Skip to content

Commit

Permalink
Merge tag 'pm+acpi-4.5-rc1-1' of git://git.kernel.org/pub/scm/linux/k…
Browse files Browse the repository at this point in the history 
…ernel/git/rafael/linux-pm

Pull oower management and ACPI updates from Rafael Wysocki:
 "As far as the number of commits goes, ACPICA takes the lead this time,
  followed by cpufreq and the device properties framework changes.

  The most significant new feature is the debugfs-based interface to the
  ACPICA's AML debugger added in the previous cycle and a new user space
  tool for accessing it.

  On the cpufreq front, the core is updated to handle governors more
  efficiently, particularly on systems where a single cpufreq policy
  object is shared between multiple CPUs, and there are quite a few
  changes in drivers (intel_pstate, cpufreq-dt etc).

  The device properties framework is updated to handle built-in (ie
  included in the kernel itself) device properties better, among other
  things by adding a fallback mechanism that will allow drivers to
  provide default properties to be used in case the plaform firmware
  doesn't provide the properties expected by them.

  The Operating Performance Points (OPP) framework gets new DT bindings
  and debugfs support.

  A new cpufreq driver for ST platforms is added and the ACPI driver for
  AMD SoCs will now support the APM X-Gene ACPI I2C device.

  The rest is mostly fixes and cleanups all over.

  Specifics:

   - Add a debugfs-based interface for interacting with the ACPICA's AML
     debugger introduced in the previous cycle and a new user space tool
     for that, fix some bugs related to the AML debugger and clean up
     the code in question (Lv Zheng, Dan Carpenter, Colin Ian King,
     Markus Elfring).

   - Update ACPICA to upstream revision 20151218 including a number of
     fixes and cleanups in the ACPICA core (Bob Moore, Lv Zheng, Labbe
     Corentin, Prarit Bhargava, Colin Ian King, David E Box, Rafael
     Wysocki).

     In particular, the previously added erroneous support for the _SUB
     object is dropped, the concatenate operator will support all ACPI
     objects now, the Debug Object handling is improved, the SuperName
     handling of parameters being control methods is fixed, the
     ObjectType operator handling is updated to follow ACPI 5.0A and the
     handling of CondRefOf and RefOf is updated accordingly, module-
     level code will be executed after loading each ACPI table now
     (instead of being run once after all tables containing AML have
     been loaded), the Operation Region handlers management is updated
     to fix some reported problems and a the ACPICA code in the kernel
     is more in line with the upstream now.

   - Update the ACPI backlight driver to provide information on whether
     or not it will generate key-presses for brightness change hotkeys
     and update some platform drivers (dell-wmi, thinkpad_acpi) to use
     that information to avoid sending double key-events to users pace
     for these, add new ACPI backlight quirks (Hans de Goede, Aaron Lu,
     Adrien Schildknecht).

   - Improve the ACPI handling of interrupt GPIOs (Christophe Ricard).

   - Fix the handling of the list of device IDs of device objects found
     in the ACPI namespace and add a helper for checking if there is a
     device object for a given device ID (Lukas Wunner).

   - Change the logic in the ACPI namespace scanning code to create
     struct acpi_device objects for all ACPI device objects found in the
     namespace even if _STA fails for them which helps to avoid device
     enumeration problems on Microsoft Surface 3 (Aaron Lu).

   - Add support for the APM X-Gene ACPI I2C device to the ACPI driver
     for AMD SoCs (Loc Ho).

   - Fix the long-standing issue with the DMA controller on Intel SoCs
     where ACPI tables have no power management support for the DMA
     controller itself, but it can be powered off automatically when the
     last (other) device on the SoC is powered off via ACPI and clean up
     the ACPI driver for Intel SoCs (acpi-lpss) after previous attempts
     to fix that problem (Andy Shevchenko).

   - Assorted ACPI fixes and cleanups (Andy Lutomirski, Colin Ian King,
     Javier Martinez Canillas, Ken Xue, Mathias Krause, Rafael Wysocki,
     Sinan Kaya).

   - Update the device properties framework for better handling of
     built-in properties, add support for built-in properties to the
     platform bus type, update the MFD subsystem's handling of device
     properties and add support for passing default configuration data
     as device properties to the intel-lpss MFD drivers, convert the
     designware I2C driver to use the unified device properties API and
     add a fallback mechanism for using default built-in properties if
     the platform firmware fails to provide the properties as expected
     by drivers (Andy Shevchenko, Mika Westerberg, Heikki Krogerus,
     Andrew Morton).

   - Add new Device Tree bindings to the Operating Performance Points
     (OPP) framework and update the exynos4412 DT binding accordingly,
     introduce debugfs support for the OPP framework (Viresh Kumar,
     Bartlomiej Zolnierkiewicz).

   - Migrate the mt8173 cpufreq driver to the new OPP bindings (Pi-Cheng
     Chen).

   - Update the cpufreq core to make the handling of governors more
     efficient, especially on systems where policy objects are shared
     between multiple CPUs (Viresh Kumar, Rafael Wysocki).

   - Fix cpufreq governor handling on configurations with
     CONFIG_HZ_PERIODIC set (Chen Yu).

   - Clean up the cpufreq core code related to the boost sysfs knob
     support and update the ACPI cpufreq driver accordingly (Rafael
     Wysocki).

   - Add a new cpufreq driver for ST platforms and corresponding Device
     Tree bindings (Lee Jones).

   - Update the intel_pstate driver to allow the P-state selection
     algorithm used by it to depend on the CPU ID of the processor it is
     running on, make it use a special P-state selection algorithm (with
     an IO wait time compensation tweak) on Atom CPUs based on the
     Airmont and Silvermont cores so as to reduce their energy
     consumption and improve intel_pstate documentation (Philippe
     Longepe, Srinivas Pandruvada).

   - Update the cpufreq-dt driver to support registering cooling devices
     that use the (P * V^2 * f) dynamic power draw formula where V is
     the voltage, f is the frequency and P is a constant coefficient
     provided by Device Tree and update the arm_big_little cpufreq
     driver to use that support (Punit Agrawal).

   - Assorted cpufreq driver (cpufreq-dt, qoriq, pcc-cpufreq,
     blackfin-cpufreq) updates (Andrzej Hajda, Hongtao Jia, Jacob
     Tanenbaum, Markus Elfring).

   - cpuidle core tweaks related to polling and measured_us calculation
     (Rik van Riel).

   - Removal of modularity from a few cpuidle drivers (clps711x, ux500,
     exynos) that cannot be built as modules in practice (Paul
     Gortmaker).

   - PM core update to prevent devices from being probed during system
     suspend/resume which is generally problematic and may lead to
     inconsistent behavior (Grygorii Strashko).

   - Assorted updates of the PM core and related code (Julia Lawall,
     Manuel Pégourié-Gonnard, Maruthi Bayyavarapu, Rafael Wysocki, Ulf
     Hansson).

   - PNP bus type updates (Christophe Le Roy, Heiner Kallweit).

   - PCI PM code cleanups (Jarkko Nikula, Julia Lawall).

   - cpupower tool updates (Jacob Tanenbaum, Thomas Renninger)"

* tag 'pm+acpi-4.5-rc1-1' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm: (177 commits)
  PM / clk: don't leave clocks enabled when driver not bound
  i2c: dw: Add APM X-Gene ACPI I2C device support
  ACPI / APD: Add APM X-Gene ACPI I2C device support
  ACPI / LPSS: change 'does not have' to 'has' in comment
  Revert "dmaengine: dw: platform: provide platform data for Intel"
  dmaengine: dw: return immediately from IRQ when DMA isn't in use
  dmaengine: dw: platform: power on device on shutdown
  ACPI / LPSS: override power state for LPSS DMA device
  PM / OPP: Use snprintf() instead of sprintf()
  Documentation: cpufreq: intel_pstate: enhance documentation
  ACPI, PCI, irq: remove redundant check for null string pointer
  ACPI / video: driver must be registered before checking for keypresses
  cpufreq-dt: fix handling regulator_get_voltage() result
  cpufreq: governor: Fix negative idle_time when configured with CONFIG_HZ_PERIODIC
  PM / sleep: Add support for read-only sysfs attributes
  ACPI: Fix white space in a structure definition
  ACPI / SBS: fix inconsistent indenting inside if statement
  PNP: respect PNP_DRIVER_RES_DO_NOT_CHANGE when detaching
  ACPI / PNP: constify device IDs
  ACPI / PCI: Simplify acpi_penalize_isa_irq()
  ...
  • Loading branch information
@torvalds
torvalds committed Jan 13, 2016
2 parents c17488d + a889f76 commit 6799060
Show file tree
Hide file tree
Showing 260 changed files with 7,312 additions and 3,389 deletions.
241 changes: 199 additions & 42 deletions Documentation/cpu-freq/intel-pstate.txt
View file
Original file line number Diff line number Diff line change
@@ -1,65 +1,222 @@
Intel P-state driver
Intel P-State driver
--------------------

This driver provides an interface to control the P state selection for
SandyBridge+ Intel processors. The driver can operate two different
modes based on the processor model, legacy mode and Hardware P state (HWP)
mode.

In legacy mode, the Intel P-state implements two internal governors,
performance and powersave, that differ from the general cpufreq governors of
the same name (the general cpufreq governors implement target(), whereas the
internal Intel P-state governors implement setpolicy()). The internal
performance governor sets the max_perf_pct and min_perf_pct to 100; that is,
the governor selects the highest available P state to maximize the performance
of the core. The internal powersave governor selects the appropriate P state
based on the current load on the CPU.

In HWP mode P state selection is implemented in the processor
itself. The driver provides the interfaces between the cpufreq core and
the processor to control P state selection based on user preferences
and reporting frequency to the cpufreq core. In this mode the
internal Intel P-state governor code is disabled.

In addition to the interfaces provided by the cpufreq core for
controlling frequency the driver provides sysfs files for
controlling P state selection. These files have been added to
/sys/devices/system/cpu/intel_pstate/

max_perf_pct: limits the maximum P state that will be requested by
the driver stated as a percentage of the available performance. The
available (P states) performance may be reduced by the no_turbo
This driver provides an interface to control the P-State selection for the
SandyBridge+ Intel processors.

The following document explains P-States:
http://events.linuxfoundation.org/sites/events/files/slides/LinuxConEurope_2015.pdf
As stated in the document, P-State doesn’t exactly mean a frequency. However, for
the sake of the relationship with cpufreq, P-State and frequency are used
interchangeably.

Understanding the cpufreq core governors and policies are important before
discussing more details about the Intel P-State driver. Based on what callbacks
a cpufreq driver provides to the cpufreq core, it can support two types of
drivers:
- with target_index() callback: In this mode, the drivers using cpufreq core
simply provide the minimum and maximum frequency limits and an additional
interface target_index() to set the current frequency. The cpufreq subsystem
has a number of scaling governors ("performance", "powersave", "ondemand",
etc.). Depending on which governor is in use, cpufreq core will call for
transitions to a specific frequency using target_index() callback.
- setpolicy() callback: In this mode, drivers do not provide target_index()
callback, so cpufreq core can't request a transition to a specific frequency.
The driver provides minimum and maximum frequency limits and callbacks to set a
policy. The policy in cpufreq sysfs is referred to as the "scaling governor".
The cpufreq core can request the driver to operate in any of the two policies:
"performance: and "powersave". The driver decides which frequency to use based
on the above policy selection considering minimum and maximum frequency limits.

The Intel P-State driver falls under the latter category, which implements the
setpolicy() callback. This driver decides what P-State to use based on the
requested policy from the cpufreq core. If the processor is capable of
selecting its next P-State internally, then the driver will offload this
responsibility to the processor (aka HWP: Hardware P-States). If not, the
driver implements algorithms to select the next P-State.

Since these policies are implemented in the driver, they are not same as the
cpufreq scaling governors implementation, even if they have the same name in
the cpufreq sysfs (scaling_governors). For example the "performance" policy is
similar to cpufreq’s "performance" governor, but "powersave" is completely
different than the cpufreq "powersave" governor. The strategy here is similar
to cpufreq "ondemand", where the requested P-State is related to the system load.

Sysfs Interface

In addition to the frequency-controlling interfaces provided by the cpufreq
core, the driver provides its own sysfs files to control the P-State selection.
These files have been added to /sys/devices/system/cpu/intel_pstate/.
Any changes made to these files are applicable to all CPUs (even in a
multi-package system).

max_perf_pct: Limits the maximum P-State that will be requested by
the driver. It states it as a percentage of the available performance. The
available (P-State) performance may be reduced by the no_turbo
setting described below.

min_perf_pct: limits the minimum P state that will be requested by
the driver stated as a percentage of the max (non-turbo)
min_perf_pct: Limits the minimum P-State that will be requested by
the driver. It states it as a percentage of the max (non-turbo)
performance level.

no_turbo: limits the driver to selecting P states below the turbo
no_turbo: Limits the driver to selecting P-State below the turbo
frequency range.

turbo_pct: displays the percentage of the total performance that
is supported by hardware that is in the turbo range. This number
turbo_pct: Displays the percentage of the total performance that
is supported by hardware that is in the turbo range. This number
is independent of whether turbo has been disabled or not.

num_pstates: displays the number of pstates that are supported
by hardware. This number is independent of whether turbo has
num_pstates: Displays the number of P-States that are supported
by hardware. This number is independent of whether turbo has
been disabled or not.

For example, if a system has these parameters:
Max 1 core turbo ratio: 0x21 (Max 1 core ratio is the maximum P-State)
Max non turbo ratio: 0x17
Minimum ratio : 0x08 (Here the ratio is called max efficiency ratio)

Sysfs will show :
max_perf_pct:100, which corresponds to 1 core ratio
min_perf_pct:24, max_efficiency_ratio / max 1 Core ratio
no_turbo:0, turbo is not disabled
num_pstates:26 = (max 1 Core ratio - Max Efficiency Ratio + 1)
turbo_pct:39 = (max 1 core ratio - max non turbo ratio) / num_pstates

Refer to "Intel® 64 and IA-32 Architectures Software Developer’s Manual
Volume 3: System Programming Guide" to understand ratios.

cpufreq sysfs for Intel P-State

Since this driver registers with cpufreq, cpufreq sysfs is also presented.
There are some important differences, which need to be considered.

scaling_cur_freq: This displays the real frequency which was used during
the last sample period instead of what is requested. Some other cpufreq driver,
like acpi-cpufreq, displays what is requested (Some changes are on the
way to fix this for acpi-cpufreq driver). The same is true for frequencies
displayed at /proc/cpuinfo.

scaling_governor: This displays current active policy. Since each CPU has a
cpufreq sysfs, it is possible to set a scaling governor to each CPU. But this
is not possible with Intel P-States, as there is one common policy for all
CPUs. Here, the last requested policy will be applicable to all CPUs. It is
suggested that one use the cpupower utility to change policy to all CPUs at the
same time.

scaling_setspeed: This attribute can never be used with Intel P-State.

scaling_max_freq/scaling_min_freq: This interface can be used similarly to
the max_perf_pct/min_perf_pct of Intel P-State sysfs. However since frequencies
are converted to nearest possible P-State, this is prone to rounding errors.
This method is not preferred to limit performance.

affected_cpus: Not used
related_cpus: Not used

For contemporary Intel processors, the frequency is controlled by the
processor itself and the P-states exposed to software are related to
processor itself and the P-State exposed to software is related to
performance levels. The idea that frequency can be set to a single
frequency is fiction for Intel Core processors. Even if the scaling
driver selects a single P state the actual frequency the processor
frequency is fictional for Intel Core processors. Even if the scaling
driver selects a single P-State, the actual frequency the processor
will run at is selected by the processor itself.

For legacy mode debugfs files have also been added to allow tuning of
the internal governor algorythm. These files are located at
/sys/kernel/debug/pstate_snb/ These files are NOT present in HWP mode.
Tuning Intel P-State driver

When HWP mode is not used, debugfs files have also been added to allow the
tuning of the internal governor algorithm. These files are located at
/sys/kernel/debug/pstate_snb/. The algorithm uses a PID (Proportional
Integral Derivative) controller. The PID tunable parameters are:

deadband
d_gain_pct
i_gain_pct
p_gain_pct
sample_rate_ms
setpoint

To adjust these parameters, some understanding of driver implementation is
necessary. There are some tweeks described here, but be very careful. Adjusting
them requires expert level understanding of power and performance relationship.
These limits are only useful when the "powersave" policy is active.

-To make the system more responsive to load changes, sample_rate_ms can
be adjusted (current default is 10ms).
-To make the system use higher performance, even if the load is lower, setpoint
can be adjusted to a lower number. This will also lead to faster ramp up time
to reach the maximum P-State.
If there are no derivative and integral coefficients, The next P-State will be
equal to:
current P-State - ((setpoint - current cpu load) * p_gain_pct)

For example, if the current PID parameters are (Which are defaults for the core
processors like SandyBridge):
deadband = 0
d_gain_pct = 0
i_gain_pct = 0
p_gain_pct = 20
sample_rate_ms = 10
setpoint = 97

If the current P-State = 0x08 and current load = 100, this will result in the
next P-State = 0x08 - ((97 - 100) * 0.2) = 8.6 (rounded to 9). Here the P-State
goes up by only 1. If during next sample interval the current load doesn't
change and still 100, then P-State goes up by one again. This process will
continue as long as the load is more than the setpoint until the maximum P-State
is reached.

For the same load at setpoint = 60, this will result in the next P-State
= 0x08 - ((60 - 100) * 0.2) = 16
So by changing the setpoint from 97 to 60, there is an increase of the
next P-State from 9 to 16. So this will make processor execute at higher
P-State for the same CPU load. If the load continues to be more than the
setpoint during next sample intervals, then P-State will go up again till the
maximum P-State is reached. But the ramp up time to reach the maximum P-State
will be much faster when the setpoint is 60 compared to 97.

Debugging Intel P-State driver

Event tracing
To debug P-State transition, the Linux event tracing interface can be used.
There are two specific events, which can be enabled (Provided the kernel
configs related to event tracing are enabled).

# cd /sys/kernel/debug/tracing/
# echo 1 > events/power/pstate_sample/enable
# echo 1 > events/power/cpu_frequency/enable
# cat trace
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
cat-5235 [002] ..s. 1177.681723: cpu_frequency: state=2900000 cpu_id=2


Using ftrace

If function level tracing is required, the Linux ftrace interface can be used.
For example if we want to check how often a function to set a P-State is
called, we can set ftrace filter to intel_pstate_set_pstate.

# cd /sys/kernel/debug/tracing/
# cat available_filter_functions | grep -i pstate
intel_pstate_set_pstate
intel_pstate_cpu_init
...

# echo intel_pstate_set_pstate > set_ftrace_filter
# echo function > current_tracer
# cat trace | head -15
# tracer: function
#
# entries-in-buffer/entries-written: 80/80 #P:4
#
# _-----=> irqs-off
# / _----=> need-resched
# | / _---=> hardirq/softirq
# || / _--=> preempt-depth
# ||| / delay
# TASK-PID CPU# |||| TIMESTAMP FUNCTION
# | | | |||| | |
Xorg-3129 [000] ..s. 2537.644844: intel_pstate_set_pstate <-intel_pstate_timer_func
gnome-terminal--4510 [002] ..s. 2537.649844: intel_pstate_set_pstate <-intel_pstate_timer_func
gnome-shell-3409 [001] ..s. 2537.650850: intel_pstate_set_pstate <-intel_pstate_timer_func
<idle>-0 [000] ..s. 2537.654843: intel_pstate_set_pstate <-intel_pstate_timer_func
4 changes: 2 additions & 2 deletions Documentation/cpu-freq/pcc-cpufreq.txt
View file
Original file line number Diff line number Diff line change
Expand Up @@ -159,8 +159,8 @@ to be strictly associated with a P-state.

2.2 cpuinfo_transition_latency:
-------------------------------
The cpuinfo_transition_latency field is 0. The PCC specification does
not include a field to expose this value currently.
The cpuinfo_transition_latency field is CPUFREQ_ETERNAL. The PCC specification
does not include a field to expose this value currently.

2.3 cpuinfo_cur_freq:
---------------------
Expand Down
17 changes: 17 additions & 0 deletions Documentation/devicetree/bindings/arm/cpus.txt
View file
Original file line number Diff line number Diff line change
Expand Up @@ -242,6 +242,23 @@ nodes to be present and contain the properties described below.
Definition: Specifies the syscon node controlling the cpu core
power domains.

- dynamic-power-coefficient
Usage: optional
Value type: <prop-encoded-array>
Definition: A u32 value that represents the running time dynamic
power coefficient in units of mW/MHz/uVolt^2. The
coefficient can either be calculated from power
measurements or derived by analysis.

The dynamic power consumption of the CPU is
proportional to the square of the Voltage (V) and
the clock frequency (f). The coefficient is used to
calculate the dynamic power as below -

Pdyn = dynamic-power-coefficient * V^2 * f

where voltage is in uV, frequency is in MHz.

Example 1 (dual-cluster big.LITTLE system 32-bit):

cpus {
Expand Down
91 changes: 91 additions & 0 deletions Documentation/devicetree/bindings/cpufreq/cpufreq-st.txt
View file
Original file line number Diff line number Diff line change
@@ -0,0 +1,91 @@
Binding for ST's CPUFreq driver
===============================

ST's CPUFreq driver attempts to read 'process' and 'version' attributes
from the SoC, then supplies the OPP framework with 'prop' and 'supported
hardware' information respectively. The framework is then able to read
the DT and operate in the usual way.

For more information about the expected DT format [See: ../opp/opp.txt].

Frequency Scaling only
----------------------

No vendor specific driver required for this.

Located in CPU's node:

- operating-points : [See: ../power/opp.txt]

Example [safe]
--------------

cpus {
cpu@0 {
/* kHz uV */
operating-points = <1500000 0
1200000 0
800000 0
500000 0>;
};
};

Dynamic Voltage and Frequency Scaling (DVFS)
--------------------------------------------

This requires the ST CPUFreq driver to supply 'process' and 'version' info.

Located in CPU's node:

- operating-points-v2 : [See ../power/opp.txt]

Example [unsafe]
----------------

cpus {
cpu@0 {
operating-points-v2 = <&cpu0_opp_table>;
};
};

cpu0_opp_table: opp_table {
compatible = "operating-points-v2";

/* ############################################################### */
/* # WARNING: Do not attempt to copy/replicate these nodes, # */
/* # they are only to be supplied by the bootloader !!! # */
/* ############################################################### */
opp0 {
/* Major Minor Substrate */
/* 2 all all */
opp-supported-hw = <0x00000004 0xffffffff 0xffffffff>;
opp-hz = /bits/ 64 <1500000000>;
clock-latency-ns = <10000000>;

opp-microvolt-pcode0 = <1200000>;
opp-microvolt-pcode1 = <1200000>;
opp-microvolt-pcode2 = <1200000>;
opp-microvolt-pcode3 = <1200000>;
opp-microvolt-pcode4 = <1170000>;
opp-microvolt-pcode5 = <1140000>;
opp-microvolt-pcode6 = <1100000>;
opp-microvolt-pcode7 = <1070000>;
};

opp1 {
/* Major Minor Substrate */
/* all all all */
opp-supported-hw = <0xffffffff 0xffffffff 0xffffffff>;
opp-hz = /bits/ 64 <1200000000>;
clock-latency-ns = <10000000>;

opp-microvolt-pcode0 = <1110000>;
opp-microvolt-pcode1 = <1150000>;
opp-microvolt-pcode2 = <1100000>;
opp-microvolt-pcode3 = <1080000>;
opp-microvolt-pcode4 = <1040000>;
opp-microvolt-pcode5 = <1020000>;
opp-microvolt-pcode6 = <980000>;
opp-microvolt-pcode7 = <930000>;
};
};
Loading

0 comments on commit 6799060

Please sign in to comment.