 provided by the platform firmware contain ``_PSS`` objects returning information
557 that can be used for CPU performance scaling (refer to the `ACPI specification`_
558 for details on the ``_PSS`` objects and the format of the information returned
561 The information returned by the ACPI ``_PSS`` objects is used by the
562 ``acpi-cpufreq`` scaling driver. On systems supported by ``intel_pstate``
563 the ``acpi-cpufreq`` driver uses the same hardware CPU performance scaling
564 interface, but the set of P-states it can use is limited by the ``_PSS``
567 On those systems each ``_PSS`` object returns a list of P-states supported by
568 the corresponding CPU which basically is a subset of the P-states range that can
569 be used by ``intel_pstate`` on the same system, with one exception: the whole
570 `turbo range <turbo_>`_ is represented by one item in it (the topmost one). By
571 convention, the frequency returned by ``_PSS`` for that item is greater by 1 MHz
572 than the frequency of the highest non-turbo P-state listed by it, but the
573 corresponding P-state representation (following the hardware specification)
574 returned for it matches the maximum supported turbo P-state (or is the
575 special value 255 meaning essentially "go as high as you can get").
577 The list of P-states returned by ``_PSS`` is reflected by the table of
578 available frequencies supplied by ``acpi-cpufreq`` to the ``CPUFreq`` core and
579 scaling governors and the minimum and maximum supported frequencies reported by
580 it come from that list as well. In particular, given the special representation
581 of the turbo range described above, this means that the maximum supported
582 frequency reported by ``acpi-cpufreq`` is higher by 1 MHz than the frequency
583 of the highest supported non-turbo P-state listed by ``_PSS`` which, of course,
584 affects decisions made by the scaling governors, except for ``powersave`` and
587 For example, if a given governor attempts to select a frequency proportional to
588 estimated CPU load and maps the load of 100% to the maximum supported frequency
589 (possibly multiplied by a constant), then it will tend to choose P-states below
590 the turbo threshold if ``acpi-cpufreq`` is used as the scaling driver, because
591 in that case the turbo range corresponds to a small fraction of the frequency
592 band it can use (1 MHz vs 1 GHz or more). In consequence, it will only go to
593 the turbo range for the highest loads and the other loads above 50% that might
594 benefit from running at turbo frequencies will be given non-turbo P-states
597 One more issue related to that may appear on systems supporting the
598 `Configurable TDP feature <turbo_>`_ allowing the platform firmware to set the
599 turbo threshold. Namely, if that is not coordinated with the lists of P-states
600 returned by ``_PSS`` properly, there may be more than one item corresponding to
601 a turbo P-state in those lists and there may be a problem with avoiding the
602 turbo range (if desirable or necessary). Usually, to avoid using turbo
603 P-states overall, ``acpi-cpufreq`` simply avoids using the topmost state listed
604 by ``_PSS``, but that is not sufficient when there are other turbo P-states in
605 the list returned by it.
607 Apart from the above, ``acpi-cpufreq`` works like ``intel_pstate`` in the
608 `passive mode <Passive Mode_>`_, except that the number of P-states it can set
609 is limited to the ones listed by the ACPI ``_PSS`` objects.
612 Kernel Command Line Options for ``intel_pstate``
613 ================================================
615 Several kernel command line options can be used to pass early-configuration-time
616 parameters to ``intel_pstate`` in order to enforce specific behavior of it. All
617 of them have to be prepended with the ``intel_pstate=`` prefix.
620 Do not register ``intel_pstate`` as the scaling driver even if the
621 processor is supported by it.
624 Register ``intel_pstate`` in the `passive mode <Passive Mode_>`_ to
627 This option implies the ``no_hwp`` one described below.
630 Register ``intel_pstate`` as the scaling driver instead of
631 ``acpi-cpufreq`` even if the latter is preferred on the given system.
633 This may prevent some platform features (such as thermal controls and
634 power capping) that rely on the availability of ACPI P-states
635 information from functioning as expected, so it should be used with
638 This option does not work with processors that are not supported by
639 ``intel_pstate`` and on platforms where the ``pcc-cpufreq`` scaling
640 driver is used instead of ``acpi-cpufreq``.
643 Do not enable the `hardware-managed P-states (HWP) feature
644 <Active Mode With HWP_>`_ even if it is supported by the processor.
647 Register ``intel_pstate`` as the scaling driver only if the
648 `hardware-managed P-states (HWP) feature <Active Mode With HWP_>`_ is
649 supported by the processor.
652 Take ACPI ``_PPC`` performance limits into account.
654 If the preferred power management profile in the FADT (Fixed ACPI
655 Description Table) is set to "Enterprise Server" or "Performance
656 Server", the ACPI ``_PPC`` limits are taken into account by default
657 and this option has no effect.
659 ``per_cpu_perf_limits``
660 Use per-logical-CPU P-State limits (see `Coordination of P-state
661 Limits`_ for details).
664 Diagnostics and Tuning
665 ======================
670 There are two static trace events that can be used for ``intel_pstate``
671 diagnostics. One of them is the ``cpu_frequency`` trace event generally used
672 by ``CPUFreq``, and the other one is the ``pstate_sample`` trace event specific
673 to ``intel_pstate``. Both of them are triggered by ``intel_pstate`` only if
674 it works in the `active mode <Active Mode_>`_.
676 The following sequence of shell commands can be used to enable them and see
677 their output (if the kernel is generally configured to support event tracing)::
679 # cd /sys/kernel/debug/tracing/
680 # echo 1 > events/power/pstate_sample/enable
681 # echo 1 > events/power/cpu_frequency/enable
683 gnome-terminal--4510 [001] ..s. 1177.680733: pstate_sample: core_busy=107 scaled=94 from=26 to=26 mperf=1143818 aperf=1230607 tsc=29838618 freq=2474476
684 cat-5235 [002] ..s. 1177.681723: cpu_frequency: state=2900000 cpu_id=2
686 If ``intel_pstate`` works in the `passive mode <Passive Mode_>`_, the
687 ``cpu_frequency`` trace event will be triggered either by the ``schedutil``
688 scaling governor (for the policies it is attached to), or by the ``CPUFreq``
689 core (for the policies with other scaling governors).
694 The ``ftrace`` interface can be used for low-level diagnostics of
695 ``intel_pstate``. For example, to check how often the function to set a
696 P-state is called, the ``ftrace`` filter can be set to to
697 :c:func:`intel_pstate_set_pstate`::
699 # cd /sys/kernel/debug/tracing/
700 # cat available_filter_functions | grep -i pstate
701 intel_pstate_set_pstate
702 intel_pstate_cpu_init
704 # echo intel_pstate_set_pstate > set_ftrace_filter
705 # echo function > current_tracer
706 # cat trace | head -15
709 # entries-in-buffer/entries-written: 80/80 #P:4
712 # / _----=> need-resched
713 # | / _---=> hardirq/softirq
714 # || / _--=> preempt-depth
716 # TASK-PID CPU# |||| TIMESTAMP FUNCTION
718 Xorg-3129 [000] ..s. 2537.644844: intel_pstate_set_pstate <-intel_pstate_timer_func
719 gnome-terminal--4510 [002] ..s. 2537.649844: intel_pstate_set_pstate <-intel_pstate_timer_func
720 gnome-shell-3409 [001] ..s. 2537.650850: intel_pstate_set_pstate <-intel_pstate_timer_func
721 <idle>-0 [000] ..s. 2537.654843: intel_pstate_set_pstate <-intel_pstate_timer_func
723 Tuning Interface in ``debugfs``
724 -------------------------------
726 The ``powersave`` algorithm provided by ``intel_pstate`` for `the Core line of
727 processors in the active mode <powersave_>`_ is based on a `PID controller`_
728 whose parameters were chosen to address a number of different use cases at the
729 same time. However, it still is possible to fine-tune it to a specific workload
730 and the ``debugfs`` interface under ``/sys/kernel/debug/pstate_snb/`` is
731 provided for this purpose. [Note that the ``pstate_snb`` directory will be
732 present only if the specific P-state selection algorithm matching the interface
733 in it actually is in use.]
735 The following files present in that directory can be used to modify the PID
736 controller parameters at run time:
745 Note, however, that achieving desirable results this way generally requires
746 expert-level understanding of the power vs performance tradeoff, so extra care
747 is recommended when attempting to do that.
750 .. _LCEU2015: http://events.linuxfoundation.org/sites/events/files/slides/LinuxConEurope_2015.pdf
751 .. _SDM: http://www.intel.com/content/www/us/en/architecture-and-technology/64-ia-32-architectures-software-developer-system-programming-manual-325384.html
752 .. _ACPI specification: http://www.uefi.org/sites/default/files/resources/ACPI_6_1.pdf
753 .. _PID controller: https://en.wikipedia.org/wiki/PID_controller
projects / karo-tx-linux.git / blob