1 CPU frequency and voltage scaling code in the Linux(TM) kernel 2 3 4 L i n u x C P U F r e q 5 6 C P U F r e q C o r e 7 8 9 Dominik Brodowski <linux@brodo.de> 10 David Kimdon <dwhedon@debian.org> 11 Rafael J. Wysocki <rafael.j.wysocki@intel.com> 12 Viresh Kumar <viresh.kumar@linaro.org> 13 14 15 16 Clock scaling allows you to change the clock speed of the CPUs on the 17 fly. This is a nice method to save battery power, because the lower 18 the clock speed, the less power the CPU consumes. 19 20 21Contents: 22--------- 231. CPUFreq core and interfaces 242. CPUFreq notifiers 253. CPUFreq Table Generation with Operating Performance Point (OPP) 26 271. General Information 28======================= 29 30The CPUFreq core code is located in drivers/cpufreq/cpufreq.c. This 31cpufreq code offers a standardized interface for the CPUFreq 32architecture drivers (those pieces of code that do actual 33frequency transitions), as well as to "notifiers". These are device 34drivers or other part of the kernel that need to be informed of 35policy changes (ex. thermal modules like ACPI) or of all 36frequency changes (ex. timing code) or even need to force certain 37speed limits (like LCD drivers on ARM architecture). Additionally, the 38kernel "constant" loops_per_jiffy is updated on frequency changes 39here. 40 41Reference counting of the cpufreq policies is done by cpufreq_cpu_get 42and cpufreq_cpu_put, which make sure that the cpufreq driver is 43correctly registered with the core, and will not be unloaded until 44cpufreq_put_cpu is called. That also ensures that the respective cpufreq 45policy doesn't get freed while being used. 46 472. CPUFreq notifiers 48==================== 49 50CPUFreq notifiers conform to the standard kernel notifier interface. 51See linux/include/linux/notifier.h for details on notifiers. 52 53There are two different CPUFreq notifiers - policy notifiers and 54transition notifiers. 55 56 572.1 CPUFreq policy notifiers 58---------------------------- 59 60These are notified when a new policy is created or removed. 61 62The phase is specified in the second argument to the notifier. The phase is 63CPUFREQ_CREATE_POLICY when the policy is first created and it is 64CPUFREQ_REMOVE_POLICY when the policy is removed. 65 66The third argument, a void *pointer, points to a struct cpufreq_policy 67consisting of several values, including min, max (the lower and upper 68frequencies (in kHz) of the new policy). 69 70 712.2 CPUFreq transition notifiers 72-------------------------------- 73 74These are notified twice for each online CPU in the policy, when the 75CPUfreq driver switches the CPU core frequency and this change has no 76any external implications. 77 78The second argument specifies the phase - CPUFREQ_PRECHANGE or 79CPUFREQ_POSTCHANGE. 80 81The third argument is a struct cpufreq_freqs with the following 82values: 83cpu - number of the affected CPU 84old - old frequency 85new - new frequency 86flags - flags of the cpufreq driver 87 883. CPUFreq Table Generation with Operating Performance Point (OPP) 89================================================================== 90For details about OPP, see Documentation/power/opp.rst 91 92dev_pm_opp_init_cpufreq_table - 93 This function provides a ready to use conversion routine to translate 94 the OPP layer's internal information about the available frequencies 95 into a format readily providable to cpufreq. 96 97 WARNING: Do not use this function in interrupt context. 98 99 Example: 100 soc_pm_init() 101 { 102 /* Do things */ 103 r = dev_pm_opp_init_cpufreq_table(dev, &freq_table); 104 if (!r) 105 policy->freq_table = freq_table; 106 /* Do other things */ 107 } 108 109 NOTE: This function is available only if CONFIG_CPU_FREQ is enabled in 110 addition to CONFIG_PM_OPP. 111 112dev_pm_opp_free_cpufreq_table - Free up the table allocated by dev_pm_opp_init_cpufreq_table 113