Originally, it is a part of the Linux Kernel source code, so it will be treated as GPLv2 (recognition that it should be).
https://www.kernel.org/doc/html/latest/index.html
Licensing documentation
The following describes the license of the Linux kernel source code (GPLv2), how to properly mark the license of individual files in the source tree, as well as links to the full license text.
https://www.kernel.org/doc/html/latest/process/license-rules.html#kernel-licensing
https://www.kernel.org/doc/html/latest/cpu-freq/cpu-drivers.html I will read.
How to Implement a new CPUFreq Processor Driver Authors:
Dominik Brodowski [email protected] Rafael J. Wysocki [email protected] Viresh Kumar [email protected]
So, you just got a brand-new CPU / chipset with datasheets and want to add cpufreq support for this CPU / chipset? Great. Here are some hints on what is necessary:
Now, you want to get a brand new CPU / chipset and datasheet and do cpufreq support for this CPU chipset, right? All right, here are some tips you need to do that.
1.1 Initialization
First of all, in an __initcall level 7 (module_init()) or later function check whether this kernel runs on the right CPU and the right chipset. If so, register a struct cpufreq_driver with the CPUfreq core using cpufreq_register_driver()
First, check that this kernel is running on the right CPU and the right chipset with __initcall level 7 (module_init ()) and later functions. If so, use cpufreq_register_driver () to register the cpufreq_driver structure for the CPUfreq core.
What shall this struct cpufreq_driver contain?
What is included in this cpufreq_driver structure?
.name - The name of this driver. .init - A pointer to the per-policy initialization function. .verify - A pointer to a “verification” function. .setpolicy or .fast_switch or .target or .target_index - See below on the differences.
--.name: The name of this driver --.init: Each policy is a pointer to an initialization function --.verify: Pointer to "verification" function --.setpolicy, .fast_switch, .target or .taget_index: These differences will be described later.
And optionally
In addition, it includes the following as options
.flags - Hints for the cpufreq core. .driver_data - cpufreq driver specific data. .resolve_freq - Returns the most appropriate frequency for a target frequency. Doesn’t change the frequency though. .get_intermediate and target_intermediate - Used to switch to stable frequency while changing CPU frequency. .get - Returns current frequency of the CPU. .bios_limit - Returns HW/BIOS max frequency limitations for the CPU. .exit - A pointer to a per-policy cleanup function called during CPU_POST_DEAD phase of cpu hotplug process. .stop_cpu - A pointer to a per-policy stop function called during CPU_DOWN_PREPARE phase of cpu hotplug process. .suspend - A pointer to a per-policy suspend function which is called with interrupts disabled and after the governor is stopped for the policy. .resume - A pointer to a per-policy resume function which is called with interrupts disabled and before the governor is started again. .ready - A pointer to a per-policy ready function which is called after the policy is fully initialized. .attr - A pointer to a NULL-terminated list of “struct freq_attr” which allow to export values to sysfs. .boost_enabled - If set, boost frequencies are enabled. .set_boost - A pointer to a per-policy function to enable/disable boost frequencies.
--.flags --Tips for cpufreq core --.driver_data --cpufreq driver specific data --.resolve_freq --Returns the optimum frequency for the target frequency. However, the frequency does not change. --.get_intermediate and target_intermediate --Used to change the CPU frequency to a stable frequency. --.get --Returns the current frequency of the CPU. --.bios_limit --Maximum CPU frequency that is the limit in HW / BIOS --. exit --Pointer to the per-policy cleanup function called during the CPU_POST_DEAD phase in the cpu hotplug process. --.stop_cpu --A pointer to the per-policy stop function that is called during the CPU_DOWN_PREPARE phase in the hotplug process. --.suspend --A pointer to the suspend function for each policy after interrupts have been disabled and goneror has been set to stop in policy. --.resume --A pointer to the per-policy resume function that is called before interrupts are disabled and the goneror is reset to started in policy. --.ready --A pointer to the ready function for each policy, which is called after the policy has completed the initialization completely. --.attr --A pointer to the null terminated list of struct freq_attr for sysfs to print the value. --.boost_enabled --If set, boost frequency is enabled. --.set_boost --a pointer to a function for each policy when boost freqiemcu is enabled / disabled.
1.2 Per-CPU Initialization
Whenever a new CPU is registered with the device model, or after the cpufreq driver registers itself, the per-policy initialization function cpufreq_driver.init is called if no cpufreq policy existed for the CPU. Note that the .init() and .exit() routines are called only once for the policy and not for each CPU managed by the policy. It takes a struct cpufreq_policy *policy as argument. What to do now?
If a cpufreq policy does not exist on the CPU when a new CPU is registered with the device model, or after the cpufreq driver registers itself, the per-policy initialization function cpufreq_driver.init () is called. Note that the .init () and .exit () routines are called only once by the policy and not for each CPU managed by the policy. The argument takes a cpufreq_policy * policy structure. What should I do then?
If necessary, activate the CPUfreq support on your CPU.
If possible, the CPU freq that supports your CPU will be activated.
Then, the driver must fill in the following values:
The driver must then enter the following values:
policy->cpuinfo.min_freq and policy->cpuinfo.max_freq the minimum and maximum frequency (in kHz) which is supported by this CPU
policy-> cpuinfo.min_freq and policy-> cpuinfo.max_freq Minimum and maximum frequencies (in kHz) supported by this CPU
policy->cpuinfo.transition_latency the time it takes on this CPU to switch between two frequencies in nanoseconds (if appropriate, else specify CPUFREQ_ETERNAL)
Time in nanoseconds required for frequency switching with this CPU (If applicable, specify CPUFREQ_ETERNAL)
policy->cur The current operating frequency of this CPU (if appropriate)
Current operating frequency running this CPU (if applicable)
policy->min, policy->max, policy->policy and, if necessary, policy->governor must contain the “default policy” for this CPU. A few moments later, cpufreq_driver.verify and either cpufreq_driver.setpolicy or cpufreq_driver.target/target_index is called with these values.
policy->min, policy->max, policy->policy and, if necessary, policy->governor You must include the "standard policy" for this CPU. After some time, these values will be used to call cpufreq_driver.verify () and cpufreq_driver.setpolicy () or cpufreq_driver.target () / target_index ().
policy->cpus Update this with the masks of the (online + offline) CPUs that do DVFS along with this CPU (i.e. that share clock/voltage rails with it).
(Online + offline) CPU mask update when running DVFS on this CPU. (That is, it shares the clock / voltage rail with the CPU).
For setting some of these values (cpuinfo.min[max]_freq, policy->min[max]), the frequency table helpers might be helpful. See the section 2 for more information on them.
Frequency tabke can be useful for setting values for cpuinfo. (min / max) _freq and polict-> (min / max). See Section 2 for more information on them.
1.3 verify
When the user decides a new policy (consisting of “policy,governor,min,max”) shall be set, this policy must be validated so that incompatible values can be corrected. For verifying these values cpufreq_verify_within_limits(struct cpufreq_policy *policy, unsigned int min_freq, unsigned int max_freq) function might be helpful. See section 2 for details on frequency table helpers.
When a user decides that a new policy consisting of policy, governor, minimum, and maximum needs to be set, this policy should be validated so that incompatible values can be fixed. The cpufreq_verify_within_limits (struct cpufreq_policy * policy, unsigned int min_freq, unsigned int max_freq) functions can help you validate these values. See Section 2 for more information on frequency table helpers.
You need to make sure that at least one valid frequency (or operating range) is within policy->min and policy->max. If necessary, increase policy->max first, and only if this is no solution, decrease policy->min.
At least one valid frequency (or scope) must be within policy-> min to policy-> max. If necessary, increase policy-> max first, and if that doesn't work, decrease policy-> min.
1.4 target or target_index or setpolicy or fast_switch?
Most cpufreq drivers or even most cpu frequency scaling algorithms only allow the CPU frequency to be set to predefined fixed values. For these, you use the ->target(), ->target_index() or ->fast_switch() callbacks.
Most cpufreq drivers and cpu frequency scaling algorithms only allow you to set the CPU frequency to a predefined fixed value. In this case, use the following function. -> target (),-> target_index () or-> fast_switch () callbacks.
Some cpufreq capable processors switch the frequency between certain limits on their own. These shall use the ->setpolicy() callback.
cpufreq-enabled processors independently switch frequencies between specific limits. These make use of the-> setpolicy () callback.
1.5. target/target_index
The target_index call has two arguments: struct cpufreq_policy *policy, and unsigned int index (into the exposed frequency table).
The call to target_index has two arguments. struct cpufreq_policy * policy, and unsigned int index (into the exposed frequency table)
The CPUfreq driver must set the new frequency when called here. The actual frequency must be determined by freq_table[index].frequency.
When called here, the CPUfreq driver will need to set a new frequency. The actual frequency should be determined based on freq_table [index] .frequency.
It should always restore to earlier frequency (i.e. policy->restore_freq) in case of errors, even if we switched to intermediate frequency earlier.
If you get an error, you should always restore the previous frequency (policy-> restore_freq), even if you switch to the previous frequency.
The target call has three arguments: struct cpufreq_policy *policy, unsigned int target_frequency, unsigned int relation.
The target is called with three arguments. struct cpufreq_policy * policy, unsigned int target_frequency, unsigned int relation.
The CPUfreq driver must set the new frequency when called here. The actual frequency must be determined using the following rules:
When called here, the CPUfreq driver will need to set a new frequency. The actual frequency should be determined using the following rules:
- keep close to “target_freq”
--Hold “target_freq” --policy-> min <= new_freq <= policy-> max (constraints must be observed!)
Here again the frequency table helper might assist you - see section 2 for details.
Again, see section 2 for the frequency table helpers to help.
1.6. fast_switch
This function is used for frequency switching from scheduler’s context. Not all drivers are expected to implement it, as sleeping from within this callback isn’t allowed. This callback must be highly optimized to do switching as fast as possible.
This function is used to switch frequencies from the scheduler context. Sleeping from within this callback is not allowed and may not be implemented by all drivers. This callback needs to be highly optimized to make the switch as fast as possible.
This function has two arguments: struct cpufreq_policy *policy and unsigned int target_frequency.
This function has two arguments. struct cpufreq_policy * policy and unsigned int target_frequency.
1.7 setpolicy
The setpolicy call only takes a struct cpufreq_policy *policy as argument. You need to set the lower limit of the in-processor or in-chipset dynamic frequency switching to policy->min, the upper limit to policy->max, and -if supported- select a performance-oriented setting when policy->policy is CPUFREQ_POLICY_PERFORMANCE, and a powersaving-oriented setting when CPUFREQ_POLICY_POWERSAVE. Also check the reference implementation in drivers/cpufreq/longrun.c
The setpolicy call takes only the cpufreq_policy * policy structure as an argument. The lower limit of dynamic frequency switching within the process or chipset should be set to policy-> min and the upper limit should be set to policy-> max. Also, if supported, you must select the performance thinking setting when policy-> policy is set to CPUFREQ_POLICY_PERFORMANCE and powersaving-oriented is set to CPUFREQ_POLICY_POWERSAVE. Check the reference implementation in drivers / cpufreq / longrun.c.
1.8 get_intermediate and target_intermediate
Only for drivers with target_index() and CPUFREQ_ASYNC_NOTIFICATION unset.
Only if CPUFREQ_ASYNC_NOTIFICATION is not set in the driver with target_index ().
get_intermediate should return a stable intermediate frequency platform wants to switch to, and target_intermediate() should set CPU to that frequency, before jumping to the frequency corresponding to ‘index’. Core will take care of sending notifications and driver doesn’t have to handle them in target_intermediate() or target_index().
get_intermediate returns a stable intermediate frequency that the platform wants to switch to. target_intermediate () must set the CPU to that frequency before jumping to the frequency corresponding to the index. Core sends a notification. The driver does not need to handle these in target_intermediate () or target_index ().
Drivers can return ‘0’ from get_intermediate() in case they don’t wish to switch to intermediate frequency for some target frequency. In that case core will directly call ->target_index().
The driver can return 0 to get_intermediate () if it does not need to be set to an intermediate frequency at the target frequency. In this case, core returns target_index () directly.
NOTE: ->target_index() should restore to policy->restore_freq in case of failures as core would send notifications for that.
Note: In case of failure in the case where core sends a notification, target_index () needs to restore policy-> restore_freq ().
As most cpufreq processors only allow for being set to a few specific frequencies, a “frequency table” with some functions might assist in some work of the processor driver.
Most of the cpufreq processors can only set a few specific frequencies, so a "frequency table" containing the functions of the currency to go to may help some work in the processor driver.
Such a “frequency table” consists of an array of struct cpufreq_frequency_table entries, with driver specific values in “driver_data”, the corresponding frequency in “frequency” and flags set.
The "frequency table" consists of an array of cpufreq_frequency_table structure entries, with driver-specific information in "driver_data", associated frequencies in "frequency", and flags.
At the end of the table, you need to add a cpufreq_frequency_table entry with frequency set to CPUFREQ_TABLE_END.
At the end of the table, an entry containing the frequency with CPUFREQ_TABLE_END set must be added to the cpufreq frequency_table entry.
And if you want to skip one entry in the table, set the frequency to CPUFREQ_ENTRY_INVALID.
Also, for the entry you want to know in the table, set CPUFREQ_ENTRY_INVALID in frequency.
The entries don’t need to be in sorted in any particular order, but if they are cpufreq core will do DVFS a bit quickly for them as search for best match is faster.
The entries do not have to be sorted in any particular order. However, if the cpufreq core does DVFS, the best search may be a little faster.
The cpufreq table is verified automatically by the core if the policy contains a valid pointer in its policy->freq_table field.
If policy contains a valid pointer contained in policy-> freq_table field, the cpufreq table is automatically validated by core.
cpufreq_frequency_table_verify() assures that at least one valid frequency is within policy->min and policy->max, and all other criteria are met. This is helpful for the ->verify call.
cpufreq_frequency_table_verify () ensures that at least one has a valid frequency that fits within policy-> min to policy-> max and that all criteria are met. This is useful for-> veryfy call.
cpufreq_frequency_table_target() is the corresponding frequency table helper for the ->target stage.
cpufreq_frequency_table_target () is a helper that corresponds to the frequency table for-> tait stage.
Just pass the values to this function, and this function returns the of the frequency table entry which contains the frequency the CPU shall be set to.
By setting a value in this function, this function returns a frequency table entry containing the frequencies to be set.
The following macros can be used as iterators over cpufreq_frequency_table:
The following macro can be used as an iterator for cpufreq_frequency_table.
cpufreq_for_each_entry(pos, table) - iterates over all entries of frequency table.
cpufreq_for_each_valid_entry(pos, table) - iterates over all entries, excluding CPUFREQ_ENTRY_INVALID frequencies.
Use arguments “pos” - a cpufreq_frequency_table * as a loop cursor and “table” - the cpufreq_frequency_table * you want to iterate over.
pos specifies cpufreq_frequency_table as the loop cursor. Then, in "table", set the frequency_table for which you want to rotate the iteration.
For example:
struct cpufreq_frequency_table *pos, *driver_freq_table;
cpufreq_for_each_entry(pos, driver_freq_table) {
/* Do something with pos */
pos->frequency = ...
}
If you need to work with the position of pos within driver_freq_table, do not subtract the pointers, as it is quite costly. Instead, use the macros cpufreq_for_each_entry_idx() and cpufreq_for_each_valid_entry_idx().
If you need to handle the position of pos in driver_freq_table, do not subtract the pointer from a cost standpoint. Instead, use the macros cpufreq_for_each_entry_idx () and cpufreq_for_each_valid_entry_idx ().
Recommended Posts