Linux Core Power Management User's Guide (v3.14)
Linux Core Power Management User's Guide (v4.1)
Power management is a wide reaching topic and reducing the power a system uses is handled by a number of drivers and techniques. Power Management can broadly be classified into two categories: Dynamic/Active Power management and Idle Power Management. 电源管理主要考虑两个方面
Dynamic Voltage and Frequency Scaling(MPU aka CPUFREQ)
Dynamic voltage and frequency scaling, or DVFS as it is commonly known, is the ability of a part to modify both the voltage and frequency it operates at based on need, user preference, or other factors. MPU DVFS is supported in the kernel by the cpufreq driver. All supported SoCs use the generic cpufreq-cpu0 driver. cpufreq驱动实现DVFS。
Design: OPP is a pair of voltage frequency value. When scaling from High OPP to Low OPP Frequency is reduced first and then the voltage. When scaling from a lower OPP to Higher OPP we scale the voltage first and then the frequency.
The frequency at which the MPU operates is selected by a driver called a governor. Each governor has a different strategy for selecting the most appropriate frequency. The following governors are available within the kernel:
- ondemand: This governor samples the load of the cpu and scales it up aggressively in order to provide the proper amount of processing power.
- conservative: This governor is similar to ondemand but uses a less aggressive method of increasing the the OPP of the MPU.
- performance: This governor statically sets the OPP of the MPU to the highest possible frequency.
- powersave: This governor statically sets the OPP of the MPU to the lowest possible frequency.
- userspace: This governor allows the user to set the desired OPP using any value found within scaling_available_frequencies by echoing it into scaling_setspeed.
More in depth documentation about each governor can be found in the linux kernel documentation here: https://www.kernel.org/doc/Documentation/cpu-freq/governors.txt
By default, cpufreq, the cpufreq-cpu0 driver, and all of the standard governors are enabled with the ondemand governor selected as the default governor.
All of the standard governors are built-in to the kernel, and by default the ondemand governor is selected.
To view available governors,
$ cat /sys/devices/system/cpu/cpu0/cpufreq/scaling_available_governors conservative userspace powersave ondemand performance
To view current governor,
$ cat /sys/devices/system/cpu/cpu0/cpufreq/scaling_governor ondemand
To set a governor,
$ echo userspace > /sys/devices/system/cpu/cpu0/cpufreq/scaling_governor
To view current OPP (frequency in kHz)
$ cat /sys/devices/system/cpu/cpu0/cpufreq/scaling_cur_freq 720000
To view supported OPP's (frequency in kHz),
$ cat /sys/devices/system/cpu/cpu0/cpufreq/scaling_available_frequencies 275000 500000 600000 720000
To change OPP (can be done only for userspace governor. If governors like ondemand is used, OPP change happens automatically based on the system load)
$ echo 275000 > /sys/devices/system/cpu/cpu0/cpufreq/scaling_setspeed
The OPP Modifier framework and DT entries used in the older v3.12 kernel have now been replaced by platform data defined in arch/arm/mach-omap2/oppXXXX_data.c files. These files allow defining of a different set of OPPs for each different SoC, each revision of that SoC, and also selective, automatic enabling based on what is detected to be supported by the specific SoC in use.
To implement Dynamic Frequency Scaling (DFS), the voltages in the table can be changed to the same fixed value to avoid any voltage scaling from taking place if the system has been designed to use a single voltage. 仅实现动态频率调节。
Dynamic Frequency Scaling
This feature, which can be enabled via patch to the SDK, enables scaling frequency INDEPENDENT of voltage. It is also referred to as DFS (as in DVFS without the 'V').
Certain systems are unable to scale voltage, either because they employ a fixed voltage regulator, or use the ZCE package of AM335x. Without being able to scale voltage, the power savings enabled via DVFS are lost. This is because the current version of the omap-cpufreq driver requires a valid MPU voltage regulator in order to operate. The purpose of this DFS feature is to enable additional power savings for systems with these sort of limitations.
When using the ZCE package of AM335x, the CORE and MPU voltage domains are tied together. Due to Advisory 1.0.22, you are not allowed to dynamically modify the CORE frequency/voltage because the EMIF cannot support it. However, to achieve maximum power savings, it may still be desirable to use a PMIC which supports dynamic voltage scaling, in order to use Adaptive Voltage Scaling (aka SmartReflex or AVS). This implementation of DFS does not affect the ability of AVS to optimize the voltage and save additional power.
Using the patch
The patch presented here has been developed for and tested on the SDK 05.07. It modifies the omap-cpufreq driver to operate without requiring a valid MPU voltage regulator. From a user perspective, changing frequency via cpufreq is accomplished with exactly the same commands as typical DVFS. For example, switching to 300 MHz is accomplished with the following command:
echo 300000 > /sys/devices/system/cpu/cpu0/cpufreq/scaling_setspeed
After applying the patch, the user must modify the kernel defconfig in order to enable the DFS feature. You should also configure the "Maximum supported DFS voltage" (shown below) to whatever the fixed voltage level is for your system, in microvolts. For example, use the value 1100000 to signify 1.1V. The software will use the voltage level that you specify to automatically disable any Operating Performance Points (OPPs) which have voltages above that level.
On AM335x, first apply the patch, then follow this procedure to modify the kernel configuration:
cd <kernel source directory> make ARCH=arm CROSS_COMPILE=arm-linux-gnueabihf- menuconfig <the menuconfig interface should appear> Select "System Type" Select "TI OMAP Common Features" Select "Dynamic Frequency Scaling" Configure "Maximum supported DFS voltage (in microvolts)" (default is 1100000, or 1.1V) Select "Exit" until you are prompted to save the configuration changes, and save them. Rebuild the kernel.