1.. SPDX-License-Identifier: GPL-2.0 2 3======================= 4Energy Model of devices 5======================= 6 71. Overview 8----------- 9 10The Energy Model (EM) framework serves as an interface between drivers knowing 11the power consumed by devices at various performance levels, and the kernel 12subsystems willing to use that information to make energy-aware decisions. 13 14The source of the information about the power consumed by devices can vary greatly 15from one platform to another. These power costs can be estimated using 16devicetree data in some cases. In others, the firmware will know better. 17Alternatively, userspace might be best positioned. And so on. In order to avoid 18each and every client subsystem to re-implement support for each and every 19possible source of information on its own, the EM framework intervenes as an 20abstraction layer which standardizes the format of power cost tables in the 21kernel, hence enabling to avoid redundant work. 22 23The power values might be expressed in micro-Watts or in an 'abstract scale'. 24Multiple subsystems might use the EM and it is up to the system integrator to 25check that the requirements for the power value scale types are met. An example 26can be found in the Energy-Aware Scheduler documentation 27Documentation/scheduler/sched-energy.rst. For some subsystems like thermal or 28powercap power values expressed in an 'abstract scale' might cause issues. 29These subsystems are more interested in estimation of power used in the past, 30thus the real micro-Watts might be needed. An example of these requirements can 31be found in the Intelligent Power Allocation in 32Documentation/driver-api/thermal/power_allocator.rst. 33Kernel subsystems might implement automatic detection to check whether EM 34registered devices have inconsistent scale (based on EM internal flag). 35Important thing to keep in mind is that when the power values are expressed in 36an 'abstract scale' deriving real energy in micro-Joules would not be possible. 37 38The figure below depicts an example of drivers (Arm-specific here, but the 39approach is applicable to any architecture) providing power costs to the EM 40framework, and interested clients reading the data from it:: 41 42 +---------------+ +-----------------+ +---------------+ 43 | Thermal (IPA) | | Scheduler (EAS) | | Other | 44 +---------------+ +-----------------+ +---------------+ 45 | | em_cpu_energy() | 46 | | em_cpu_get() | 47 +---------+ | +---------+ 48 | | | 49 v v v 50 +---------------------+ 51 | Energy Model | 52 | Framework | 53 +---------------------+ 54 ^ ^ ^ 55 | | | em_dev_register_perf_domain() 56 +----------+ | +---------+ 57 | | | 58 +---------------+ +---------------+ +--------------+ 59 | cpufreq-dt | | arm_scmi | | Other | 60 +---------------+ +---------------+ +--------------+ 61 ^ ^ ^ 62 | | | 63 +--------------+ +---------------+ +--------------+ 64 | Device Tree | | Firmware | | ? | 65 +--------------+ +---------------+ +--------------+ 66 67In case of CPU devices the EM framework manages power cost tables per 68'performance domain' in the system. A performance domain is a group of CPUs 69whose performance is scaled together. Performance domains generally have a 701-to-1 mapping with CPUFreq policies. All CPUs in a performance domain are 71required to have the same micro-architecture. CPUs in different performance 72domains can have different micro-architectures. 73 74 752. Core APIs 76------------ 77 782.1 Config options 79^^^^^^^^^^^^^^^^^^ 80 81CONFIG_ENERGY_MODEL must be enabled to use the EM framework. 82 83 842.2 Registration of performance domains 85^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ 86 87Registration of 'advanced' EM 88~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 89 90The 'advanced' EM gets it's name due to the fact that the driver is allowed 91to provide more precised power model. It's not limited to some implemented math 92formula in the framework (like it's in 'simple' EM case). It can better reflect 93the real power measurements performed for each performance state. Thus, this 94registration method should be preferred in case considering EM static power 95(leakage) is important. 96 97Drivers are expected to register performance domains into the EM framework by 98calling the following API:: 99 100 int em_dev_register_perf_domain(struct device *dev, unsigned int nr_states, 101 struct em_data_callback *cb, cpumask_t *cpus, bool microwatts); 102 103Drivers must provide a callback function returning <frequency, power> tuples 104for each performance state. The callback function provided by the driver is free 105to fetch data from any relevant location (DT, firmware, ...), and by any mean 106deemed necessary. Only for CPU devices, drivers must specify the CPUs of the 107performance domains using cpumask. For other devices than CPUs the last 108argument must be set to NULL. 109The last argument 'microwatts' is important to set with correct value. Kernel 110subsystems which use EM might rely on this flag to check if all EM devices use 111the same scale. If there are different scales, these subsystems might decide 112to return warning/error, stop working or panic. 113See Section 3. for an example of driver implementing this 114callback, or Section 2.4 for further documentation on this API 115 116Registration of EM using DT 117~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 118 119The EM can also be registered using OPP framework and information in DT 120"operating-points-v2". Each OPP entry in DT can be extended with a property 121"opp-microwatt" containing micro-Watts power value. This OPP DT property 122allows a platform to register EM power values which are reflecting total power 123(static + dynamic). These power values might be coming directly from 124experiments and measurements. 125 126Registration of 'artificial' EM 127~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 128 129There is an option to provide a custom callback for drivers missing detailed 130knowledge about power value for each performance state. The callback 131.get_cost() is optional and provides the 'cost' values used by the EAS. 132This is useful for platforms that only provide information on relative 133efficiency between CPU types, where one could use the information to 134create an abstract power model. But even an abstract power model can 135sometimes be hard to fit in, given the input power value size restrictions. 136The .get_cost() allows to provide the 'cost' values which reflect the 137efficiency of the CPUs. This would allow to provide EAS information which 138has different relation than what would be forced by the EM internal 139formulas calculating 'cost' values. To register an EM for such platform, the 140driver must set the flag 'microwatts' to 0, provide .get_power() callback 141and provide .get_cost() callback. The EM framework would handle such platform 142properly during registration. A flag EM_PERF_DOMAIN_ARTIFICIAL is set for such 143platform. Special care should be taken by other frameworks which are using EM 144to test and treat this flag properly. 145 146Registration of 'simple' EM 147~~~~~~~~~~~~~~~~~~~~~~~~~~~ 148 149The 'simple' EM is registered using the framework helper function 150cpufreq_register_em_with_opp(). It implements a power model which is tight to 151math formula:: 152 153 Power = C * V^2 * f 154 155The EM which is registered using this method might not reflect correctly the 156physics of a real device, e.g. when static power (leakage) is important. 157 158 1592.3 Accessing performance domains 160^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ 161 162There are two API functions which provide the access to the energy model: 163em_cpu_get() which takes CPU id as an argument and em_pd_get() with device 164pointer as an argument. It depends on the subsystem which interface it is 165going to use, but in case of CPU devices both functions return the same 166performance domain. 167 168Subsystems interested in the energy model of a CPU can retrieve it using the 169em_cpu_get() API. The energy model tables are allocated once upon creation of 170the performance domains, and kept in memory untouched. 171 172The energy consumed by a performance domain can be estimated using the 173em_cpu_energy() API. The estimation is performed assuming that the schedutil 174CPUfreq governor is in use in case of CPU device. Currently this calculation is 175not provided for other type of devices. 176 177More details about the above APIs can be found in ``<linux/energy_model.h>`` 178or in Section 2.4 179 180 1812.4 Description details of this API 182^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ 183.. kernel-doc:: include/linux/energy_model.h 184 :internal: 185 186.. kernel-doc:: kernel/power/energy_model.c 187 :export: 188 189 1903. Example driver 191----------------- 192 193The CPUFreq framework supports dedicated callback for registering 194the EM for a given CPU(s) 'policy' object: cpufreq_driver::register_em(). 195That callback has to be implemented properly for a given driver, 196because the framework would call it at the right time during setup. 197This section provides a simple example of a CPUFreq driver registering a 198performance domain in the Energy Model framework using the (fake) 'foo' 199protocol. The driver implements an est_power() function to be provided to the 200EM framework:: 201 202 -> drivers/cpufreq/foo_cpufreq.c 203 204 01 static int est_power(struct device *dev, unsigned long *mW, 205 02 unsigned long *KHz) 206 03 { 207 04 long freq, power; 208 05 209 06 /* Use the 'foo' protocol to ceil the frequency */ 210 07 freq = foo_get_freq_ceil(dev, *KHz); 211 08 if (freq < 0); 212 09 return freq; 213 10 214 11 /* Estimate the power cost for the dev at the relevant freq. */ 215 12 power = foo_estimate_power(dev, freq); 216 13 if (power < 0); 217 14 return power; 218 15 219 16 /* Return the values to the EM framework */ 220 17 *mW = power; 221 18 *KHz = freq; 222 19 223 20 return 0; 224 21 } 225 22 226 23 static void foo_cpufreq_register_em(struct cpufreq_policy *policy) 227 24 { 228 25 struct em_data_callback em_cb = EM_DATA_CB(est_power); 229 26 struct device *cpu_dev; 230 27 int nr_opp; 231 28 232 29 cpu_dev = get_cpu_device(cpumask_first(policy->cpus)); 233 30 234 31 /* Find the number of OPPs for this policy */ 235 32 nr_opp = foo_get_nr_opp(policy); 236 33 237 34 /* And register the new performance domain */ 238 35 em_dev_register_perf_domain(cpu_dev, nr_opp, &em_cb, policy->cpus, 239 36 true); 240 37 } 241 38 242 39 static struct cpufreq_driver foo_cpufreq_driver = { 243 40 .register_em = foo_cpufreq_register_em, 244 41 }; 245