linux/drivers/thermal/devfreq_cooling.c

576 lines
15 KiB
C
Raw Normal View History

// SPDX-License-Identifier: GPL-2.0
/*
* devfreq_cooling: Thermal cooling device implementation for devices using
* devfreq
*
* Copyright (C) 2014-2015 ARM Limited
*
* TODO:
* - If OPPs are added or removed after devfreq cooling has
* registered, the devfreq cooling won't react to it.
*/
#include <linux/devfreq.h>
#include <linux/devfreq_cooling.h>
#include <linux/energy_model.h>
#include <linux/export.h>
#include <linux/slab.h>
#include <linux/pm_opp.h>
#include <linux/pm_qos.h>
#include <linux/thermal.h>
#include <linux/units.h>
#include "thermal_trace.h"
#define SCALE_ERROR_MITIGATION 100
/**
* struct devfreq_cooling_device - Devfreq cooling device
* devfreq_cooling_device registered.
* @cdev: Pointer to associated thermal cooling device.
* @cooling_ops: devfreq callbacks to thermal cooling device ops
* @devfreq: Pointer to associated devfreq device.
* @cooling_state: Current cooling state.
* @freq_table: Pointer to a table with the frequencies sorted in descending
* order. You can index the table by cooling device state
* @max_state: It is the last index, that is, one less than the number of the
* OPPs
* @power_ops: Pointer to devfreq_cooling_power, a more precised model.
* @res_util: Resource utilization scaling factor for the power.
* It is multiplied by 100 to minimize the error. It is used
* for estimation of the power budget instead of using
* 'utilization' (which is 'busy_time' / 'total_time').
* The 'res_util' range is from 100 to power * 100 for the
* corresponding 'state'.
* @capped_state: index to cooling state with in dynamic power budget
* @req_max_freq: PM QoS request for limiting the maximum frequency
* of the devfreq device.
* @em_pd: Energy Model for the associated Devfreq device
*/
struct devfreq_cooling_device {
struct thermal_cooling_device *cdev;
struct thermal_cooling_device_ops cooling_ops;
struct devfreq *devfreq;
unsigned long cooling_state;
u32 *freq_table;
size_t max_state;
struct devfreq_cooling_power *power_ops;
u32 res_util;
int capped_state;
struct dev_pm_qos_request req_max_freq;
struct em_perf_domain *em_pd;
};
static int devfreq_cooling_get_max_state(struct thermal_cooling_device *cdev,
unsigned long *state)
{
struct devfreq_cooling_device *dfc = cdev->devdata;
*state = dfc->max_state;
return 0;
}
static int devfreq_cooling_get_cur_state(struct thermal_cooling_device *cdev,
unsigned long *state)
{
struct devfreq_cooling_device *dfc = cdev->devdata;
*state = dfc->cooling_state;
return 0;
}
static int devfreq_cooling_set_cur_state(struct thermal_cooling_device *cdev,
unsigned long state)
{
struct devfreq_cooling_device *dfc = cdev->devdata;
struct devfreq *df = dfc->devfreq;
struct device *dev = df->dev.parent;
struct em_perf_state *table;
unsigned long freq;
int perf_idx;
if (state == dfc->cooling_state)
return 0;
dev_dbg(dev, "Setting cooling state %lu\n", state);
if (state > dfc->max_state)
return -EINVAL;
if (dfc->em_pd) {
perf_idx = dfc->max_state - state;
rcu_read_lock();
table = em_perf_state_from_pd(dfc->em_pd);
freq = table[perf_idx].frequency * 1000;
rcu_read_unlock();
} else {
freq = dfc->freq_table[state];
}
dev_pm_qos_update_request(&dfc->req_max_freq,
DIV_ROUND_UP(freq, HZ_PER_KHZ));
dfc->cooling_state = state;
return 0;
}
/**
* get_perf_idx() - get the performance index corresponding to a frequency
* @em_pd: Pointer to device's Energy Model
* @freq: frequency in kHz
*
* Return: the performance index associated with the @freq, or
* -EINVAL if it wasn't found.
*/
static int get_perf_idx(struct em_perf_domain *em_pd, unsigned long freq)
{
struct em_perf_state *table;
int i, idx = -EINVAL;
rcu_read_lock();
table = em_perf_state_from_pd(em_pd);
for (i = 0; i < em_pd->nr_perf_states; i++) {
if (table[i].frequency != freq)
continue;
idx = i;
break;
}
rcu_read_unlock();
return idx;
}
static unsigned long get_voltage(struct devfreq *df, unsigned long freq)
{
struct device *dev = df->dev.parent;
unsigned long voltage;
struct dev_pm_opp *opp;
opp = dev_pm_opp_find_freq_exact(dev, freq, true);
if (PTR_ERR(opp) == -ERANGE)
opp = dev_pm_opp_find_freq_exact(dev, freq, false);
if (IS_ERR(opp)) {
dev_err_ratelimited(dev, "Failed to find OPP for frequency %lu: %ld\n",
freq, PTR_ERR(opp));
return 0;
}
voltage = dev_pm_opp_get_voltage(opp) / 1000; /* mV */
dev_pm_opp_put(opp);
if (voltage == 0) {
dev_err_ratelimited(dev,
"Failed to get voltage for frequency %lu\n",
freq);
}
return voltage;
}
static void _normalize_load(struct devfreq_dev_status *status)
{
if (status->total_time > 0xfffff) {
status->total_time >>= 10;
status->busy_time >>= 10;
}
status->busy_time <<= 10;
status->busy_time /= status->total_time ? : 1;
status->busy_time = status->busy_time ? : 1;
status->total_time = 1024;
}
static int devfreq_cooling_get_requested_power(struct thermal_cooling_device *cdev,
u32 *power)
{
struct devfreq_cooling_device *dfc = cdev->devdata;
struct devfreq *df = dfc->devfreq;
struct devfreq_dev_status status;
struct em_perf_state *table;
unsigned long state;
unsigned long freq;
unsigned long voltage;
int res, perf_idx;
mutex_lock(&df->lock);
status = df->last_status;
mutex_unlock(&df->lock);
freq = status.current_frequency;
if (dfc->power_ops && dfc->power_ops->get_real_power) {
voltage = get_voltage(df, freq);
if (voltage == 0) {
res = -EINVAL;
goto fail;
}
res = dfc->power_ops->get_real_power(df, power, freq, voltage);
if (!res) {
state = dfc->max_state - dfc->capped_state;
PM: EM: convert power field to micro-Watts precision and align drivers The milli-Watts precision causes rounding errors while calculating efficiency cost for each OPP. This is especially visible in the 'simple' Energy Model (EM), where the power for each OPP is provided from OPP framework. This can cause some OPPs to be marked inefficient, while using micro-Watts precision that might not happen. Update all EM users which access 'power' field and assume the value is in milli-Watts. Solve also an issue with potential overflow in calculation of energy estimation on 32bit machine. It's needed now since the power value (thus the 'cost' as well) are higher. Example calculation which shows the rounding error and impact: power = 'dyn-power-coeff' * volt_mV * volt_mV * freq_MHz power_a_uW = (100 * 600mW * 600mW * 500MHz) / 10^6 = 18000 power_a_mW = (100 * 600mW * 600mW * 500MHz) / 10^9 = 18 power_b_uW = (100 * 605mW * 605mW * 600MHz) / 10^6 = 21961 power_b_mW = (100 * 605mW * 605mW * 600MHz) / 10^9 = 21 max_freq = 2000MHz cost_a_mW = 18 * 2000MHz/500MHz = 72 cost_a_uW = 18000 * 2000MHz/500MHz = 72000 cost_b_mW = 21 * 2000MHz/600MHz = 70 // <- artificially better cost_b_uW = 21961 * 2000MHz/600MHz = 73203 The 'cost_b_mW' (which is based on old milli-Watts) is misleadingly better that the 'cost_b_uW' (this patch uses micro-Watts) and such would have impact on the 'inefficient OPPs' information in the Cpufreq framework. This patch set removes the rounding issue. Signed-off-by: Lukasz Luba <lukasz.luba@arm.com> Acked-by: Daniel Lezcano <daniel.lezcano@linaro.org> Acked-by: Viresh Kumar <viresh.kumar@linaro.org> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
2022-07-07 15:15:52 +08:00
/* Convert EM power into milli-Watts first */
rcu_read_lock();
table = em_perf_state_from_pd(dfc->em_pd);
dfc->res_util = table[state].power;
rcu_read_unlock();
PM: EM: convert power field to micro-Watts precision and align drivers The milli-Watts precision causes rounding errors while calculating efficiency cost for each OPP. This is especially visible in the 'simple' Energy Model (EM), where the power for each OPP is provided from OPP framework. This can cause some OPPs to be marked inefficient, while using micro-Watts precision that might not happen. Update all EM users which access 'power' field and assume the value is in milli-Watts. Solve also an issue with potential overflow in calculation of energy estimation on 32bit machine. It's needed now since the power value (thus the 'cost' as well) are higher. Example calculation which shows the rounding error and impact: power = 'dyn-power-coeff' * volt_mV * volt_mV * freq_MHz power_a_uW = (100 * 600mW * 600mW * 500MHz) / 10^6 = 18000 power_a_mW = (100 * 600mW * 600mW * 500MHz) / 10^9 = 18 power_b_uW = (100 * 605mW * 605mW * 600MHz) / 10^6 = 21961 power_b_mW = (100 * 605mW * 605mW * 600MHz) / 10^9 = 21 max_freq = 2000MHz cost_a_mW = 18 * 2000MHz/500MHz = 72 cost_a_uW = 18000 * 2000MHz/500MHz = 72000 cost_b_mW = 21 * 2000MHz/600MHz = 70 // <- artificially better cost_b_uW = 21961 * 2000MHz/600MHz = 73203 The 'cost_b_mW' (which is based on old milli-Watts) is misleadingly better that the 'cost_b_uW' (this patch uses micro-Watts) and such would have impact on the 'inefficient OPPs' information in the Cpufreq framework. This patch set removes the rounding issue. Signed-off-by: Lukasz Luba <lukasz.luba@arm.com> Acked-by: Daniel Lezcano <daniel.lezcano@linaro.org> Acked-by: Viresh Kumar <viresh.kumar@linaro.org> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
2022-07-07 15:15:52 +08:00
dfc->res_util /= MICROWATT_PER_MILLIWATT;
dfc->res_util *= SCALE_ERROR_MITIGATION;
if (*power > 1)
dfc->res_util /= *power;
} else {
goto fail;
}
} else {
/* Energy Model frequencies are in kHz */
perf_idx = get_perf_idx(dfc->em_pd, freq / 1000);
if (perf_idx < 0) {
res = -EAGAIN;
goto fail;
}
_normalize_load(&status);
PM: EM: convert power field to micro-Watts precision and align drivers The milli-Watts precision causes rounding errors while calculating efficiency cost for each OPP. This is especially visible in the 'simple' Energy Model (EM), where the power for each OPP is provided from OPP framework. This can cause some OPPs to be marked inefficient, while using micro-Watts precision that might not happen. Update all EM users which access 'power' field and assume the value is in milli-Watts. Solve also an issue with potential overflow in calculation of energy estimation on 32bit machine. It's needed now since the power value (thus the 'cost' as well) are higher. Example calculation which shows the rounding error and impact: power = 'dyn-power-coeff' * volt_mV * volt_mV * freq_MHz power_a_uW = (100 * 600mW * 600mW * 500MHz) / 10^6 = 18000 power_a_mW = (100 * 600mW * 600mW * 500MHz) / 10^9 = 18 power_b_uW = (100 * 605mW * 605mW * 600MHz) / 10^6 = 21961 power_b_mW = (100 * 605mW * 605mW * 600MHz) / 10^9 = 21 max_freq = 2000MHz cost_a_mW = 18 * 2000MHz/500MHz = 72 cost_a_uW = 18000 * 2000MHz/500MHz = 72000 cost_b_mW = 21 * 2000MHz/600MHz = 70 // <- artificially better cost_b_uW = 21961 * 2000MHz/600MHz = 73203 The 'cost_b_mW' (which is based on old milli-Watts) is misleadingly better that the 'cost_b_uW' (this patch uses micro-Watts) and such would have impact on the 'inefficient OPPs' information in the Cpufreq framework. This patch set removes the rounding issue. Signed-off-by: Lukasz Luba <lukasz.luba@arm.com> Acked-by: Daniel Lezcano <daniel.lezcano@linaro.org> Acked-by: Viresh Kumar <viresh.kumar@linaro.org> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
2022-07-07 15:15:52 +08:00
/* Convert EM power into milli-Watts first */
rcu_read_lock();
table = em_perf_state_from_pd(dfc->em_pd);
*power = table[perf_idx].power;
rcu_read_unlock();
PM: EM: convert power field to micro-Watts precision and align drivers The milli-Watts precision causes rounding errors while calculating efficiency cost for each OPP. This is especially visible in the 'simple' Energy Model (EM), where the power for each OPP is provided from OPP framework. This can cause some OPPs to be marked inefficient, while using micro-Watts precision that might not happen. Update all EM users which access 'power' field and assume the value is in milli-Watts. Solve also an issue with potential overflow in calculation of energy estimation on 32bit machine. It's needed now since the power value (thus the 'cost' as well) are higher. Example calculation which shows the rounding error and impact: power = 'dyn-power-coeff' * volt_mV * volt_mV * freq_MHz power_a_uW = (100 * 600mW * 600mW * 500MHz) / 10^6 = 18000 power_a_mW = (100 * 600mW * 600mW * 500MHz) / 10^9 = 18 power_b_uW = (100 * 605mW * 605mW * 600MHz) / 10^6 = 21961 power_b_mW = (100 * 605mW * 605mW * 600MHz) / 10^9 = 21 max_freq = 2000MHz cost_a_mW = 18 * 2000MHz/500MHz = 72 cost_a_uW = 18000 * 2000MHz/500MHz = 72000 cost_b_mW = 21 * 2000MHz/600MHz = 70 // <- artificially better cost_b_uW = 21961 * 2000MHz/600MHz = 73203 The 'cost_b_mW' (which is based on old milli-Watts) is misleadingly better that the 'cost_b_uW' (this patch uses micro-Watts) and such would have impact on the 'inefficient OPPs' information in the Cpufreq framework. This patch set removes the rounding issue. Signed-off-by: Lukasz Luba <lukasz.luba@arm.com> Acked-by: Daniel Lezcano <daniel.lezcano@linaro.org> Acked-by: Viresh Kumar <viresh.kumar@linaro.org> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
2022-07-07 15:15:52 +08:00
*power /= MICROWATT_PER_MILLIWATT;
/* Scale power for utilization */
*power *= status.busy_time;
*power >>= 10;
}
trace_thermal_power_devfreq_get_power(cdev, &status, freq, *power);
return 0;
fail:
/* It is safe to set max in this case */
dfc->res_util = SCALE_ERROR_MITIGATION;
return res;
}
static int devfreq_cooling_state2power(struct thermal_cooling_device *cdev,
unsigned long state, u32 *power)
{
struct devfreq_cooling_device *dfc = cdev->devdata;
struct em_perf_state *table;
int perf_idx;
if (state > dfc->max_state)
return -EINVAL;
perf_idx = dfc->max_state - state;
rcu_read_lock();
table = em_perf_state_from_pd(dfc->em_pd);
*power = table[perf_idx].power;
rcu_read_unlock();
PM: EM: convert power field to micro-Watts precision and align drivers The milli-Watts precision causes rounding errors while calculating efficiency cost for each OPP. This is especially visible in the 'simple' Energy Model (EM), where the power for each OPP is provided from OPP framework. This can cause some OPPs to be marked inefficient, while using micro-Watts precision that might not happen. Update all EM users which access 'power' field and assume the value is in milli-Watts. Solve also an issue with potential overflow in calculation of energy estimation on 32bit machine. It's needed now since the power value (thus the 'cost' as well) are higher. Example calculation which shows the rounding error and impact: power = 'dyn-power-coeff' * volt_mV * volt_mV * freq_MHz power_a_uW = (100 * 600mW * 600mW * 500MHz) / 10^6 = 18000 power_a_mW = (100 * 600mW * 600mW * 500MHz) / 10^9 = 18 power_b_uW = (100 * 605mW * 605mW * 600MHz) / 10^6 = 21961 power_b_mW = (100 * 605mW * 605mW * 600MHz) / 10^9 = 21 max_freq = 2000MHz cost_a_mW = 18 * 2000MHz/500MHz = 72 cost_a_uW = 18000 * 2000MHz/500MHz = 72000 cost_b_mW = 21 * 2000MHz/600MHz = 70 // <- artificially better cost_b_uW = 21961 * 2000MHz/600MHz = 73203 The 'cost_b_mW' (which is based on old milli-Watts) is misleadingly better that the 'cost_b_uW' (this patch uses micro-Watts) and such would have impact on the 'inefficient OPPs' information in the Cpufreq framework. This patch set removes the rounding issue. Signed-off-by: Lukasz Luba <lukasz.luba@arm.com> Acked-by: Daniel Lezcano <daniel.lezcano@linaro.org> Acked-by: Viresh Kumar <viresh.kumar@linaro.org> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
2022-07-07 15:15:52 +08:00
*power /= MICROWATT_PER_MILLIWATT;
return 0;
}
static int devfreq_cooling_power2state(struct thermal_cooling_device *cdev,
u32 power, unsigned long *state)
{
struct devfreq_cooling_device *dfc = cdev->devdata;
struct devfreq *df = dfc->devfreq;
struct devfreq_dev_status status;
PM: EM: convert power field to micro-Watts precision and align drivers The milli-Watts precision causes rounding errors while calculating efficiency cost for each OPP. This is especially visible in the 'simple' Energy Model (EM), where the power for each OPP is provided from OPP framework. This can cause some OPPs to be marked inefficient, while using micro-Watts precision that might not happen. Update all EM users which access 'power' field and assume the value is in milli-Watts. Solve also an issue with potential overflow in calculation of energy estimation on 32bit machine. It's needed now since the power value (thus the 'cost' as well) are higher. Example calculation which shows the rounding error and impact: power = 'dyn-power-coeff' * volt_mV * volt_mV * freq_MHz power_a_uW = (100 * 600mW * 600mW * 500MHz) / 10^6 = 18000 power_a_mW = (100 * 600mW * 600mW * 500MHz) / 10^9 = 18 power_b_uW = (100 * 605mW * 605mW * 600MHz) / 10^6 = 21961 power_b_mW = (100 * 605mW * 605mW * 600MHz) / 10^9 = 21 max_freq = 2000MHz cost_a_mW = 18 * 2000MHz/500MHz = 72 cost_a_uW = 18000 * 2000MHz/500MHz = 72000 cost_b_mW = 21 * 2000MHz/600MHz = 70 // <- artificially better cost_b_uW = 21961 * 2000MHz/600MHz = 73203 The 'cost_b_mW' (which is based on old milli-Watts) is misleadingly better that the 'cost_b_uW' (this patch uses micro-Watts) and such would have impact on the 'inefficient OPPs' information in the Cpufreq framework. This patch set removes the rounding issue. Signed-off-by: Lukasz Luba <lukasz.luba@arm.com> Acked-by: Daniel Lezcano <daniel.lezcano@linaro.org> Acked-by: Viresh Kumar <viresh.kumar@linaro.org> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
2022-07-07 15:15:52 +08:00
unsigned long freq, em_power_mw;
struct em_perf_state *table;
s32 est_power;
int i;
mutex_lock(&df->lock);
status = df->last_status;
mutex_unlock(&df->lock);
freq = status.current_frequency;
if (dfc->power_ops && dfc->power_ops->get_real_power) {
/* Scale for resource utilization */
est_power = power * dfc->res_util;
est_power /= SCALE_ERROR_MITIGATION;
} else {
/* Scale dynamic power for utilization */
_normalize_load(&status);
est_power = power << 10;
est_power /= status.busy_time;
}
/*
* Find the first cooling state that is within the power
* budget. The EM power table is sorted ascending.
*/
rcu_read_lock();
table = em_perf_state_from_pd(dfc->em_pd);
PM: EM: convert power field to micro-Watts precision and align drivers The milli-Watts precision causes rounding errors while calculating efficiency cost for each OPP. This is especially visible in the 'simple' Energy Model (EM), where the power for each OPP is provided from OPP framework. This can cause some OPPs to be marked inefficient, while using micro-Watts precision that might not happen. Update all EM users which access 'power' field and assume the value is in milli-Watts. Solve also an issue with potential overflow in calculation of energy estimation on 32bit machine. It's needed now since the power value (thus the 'cost' as well) are higher. Example calculation which shows the rounding error and impact: power = 'dyn-power-coeff' * volt_mV * volt_mV * freq_MHz power_a_uW = (100 * 600mW * 600mW * 500MHz) / 10^6 = 18000 power_a_mW = (100 * 600mW * 600mW * 500MHz) / 10^9 = 18 power_b_uW = (100 * 605mW * 605mW * 600MHz) / 10^6 = 21961 power_b_mW = (100 * 605mW * 605mW * 600MHz) / 10^9 = 21 max_freq = 2000MHz cost_a_mW = 18 * 2000MHz/500MHz = 72 cost_a_uW = 18000 * 2000MHz/500MHz = 72000 cost_b_mW = 21 * 2000MHz/600MHz = 70 // <- artificially better cost_b_uW = 21961 * 2000MHz/600MHz = 73203 The 'cost_b_mW' (which is based on old milli-Watts) is misleadingly better that the 'cost_b_uW' (this patch uses micro-Watts) and such would have impact on the 'inefficient OPPs' information in the Cpufreq framework. This patch set removes the rounding issue. Signed-off-by: Lukasz Luba <lukasz.luba@arm.com> Acked-by: Daniel Lezcano <daniel.lezcano@linaro.org> Acked-by: Viresh Kumar <viresh.kumar@linaro.org> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
2022-07-07 15:15:52 +08:00
for (i = dfc->max_state; i > 0; i--) {
/* Convert EM power to milli-Watts to make safe comparison */
em_power_mw = table[i].power;
PM: EM: convert power field to micro-Watts precision and align drivers The milli-Watts precision causes rounding errors while calculating efficiency cost for each OPP. This is especially visible in the 'simple' Energy Model (EM), where the power for each OPP is provided from OPP framework. This can cause some OPPs to be marked inefficient, while using micro-Watts precision that might not happen. Update all EM users which access 'power' field and assume the value is in milli-Watts. Solve also an issue with potential overflow in calculation of energy estimation on 32bit machine. It's needed now since the power value (thus the 'cost' as well) are higher. Example calculation which shows the rounding error and impact: power = 'dyn-power-coeff' * volt_mV * volt_mV * freq_MHz power_a_uW = (100 * 600mW * 600mW * 500MHz) / 10^6 = 18000 power_a_mW = (100 * 600mW * 600mW * 500MHz) / 10^9 = 18 power_b_uW = (100 * 605mW * 605mW * 600MHz) / 10^6 = 21961 power_b_mW = (100 * 605mW * 605mW * 600MHz) / 10^9 = 21 max_freq = 2000MHz cost_a_mW = 18 * 2000MHz/500MHz = 72 cost_a_uW = 18000 * 2000MHz/500MHz = 72000 cost_b_mW = 21 * 2000MHz/600MHz = 70 // <- artificially better cost_b_uW = 21961 * 2000MHz/600MHz = 73203 The 'cost_b_mW' (which is based on old milli-Watts) is misleadingly better that the 'cost_b_uW' (this patch uses micro-Watts) and such would have impact on the 'inefficient OPPs' information in the Cpufreq framework. This patch set removes the rounding issue. Signed-off-by: Lukasz Luba <lukasz.luba@arm.com> Acked-by: Daniel Lezcano <daniel.lezcano@linaro.org> Acked-by: Viresh Kumar <viresh.kumar@linaro.org> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
2022-07-07 15:15:52 +08:00
em_power_mw /= MICROWATT_PER_MILLIWATT;
if (est_power >= em_power_mw)
break;
PM: EM: convert power field to micro-Watts precision and align drivers The milli-Watts precision causes rounding errors while calculating efficiency cost for each OPP. This is especially visible in the 'simple' Energy Model (EM), where the power for each OPP is provided from OPP framework. This can cause some OPPs to be marked inefficient, while using micro-Watts precision that might not happen. Update all EM users which access 'power' field and assume the value is in milli-Watts. Solve also an issue with potential overflow in calculation of energy estimation on 32bit machine. It's needed now since the power value (thus the 'cost' as well) are higher. Example calculation which shows the rounding error and impact: power = 'dyn-power-coeff' * volt_mV * volt_mV * freq_MHz power_a_uW = (100 * 600mW * 600mW * 500MHz) / 10^6 = 18000 power_a_mW = (100 * 600mW * 600mW * 500MHz) / 10^9 = 18 power_b_uW = (100 * 605mW * 605mW * 600MHz) / 10^6 = 21961 power_b_mW = (100 * 605mW * 605mW * 600MHz) / 10^9 = 21 max_freq = 2000MHz cost_a_mW = 18 * 2000MHz/500MHz = 72 cost_a_uW = 18000 * 2000MHz/500MHz = 72000 cost_b_mW = 21 * 2000MHz/600MHz = 70 // <- artificially better cost_b_uW = 21961 * 2000MHz/600MHz = 73203 The 'cost_b_mW' (which is based on old milli-Watts) is misleadingly better that the 'cost_b_uW' (this patch uses micro-Watts) and such would have impact on the 'inefficient OPPs' information in the Cpufreq framework. This patch set removes the rounding issue. Signed-off-by: Lukasz Luba <lukasz.luba@arm.com> Acked-by: Daniel Lezcano <daniel.lezcano@linaro.org> Acked-by: Viresh Kumar <viresh.kumar@linaro.org> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
2022-07-07 15:15:52 +08:00
}
rcu_read_unlock();
*state = dfc->max_state - i;
dfc->capped_state = *state;
trace_thermal_power_devfreq_limit(cdev, freq, *state, power);
return 0;
}
/**
* devfreq_cooling_gen_tables() - Generate frequency table.
* @dfc: Pointer to devfreq cooling device.
* @num_opps: Number of OPPs
*
* Generate frequency table which holds the frequencies in descending
* order. That way its indexed by cooling device state. This is for
* compatibility with drivers which do not register Energy Model.
*
* Return: 0 on success, negative error code on failure.
*/
static int devfreq_cooling_gen_tables(struct devfreq_cooling_device *dfc,
int num_opps)
{
struct devfreq *df = dfc->devfreq;
struct device *dev = df->dev.parent;
unsigned long freq;
int i;
dfc->freq_table = kcalloc(num_opps, sizeof(*dfc->freq_table),
GFP_KERNEL);
if (!dfc->freq_table)
return -ENOMEM;
for (i = 0, freq = ULONG_MAX; i < num_opps; i++, freq--) {
struct dev_pm_opp *opp;
opp = dev_pm_opp_find_freq_floor(dev, &freq);
if (IS_ERR(opp)) {
kfree(dfc->freq_table);
return PTR_ERR(opp);
}
dev_pm_opp_put(opp);
dfc->freq_table[i] = freq;
}
return 0;
}
/**
* of_devfreq_cooling_register_power() - Register devfreq cooling device,
* with OF and power information.
* @np: Pointer to OF device_node.
* @df: Pointer to devfreq device.
* @dfc_power: Pointer to devfreq_cooling_power.
*
* Register a devfreq cooling device. The available OPPs must be
* registered on the device.
*
* If @dfc_power is provided, the cooling device is registered with the
* power extensions. For the power extensions to work correctly,
* devfreq should use the simple_ondemand governor, other governors
* are not currently supported.
*/
struct thermal_cooling_device *
of_devfreq_cooling_register_power(struct device_node *np, struct devfreq *df,
struct devfreq_cooling_power *dfc_power)
{
struct thermal_cooling_device *cdev;
struct device *dev = df->dev.parent;
struct devfreq_cooling_device *dfc;
struct em_perf_domain *em;
struct thermal_cooling_device_ops *ops;
char *name;
int err, num_opps;
dfc = kzalloc(sizeof(*dfc), GFP_KERNEL);
if (!dfc)
return ERR_PTR(-ENOMEM);
dfc->devfreq = df;
ops = &dfc->cooling_ops;
ops->get_max_state = devfreq_cooling_get_max_state;
ops->get_cur_state = devfreq_cooling_get_cur_state;
ops->set_cur_state = devfreq_cooling_set_cur_state;
em = em_pd_get(dev);
if (em && !em_is_artificial(em)) {
dfc->em_pd = em;
ops->get_requested_power =
devfreq_cooling_get_requested_power;
ops->state2power = devfreq_cooling_state2power;
ops->power2state = devfreq_cooling_power2state;
dfc->power_ops = dfc_power;
num_opps = em_pd_nr_perf_states(dfc->em_pd);
} else {
/* Backward compatibility for drivers which do not use IPA */
dev_dbg(dev, "missing proper EM for cooling device\n");
num_opps = dev_pm_opp_get_opp_count(dev);
err = devfreq_cooling_gen_tables(dfc, num_opps);
if (err)
goto free_dfc;
}
if (num_opps <= 0) {
err = -EINVAL;
goto free_dfc;
}
/* max_state is an index, not a counter */
dfc->max_state = num_opps - 1;
err = dev_pm_qos_add_request(dev, &dfc->req_max_freq,
DEV_PM_QOS_MAX_FREQUENCY,
PM_QOS_MAX_FREQUENCY_DEFAULT_VALUE);
if (err < 0)
goto free_table;
err = -ENOMEM;
name = kasprintf(GFP_KERNEL, "devfreq-%s", dev_name(dev));
if (!name)
goto remove_qos_req;
cdev = thermal_of_cooling_device_register(np, name, dfc, ops);
kfree(name);
if (IS_ERR(cdev)) {
err = PTR_ERR(cdev);
dev_err(dev,
"Failed to register devfreq cooling device (%d)\n",
err);
goto remove_qos_req;
}
dfc->cdev = cdev;
return cdev;
remove_qos_req:
dev_pm_qos_remove_request(&dfc->req_max_freq);
free_table:
kfree(dfc->freq_table);
free_dfc:
kfree(dfc);
return ERR_PTR(err);
}
EXPORT_SYMBOL_GPL(of_devfreq_cooling_register_power);
/**
* of_devfreq_cooling_register() - Register devfreq cooling device,
* with OF information.
* @np: Pointer to OF device_node.
* @df: Pointer to devfreq device.
*/
struct thermal_cooling_device *
of_devfreq_cooling_register(struct device_node *np, struct devfreq *df)
{
return of_devfreq_cooling_register_power(np, df, NULL);
}
EXPORT_SYMBOL_GPL(of_devfreq_cooling_register);
/**
* devfreq_cooling_register() - Register devfreq cooling device.
* @df: Pointer to devfreq device.
*/
struct thermal_cooling_device *devfreq_cooling_register(struct devfreq *df)
{
return of_devfreq_cooling_register(NULL, df);
}
EXPORT_SYMBOL_GPL(devfreq_cooling_register);
/**
* devfreq_cooling_em_register() - Register devfreq cooling device with
* power information and automatically register Energy Model (EM)
* @df: Pointer to devfreq device.
* @dfc_power: Pointer to devfreq_cooling_power.
*
* Register a devfreq cooling device and automatically register EM. The
* available OPPs must be registered for the device.
*
* If @dfc_power is provided, the cooling device is registered with the
* power extensions. It is using the simple Energy Model which requires
* "dynamic-power-coefficient" a devicetree property. To not break drivers
* which miss that DT property, the function won't bail out when the EM
* registration failed. The cooling device will be registered if everything
* else is OK.
*/
struct thermal_cooling_device *
devfreq_cooling_em_register(struct devfreq *df,
struct devfreq_cooling_power *dfc_power)
{
struct thermal_cooling_device *cdev;
struct device *dev;
int ret;
if (IS_ERR_OR_NULL(df))
return ERR_PTR(-EINVAL);
dev = df->dev.parent;
ret = dev_pm_opp_of_register_em(dev, NULL);
if (ret)
dev_dbg(dev, "Unable to register EM for devfreq cooling device (%d)\n",
ret);
cdev = of_devfreq_cooling_register_power(dev->of_node, df, dfc_power);
if (IS_ERR_OR_NULL(cdev))
em_dev_unregister_perf_domain(dev);
return cdev;
}
EXPORT_SYMBOL_GPL(devfreq_cooling_em_register);
/**
* devfreq_cooling_unregister() - Unregister devfreq cooling device.
* @cdev: Pointer to devfreq cooling device to unregister.
*
* Unregisters devfreq cooling device and related Energy Model if it was
* present.
*/
void devfreq_cooling_unregister(struct thermal_cooling_device *cdev)
{
struct devfreq_cooling_device *dfc;
struct device *dev;
if (IS_ERR_OR_NULL(cdev))
return;
dfc = cdev->devdata;
dev = dfc->devfreq->dev.parent;
thermal_cooling_device_unregister(dfc->cdev);
dev_pm_qos_remove_request(&dfc->req_max_freq);
em_dev_unregister_perf_domain(dev);
kfree(dfc->freq_table);
kfree(dfc);
}
EXPORT_SYMBOL_GPL(devfreq_cooling_unregister);