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 milli-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 milli-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 milli-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 87Drivers are expected to register performance domains into the EM framework by 88calling the following API:: 89 90 int em_dev_register_perf_domain(struct device *dev, unsigned int nr_states, 91 struct em_data_callback *cb, cpumask_t *cpus, bool milliwatts); 92 93Drivers must provide a callback function returning <frequency, power> tuples 94for each performance state. The callback function provided by the driver is free 95to fetch data from any relevant location (DT, firmware, ...), and by any mean 96deemed necessary. Only for CPU devices, drivers must specify the CPUs of the 97performance domains using cpumask. For other devices than CPUs the last 98argument must be set to NULL. 99The last argument 'milliwatts' is important to set with correct value. Kernel 100subsystems which use EM might rely on this flag to check if all EM devices use 101the same scale. If there are different scales, these subsystems might decide 102to: return warning/error, stop working or panic. 103See Section 3. for an example of driver implementing this 104callback, and kernel/power/energy_model.c for further documentation on this 105API. 106 107 1082.3 Accessing performance domains 109^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ 110 111There are two API functions which provide the access to the energy model: 112em_cpu_get() which takes CPU id as an argument and em_pd_get() with device 113pointer as an argument. It depends on the subsystem which interface it is 114going to use, but in case of CPU devices both functions return the same 115performance domain. 116 117Subsystems interested in the energy model of a CPU can retrieve it using the 118em_cpu_get() API. The energy model tables are allocated once upon creation of 119the performance domains, and kept in memory untouched. 120 121The energy consumed by a performance domain can be estimated using the 122em_cpu_energy() API. The estimation is performed assuming that the schedutil 123CPUfreq governor is in use in case of CPU device. Currently this calculation is 124not provided for other type of devices. 125 126More details about the above APIs can be found in include/linux/energy_model.h. 127 128 1293. Example driver 130----------------- 131 132This section provides a simple example of a CPUFreq driver registering a 133performance domain in the Energy Model framework using the (fake) 'foo' 134protocol. The driver implements an est_power() function to be provided to the 135EM framework:: 136 137 -> drivers/cpufreq/foo_cpufreq.c 138 139 01 static int est_power(unsigned long *mW, unsigned long *KHz, 140 02 struct device *dev) 141 03 { 142 04 long freq, power; 143 05 144 06 /* Use the 'foo' protocol to ceil the frequency */ 145 07 freq = foo_get_freq_ceil(dev, *KHz); 146 08 if (freq < 0); 147 09 return freq; 148 10 149 11 /* Estimate the power cost for the dev at the relevant freq. */ 150 12 power = foo_estimate_power(dev, freq); 151 13 if (power < 0); 152 14 return power; 153 15 154 16 /* Return the values to the EM framework */ 155 17 *mW = power; 156 18 *KHz = freq; 157 19 158 20 return 0; 159 21 } 160 22 161 23 static int foo_cpufreq_init(struct cpufreq_policy *policy) 162 24 { 163 25 struct em_data_callback em_cb = EM_DATA_CB(est_power); 164 26 struct device *cpu_dev; 165 27 int nr_opp, ret; 166 28 167 29 cpu_dev = get_cpu_device(cpumask_first(policy->cpus)); 168 30 169 31 /* Do the actual CPUFreq init work ... */ 170 32 ret = do_foo_cpufreq_init(policy); 171 33 if (ret) 172 34 return ret; 173 35 174 36 /* Find the number of OPPs for this policy */ 175 37 nr_opp = foo_get_nr_opp(policy); 176 38 177 39 /* And register the new performance domain */ 178 40 em_dev_register_perf_domain(cpu_dev, nr_opp, &em_cb, policy->cpus, 179 41 true); 180 42 181 43 return 0; 182 44 } 183