linux_dsm_epyc7002/drivers/cpufreq/cpufreq-dt.c

395 lines
9.5 KiB
C
Raw Normal View History

// SPDX-License-Identifier: GPL-2.0-only
/*
* Copyright (C) 2012 Freescale Semiconductor, Inc.
*
* Copyright (C) 2014 Linaro.
* Viresh Kumar <viresh.kumar@linaro.org>
*/
#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
#include <linux/clk.h>
#include <linux/cpu.h>
#include <linux/cpufreq.h>
#include <linux/cpumask.h>
#include <linux/err.h>
cpufreq: dt: Refactor initialization to handle probe deferral properly cpufreq-dt is currently unable to handle -EPROBE_DEFER properly because the error code is not propagated for the cpufreq_driver->init() callback. Instead, it attempts to avoid the situation by temporarily requesting all resources within resources_available() and releasing them again immediately after. This has several disadvantages: - Whenever we add something like interconnect handling to the OPP core we need to patch cpufreq-dt to request these resources early. - resources_available() is only run for CPU0, but other clusters may eventually depend on other resources that are not available yet. (See FIXME comment removed by this commit...) - All resources need to be looked up several times. Now that the OPP core can propagate -EPROBE_DEFER during initialization, it would be nice to avoid all that trouble and just propagate its error code when necessary. This commit refactors the cpufreq-dt driver to initialize private_data before registering the cpufreq driver. We do this by iterating over all possible CPUs and ensure that all resources are initialized: 1. dev_pm_opp_get_opp_table() ensures the OPP table is allocated and initialized with clock and interconnects. 2. dev_pm_opp_set_regulators() requests the regulators and assigns them to the OPP table. 3. We call dev_pm_opp_of_get_sharing_cpus() early so that we only initialize the OPP table once for each shared policy. With these changes, we actually end up saving a few lines of code, the resources are no longer looked up multiple times and everything should be much more robust. Signed-off-by: Stephan Gerhold <stephan@gerhold.net> [ Viresh: Use list_head structure for maintaining the list and minor changes ] Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
2020-07-27 16:30:47 +07:00
#include <linux/list.h>
#include <linux/module.h>
#include <linux/of.h>
#include <linux/pm_opp.h>
#include <linux/platform_device.h>
#include <linux/regulator/consumer.h>
#include <linux/slab.h>
#include <linux/thermal.h>
#include "cpufreq-dt.h"
struct private_data {
cpufreq: dt: Refactor initialization to handle probe deferral properly cpufreq-dt is currently unable to handle -EPROBE_DEFER properly because the error code is not propagated for the cpufreq_driver->init() callback. Instead, it attempts to avoid the situation by temporarily requesting all resources within resources_available() and releasing them again immediately after. This has several disadvantages: - Whenever we add something like interconnect handling to the OPP core we need to patch cpufreq-dt to request these resources early. - resources_available() is only run for CPU0, but other clusters may eventually depend on other resources that are not available yet. (See FIXME comment removed by this commit...) - All resources need to be looked up several times. Now that the OPP core can propagate -EPROBE_DEFER during initialization, it would be nice to avoid all that trouble and just propagate its error code when necessary. This commit refactors the cpufreq-dt driver to initialize private_data before registering the cpufreq driver. We do this by iterating over all possible CPUs and ensure that all resources are initialized: 1. dev_pm_opp_get_opp_table() ensures the OPP table is allocated and initialized with clock and interconnects. 2. dev_pm_opp_set_regulators() requests the regulators and assigns them to the OPP table. 3. We call dev_pm_opp_of_get_sharing_cpus() early so that we only initialize the OPP table once for each shared policy. With these changes, we actually end up saving a few lines of code, the resources are no longer looked up multiple times and everything should be much more robust. Signed-off-by: Stephan Gerhold <stephan@gerhold.net> [ Viresh: Use list_head structure for maintaining the list and minor changes ] Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
2020-07-27 16:30:47 +07:00
struct list_head node;
cpumask_var_t cpus;
struct device *cpu_dev;
cpufreq: dt: Refactor initialization to handle probe deferral properly cpufreq-dt is currently unable to handle -EPROBE_DEFER properly because the error code is not propagated for the cpufreq_driver->init() callback. Instead, it attempts to avoid the situation by temporarily requesting all resources within resources_available() and releasing them again immediately after. This has several disadvantages: - Whenever we add something like interconnect handling to the OPP core we need to patch cpufreq-dt to request these resources early. - resources_available() is only run for CPU0, but other clusters may eventually depend on other resources that are not available yet. (See FIXME comment removed by this commit...) - All resources need to be looked up several times. Now that the OPP core can propagate -EPROBE_DEFER during initialization, it would be nice to avoid all that trouble and just propagate its error code when necessary. This commit refactors the cpufreq-dt driver to initialize private_data before registering the cpufreq driver. We do this by iterating over all possible CPUs and ensure that all resources are initialized: 1. dev_pm_opp_get_opp_table() ensures the OPP table is allocated and initialized with clock and interconnects. 2. dev_pm_opp_set_regulators() requests the regulators and assigns them to the OPP table. 3. We call dev_pm_opp_of_get_sharing_cpus() early so that we only initialize the OPP table once for each shared policy. With these changes, we actually end up saving a few lines of code, the resources are no longer looked up multiple times and everything should be much more robust. Signed-off-by: Stephan Gerhold <stephan@gerhold.net> [ Viresh: Use list_head structure for maintaining the list and minor changes ] Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
2020-07-27 16:30:47 +07:00
struct opp_table *opp_table;
struct opp_table *reg_opp_table;
bool have_static_opps;
};
cpufreq: dt: Refactor initialization to handle probe deferral properly cpufreq-dt is currently unable to handle -EPROBE_DEFER properly because the error code is not propagated for the cpufreq_driver->init() callback. Instead, it attempts to avoid the situation by temporarily requesting all resources within resources_available() and releasing them again immediately after. This has several disadvantages: - Whenever we add something like interconnect handling to the OPP core we need to patch cpufreq-dt to request these resources early. - resources_available() is only run for CPU0, but other clusters may eventually depend on other resources that are not available yet. (See FIXME comment removed by this commit...) - All resources need to be looked up several times. Now that the OPP core can propagate -EPROBE_DEFER during initialization, it would be nice to avoid all that trouble and just propagate its error code when necessary. This commit refactors the cpufreq-dt driver to initialize private_data before registering the cpufreq driver. We do this by iterating over all possible CPUs and ensure that all resources are initialized: 1. dev_pm_opp_get_opp_table() ensures the OPP table is allocated and initialized with clock and interconnects. 2. dev_pm_opp_set_regulators() requests the regulators and assigns them to the OPP table. 3. We call dev_pm_opp_of_get_sharing_cpus() early so that we only initialize the OPP table once for each shared policy. With these changes, we actually end up saving a few lines of code, the resources are no longer looked up multiple times and everything should be much more robust. Signed-off-by: Stephan Gerhold <stephan@gerhold.net> [ Viresh: Use list_head structure for maintaining the list and minor changes ] Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
2020-07-27 16:30:47 +07:00
static LIST_HEAD(priv_list);
static struct freq_attr *cpufreq_dt_attr[] = {
&cpufreq_freq_attr_scaling_available_freqs,
NULL, /* Extra space for boost-attr if required */
NULL,
};
cpufreq: dt: Refactor initialization to handle probe deferral properly cpufreq-dt is currently unable to handle -EPROBE_DEFER properly because the error code is not propagated for the cpufreq_driver->init() callback. Instead, it attempts to avoid the situation by temporarily requesting all resources within resources_available() and releasing them again immediately after. This has several disadvantages: - Whenever we add something like interconnect handling to the OPP core we need to patch cpufreq-dt to request these resources early. - resources_available() is only run for CPU0, but other clusters may eventually depend on other resources that are not available yet. (See FIXME comment removed by this commit...) - All resources need to be looked up several times. Now that the OPP core can propagate -EPROBE_DEFER during initialization, it would be nice to avoid all that trouble and just propagate its error code when necessary. This commit refactors the cpufreq-dt driver to initialize private_data before registering the cpufreq driver. We do this by iterating over all possible CPUs and ensure that all resources are initialized: 1. dev_pm_opp_get_opp_table() ensures the OPP table is allocated and initialized with clock and interconnects. 2. dev_pm_opp_set_regulators() requests the regulators and assigns them to the OPP table. 3. We call dev_pm_opp_of_get_sharing_cpus() early so that we only initialize the OPP table once for each shared policy. With these changes, we actually end up saving a few lines of code, the resources are no longer looked up multiple times and everything should be much more robust. Signed-off-by: Stephan Gerhold <stephan@gerhold.net> [ Viresh: Use list_head structure for maintaining the list and minor changes ] Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
2020-07-27 16:30:47 +07:00
static struct private_data *cpufreq_dt_find_data(int cpu)
{
struct private_data *priv;
list_for_each_entry(priv, &priv_list, node) {
if (cpumask_test_cpu(cpu, priv->cpus))
return priv;
}
return NULL;
}
static int set_target(struct cpufreq_policy *policy, unsigned int index)
{
struct private_data *priv = policy->driver_data;
unsigned long freq = policy->freq_table[index].frequency;
cpufreq: move invariance setter calls in cpufreq core To properly scale its per-entity load-tracking signals, the task scheduler needs to be given a frequency scale factor, i.e. some image of the current frequency the CPU is running at. Currently, this scale can be computed either by using counters (APERF/MPERF on x86, AMU on arm64), or by piggy-backing on the frequency selection done by cpufreq. For the latter, drivers have to explicitly set the scale factor themselves, despite it being purely boiler-plate code: the required information depends entirely on the kind of frequency switch callback implemented by the driver, i.e. either of: target_index(), target(), fast_switch() and setpolicy(). The fitness of those callbacks with regard to driving the Frequency Invariance Engine (FIE) is studied below: target_index() ============== Documentation states that the chosen frequency "must be determined by freq_table[index].frequency". It isn't clear if it *has* to be that frequency, or if it can use that frequency value to do some computation that ultimately leads to a different frequency selection. All drivers go for the former, while the vexpress-spc-cpufreq has an atypical implementation which is handled separately. Therefore, the hook works on the assumption the core can use freq_table[index].frequency. target() ======= This has been flagged as deprecated since: commit 9c0ebcf78fde ("cpufreq: Implement light weight ->target_index() routine") It also doesn't have that many users: gx-suspmod.c:439: .target = cpufreq_gx_target, s3c24xx-cpufreq.c:428: .target = s3c_cpufreq_target, intel_pstate.c:2528: .target = intel_cpufreq_target, cppc_cpufreq.c:401: .target = cppc_cpufreq_set_target, cpufreq-nforce2.c:371: .target = nforce2_target, sh-cpufreq.c:163: .target = sh_cpufreq_target, pcc-cpufreq.c:573: .target = pcc_cpufreq_target, Similarly to the path taken for target_index() calls in the cpufreq core during a frequency change, all of the drivers above will mark the end of a frequency change by a call to cpufreq_freq_transition_end(). Therefore, cpufreq_freq_transition_end() can be used as the location for the arch_set_freq_scale() call to potentially inform the scheduler of the frequency change. This change maintains the previous functionality for the drivers that implement the target_index() callback, while also adding support for the few drivers that implement the deprecated target() callback. fast_switch() ============= This callback *has* to return the frequency that was selected. setpolicy() =========== This callback does not have any designated way of informing what was the end choice. But there are only two drivers using setpolicy(), and none of them have current FIE support: drivers/cpufreq/longrun.c:281: .setpolicy = longrun_set_policy, drivers/cpufreq/intel_pstate.c:2215: .setpolicy = intel_pstate_set_policy, The intel_pstate is known to use counter-driven frequency invariance. Conclusion ========== Given that the significant majority of current FIE enabled drivers use callbacks that lend themselves to triggering the setting of the FIE scale factor in a generic way, move the invariance setter calls to cpufreq core. As a result of setting the frequency scale factor in cpufreq core, after callbacks that lend themselves to trigger it, remove this functionality from the driver side. To be noted that despite marking a successful frequency change, many cpufreq drivers will consider the new frequency as the requested frequency, although this is might not be the one granted by the hardware. Therefore, the call to arch_set_freq_scale() is a "best effort" one, and it is up to the architecture if the new frequency is used in the new frequency scale factor setting (determined by the implementation of arch_set_freq_scale()) or eventually used by the scheduler (determined by the implementation of arch_scale_freq_capacity()). The architecture is in a better position to decide if it has better methods to obtain more accurate information regarding the current frequency and use that information instead (for example, the use of counters). Also, the implementation to arch_set_freq_scale() will now have to handle error conditions (current frequency == 0) in order to prevent the overhead in cpufreq core when the default arch_set_freq_scale() implementation is used. Signed-off-by: Ionela Voinescu <ionela.voinescu@arm.com> Suggested-by: Valentin Schneider <valentin.schneider@arm.com> Acked-by: Viresh Kumar <viresh.kumar@linaro.org> Acked-by: Sudeep Holla <sudeep.holla@arm.com> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
2020-09-02 03:55:46 +07:00
return dev_pm_opp_set_rate(priv->cpu_dev, freq * 1000);
}
/*
* An earlier version of opp-v1 bindings used to name the regulator
* "cpu0-supply", we still need to handle that for backwards compatibility.
*/
static const char *find_supply_name(struct device *dev)
{
struct device_node *np;
struct property *pp;
int cpu = dev->id;
const char *name = NULL;
np = of_node_get(dev->of_node);
/* This must be valid for sure */
if (WARN_ON(!np))
return NULL;
/* Try "cpu0" for older DTs */
if (!cpu) {
pp = of_find_property(np, "cpu0-supply", NULL);
if (pp) {
name = "cpu0";
goto node_put;
}
}
pp = of_find_property(np, "cpu-supply", NULL);
if (pp) {
name = "cpu";
goto node_put;
}
dev_dbg(dev, "no regulator for cpu%d\n", cpu);
node_put:
of_node_put(np);
return name;
}
static int cpufreq_init(struct cpufreq_policy *policy)
{
struct cpufreq_frequency_table *freq_table;
struct private_data *priv;
struct device *cpu_dev;
struct clk *cpu_clk;
unsigned int transition_latency;
int ret;
cpufreq: dt: Refactor initialization to handle probe deferral properly cpufreq-dt is currently unable to handle -EPROBE_DEFER properly because the error code is not propagated for the cpufreq_driver->init() callback. Instead, it attempts to avoid the situation by temporarily requesting all resources within resources_available() and releasing them again immediately after. This has several disadvantages: - Whenever we add something like interconnect handling to the OPP core we need to patch cpufreq-dt to request these resources early. - resources_available() is only run for CPU0, but other clusters may eventually depend on other resources that are not available yet. (See FIXME comment removed by this commit...) - All resources need to be looked up several times. Now that the OPP core can propagate -EPROBE_DEFER during initialization, it would be nice to avoid all that trouble and just propagate its error code when necessary. This commit refactors the cpufreq-dt driver to initialize private_data before registering the cpufreq driver. We do this by iterating over all possible CPUs and ensure that all resources are initialized: 1. dev_pm_opp_get_opp_table() ensures the OPP table is allocated and initialized with clock and interconnects. 2. dev_pm_opp_set_regulators() requests the regulators and assigns them to the OPP table. 3. We call dev_pm_opp_of_get_sharing_cpus() early so that we only initialize the OPP table once for each shared policy. With these changes, we actually end up saving a few lines of code, the resources are no longer looked up multiple times and everything should be much more robust. Signed-off-by: Stephan Gerhold <stephan@gerhold.net> [ Viresh: Use list_head structure for maintaining the list and minor changes ] Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
2020-07-27 16:30:47 +07:00
priv = cpufreq_dt_find_data(policy->cpu);
if (!priv) {
pr_err("failed to find data for cpu%d\n", policy->cpu);
return -ENODEV;
}
cpufreq: dt: Refactor initialization to handle probe deferral properly cpufreq-dt is currently unable to handle -EPROBE_DEFER properly because the error code is not propagated for the cpufreq_driver->init() callback. Instead, it attempts to avoid the situation by temporarily requesting all resources within resources_available() and releasing them again immediately after. This has several disadvantages: - Whenever we add something like interconnect handling to the OPP core we need to patch cpufreq-dt to request these resources early. - resources_available() is only run for CPU0, but other clusters may eventually depend on other resources that are not available yet. (See FIXME comment removed by this commit...) - All resources need to be looked up several times. Now that the OPP core can propagate -EPROBE_DEFER during initialization, it would be nice to avoid all that trouble and just propagate its error code when necessary. This commit refactors the cpufreq-dt driver to initialize private_data before registering the cpufreq driver. We do this by iterating over all possible CPUs and ensure that all resources are initialized: 1. dev_pm_opp_get_opp_table() ensures the OPP table is allocated and initialized with clock and interconnects. 2. dev_pm_opp_set_regulators() requests the regulators and assigns them to the OPP table. 3. We call dev_pm_opp_of_get_sharing_cpus() early so that we only initialize the OPP table once for each shared policy. With these changes, we actually end up saving a few lines of code, the resources are no longer looked up multiple times and everything should be much more robust. Signed-off-by: Stephan Gerhold <stephan@gerhold.net> [ Viresh: Use list_head structure for maintaining the list and minor changes ] Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
2020-07-27 16:30:47 +07:00
cpu_dev = priv->cpu_dev;
cpumask_copy(policy->cpus, priv->cpus);
cpu_clk = clk_get(cpu_dev, NULL);
if (IS_ERR(cpu_clk)) {
ret = PTR_ERR(cpu_clk);
dev_err(cpu_dev, "%s: failed to get clk: %d\n", __func__, ret);
return ret;
}
/*
* Initialize OPP tables for all policy->cpus. They will be shared by
* all CPUs which have marked their CPUs shared with OPP bindings.
*
* For platforms not using operating-points-v2 bindings, we do this
* before updating policy->cpus. Otherwise, we will end up creating
* duplicate OPPs for policy->cpus.
*
* OPPs might be populated at runtime, don't check for error here
*/
if (!dev_pm_opp_of_cpumask_add_table(policy->cpus))
priv->have_static_opps = true;
/*
* But we need OPP table to function so if it is not there let's
* give platform code chance to provide it for us.
*/
ret = dev_pm_opp_get_opp_count(cpu_dev);
if (ret <= 0) {
cpufreq: dt: Refactor initialization to handle probe deferral properly cpufreq-dt is currently unable to handle -EPROBE_DEFER properly because the error code is not propagated for the cpufreq_driver->init() callback. Instead, it attempts to avoid the situation by temporarily requesting all resources within resources_available() and releasing them again immediately after. This has several disadvantages: - Whenever we add something like interconnect handling to the OPP core we need to patch cpufreq-dt to request these resources early. - resources_available() is only run for CPU0, but other clusters may eventually depend on other resources that are not available yet. (See FIXME comment removed by this commit...) - All resources need to be looked up several times. Now that the OPP core can propagate -EPROBE_DEFER during initialization, it would be nice to avoid all that trouble and just propagate its error code when necessary. This commit refactors the cpufreq-dt driver to initialize private_data before registering the cpufreq driver. We do this by iterating over all possible CPUs and ensure that all resources are initialized: 1. dev_pm_opp_get_opp_table() ensures the OPP table is allocated and initialized with clock and interconnects. 2. dev_pm_opp_set_regulators() requests the regulators and assigns them to the OPP table. 3. We call dev_pm_opp_of_get_sharing_cpus() early so that we only initialize the OPP table once for each shared policy. With these changes, we actually end up saving a few lines of code, the resources are no longer looked up multiple times and everything should be much more robust. Signed-off-by: Stephan Gerhold <stephan@gerhold.net> [ Viresh: Use list_head structure for maintaining the list and minor changes ] Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
2020-07-27 16:30:47 +07:00
dev_err(cpu_dev, "OPP table can't be empty\n");
ret = -ENODEV;
goto out_free_opp;
}
ret = dev_pm_opp_init_cpufreq_table(cpu_dev, &freq_table);
if (ret) {
dev_err(cpu_dev, "failed to init cpufreq table: %d\n", ret);
goto out_free_opp;
}
policy->driver_data = priv;
policy->clk = cpu_clk;
policy->freq_table = freq_table;
policy->suspend_freq = dev_pm_opp_get_suspend_opp_freq(cpu_dev) / 1000;
/* Support turbo/boost mode */
if (policy_has_boost_freq(policy)) {
/* This gets disabled by core on driver unregister */
ret = cpufreq_enable_boost_support();
if (ret)
goto out_free_cpufreq_table;
cpufreq_dt_attr[1] = &cpufreq_freq_attr_scaling_boost_freqs;
}
transition_latency = dev_pm_opp_get_max_transition_latency(cpu_dev);
if (!transition_latency)
transition_latency = CPUFREQ_ETERNAL;
policy->cpuinfo.transition_latency = transition_latency;
policy->dvfs_possible_from_any_cpu = true;
dev_pm_opp_of_register_em(cpu_dev, policy->cpus);
return 0;
out_free_cpufreq_table:
dev_pm_opp_free_cpufreq_table(cpu_dev, &freq_table);
out_free_opp:
if (priv->have_static_opps)
dev_pm_opp_of_cpumask_remove_table(policy->cpus);
clk_put(cpu_clk);
return ret;
}
static int cpufreq_online(struct cpufreq_policy *policy)
{
/* We did light-weight tear down earlier, nothing to do here */
return 0;
}
static int cpufreq_offline(struct cpufreq_policy *policy)
{
/*
* Preserve policy->driver_data and don't free resources on light-weight
* tear down.
*/
return 0;
}
static int cpufreq_exit(struct cpufreq_policy *policy)
{
struct private_data *priv = policy->driver_data;
dev_pm_opp_free_cpufreq_table(priv->cpu_dev, &policy->freq_table);
if (priv->have_static_opps)
dev_pm_opp_of_cpumask_remove_table(policy->related_cpus);
clk_put(policy->clk);
return 0;
}
static struct cpufreq_driver dt_cpufreq_driver = {
.flags = CPUFREQ_STICKY | CPUFREQ_NEED_INITIAL_FREQ_CHECK |
CPUFREQ_IS_COOLING_DEV,
.verify = cpufreq_generic_frequency_table_verify,
.target_index = set_target,
.get = cpufreq_generic_get,
.init = cpufreq_init,
.exit = cpufreq_exit,
.online = cpufreq_online,
.offline = cpufreq_offline,
.name = "cpufreq-dt",
.attr = cpufreq_dt_attr,
.suspend = cpufreq_generic_suspend,
};
cpufreq: dt: Refactor initialization to handle probe deferral properly cpufreq-dt is currently unable to handle -EPROBE_DEFER properly because the error code is not propagated for the cpufreq_driver->init() callback. Instead, it attempts to avoid the situation by temporarily requesting all resources within resources_available() and releasing them again immediately after. This has several disadvantages: - Whenever we add something like interconnect handling to the OPP core we need to patch cpufreq-dt to request these resources early. - resources_available() is only run for CPU0, but other clusters may eventually depend on other resources that are not available yet. (See FIXME comment removed by this commit...) - All resources need to be looked up several times. Now that the OPP core can propagate -EPROBE_DEFER during initialization, it would be nice to avoid all that trouble and just propagate its error code when necessary. This commit refactors the cpufreq-dt driver to initialize private_data before registering the cpufreq driver. We do this by iterating over all possible CPUs and ensure that all resources are initialized: 1. dev_pm_opp_get_opp_table() ensures the OPP table is allocated and initialized with clock and interconnects. 2. dev_pm_opp_set_regulators() requests the regulators and assigns them to the OPP table. 3. We call dev_pm_opp_of_get_sharing_cpus() early so that we only initialize the OPP table once for each shared policy. With these changes, we actually end up saving a few lines of code, the resources are no longer looked up multiple times and everything should be much more robust. Signed-off-by: Stephan Gerhold <stephan@gerhold.net> [ Viresh: Use list_head structure for maintaining the list and minor changes ] Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
2020-07-27 16:30:47 +07:00
static int dt_cpufreq_early_init(struct device *dev, int cpu)
{
cpufreq: dt: Refactor initialization to handle probe deferral properly cpufreq-dt is currently unable to handle -EPROBE_DEFER properly because the error code is not propagated for the cpufreq_driver->init() callback. Instead, it attempts to avoid the situation by temporarily requesting all resources within resources_available() and releasing them again immediately after. This has several disadvantages: - Whenever we add something like interconnect handling to the OPP core we need to patch cpufreq-dt to request these resources early. - resources_available() is only run for CPU0, but other clusters may eventually depend on other resources that are not available yet. (See FIXME comment removed by this commit...) - All resources need to be looked up several times. Now that the OPP core can propagate -EPROBE_DEFER during initialization, it would be nice to avoid all that trouble and just propagate its error code when necessary. This commit refactors the cpufreq-dt driver to initialize private_data before registering the cpufreq driver. We do this by iterating over all possible CPUs and ensure that all resources are initialized: 1. dev_pm_opp_get_opp_table() ensures the OPP table is allocated and initialized with clock and interconnects. 2. dev_pm_opp_set_regulators() requests the regulators and assigns them to the OPP table. 3. We call dev_pm_opp_of_get_sharing_cpus() early so that we only initialize the OPP table once for each shared policy. With these changes, we actually end up saving a few lines of code, the resources are no longer looked up multiple times and everything should be much more robust. Signed-off-by: Stephan Gerhold <stephan@gerhold.net> [ Viresh: Use list_head structure for maintaining the list and minor changes ] Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
2020-07-27 16:30:47 +07:00
struct private_data *priv;
struct device *cpu_dev;
const char *reg_name;
int ret;
cpufreq: dt: Refactor initialization to handle probe deferral properly cpufreq-dt is currently unable to handle -EPROBE_DEFER properly because the error code is not propagated for the cpufreq_driver->init() callback. Instead, it attempts to avoid the situation by temporarily requesting all resources within resources_available() and releasing them again immediately after. This has several disadvantages: - Whenever we add something like interconnect handling to the OPP core we need to patch cpufreq-dt to request these resources early. - resources_available() is only run for CPU0, but other clusters may eventually depend on other resources that are not available yet. (See FIXME comment removed by this commit...) - All resources need to be looked up several times. Now that the OPP core can propagate -EPROBE_DEFER during initialization, it would be nice to avoid all that trouble and just propagate its error code when necessary. This commit refactors the cpufreq-dt driver to initialize private_data before registering the cpufreq driver. We do this by iterating over all possible CPUs and ensure that all resources are initialized: 1. dev_pm_opp_get_opp_table() ensures the OPP table is allocated and initialized with clock and interconnects. 2. dev_pm_opp_set_regulators() requests the regulators and assigns them to the OPP table. 3. We call dev_pm_opp_of_get_sharing_cpus() early so that we only initialize the OPP table once for each shared policy. With these changes, we actually end up saving a few lines of code, the resources are no longer looked up multiple times and everything should be much more robust. Signed-off-by: Stephan Gerhold <stephan@gerhold.net> [ Viresh: Use list_head structure for maintaining the list and minor changes ] Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
2020-07-27 16:30:47 +07:00
/* Check if this CPU is already covered by some other policy */
if (cpufreq_dt_find_data(cpu))
return 0;
cpu_dev = get_cpu_device(cpu);
if (!cpu_dev)
return -EPROBE_DEFER;
priv = devm_kzalloc(dev, sizeof(*priv), GFP_KERNEL);
if (!priv)
return -ENOMEM;
if (!alloc_cpumask_var(&priv->cpus, GFP_KERNEL))
return -ENOMEM;
priv->cpu_dev = cpu_dev;
/* Try to get OPP table early to ensure resources are available */
priv->opp_table = dev_pm_opp_get_opp_table(cpu_dev);
if (IS_ERR(priv->opp_table)) {
ret = PTR_ERR(priv->opp_table);
if (ret != -EPROBE_DEFER)
dev_err(cpu_dev, "failed to get OPP table: %d\n", ret);
goto free_cpumask;
}
/*
cpufreq: dt: Refactor initialization to handle probe deferral properly cpufreq-dt is currently unable to handle -EPROBE_DEFER properly because the error code is not propagated for the cpufreq_driver->init() callback. Instead, it attempts to avoid the situation by temporarily requesting all resources within resources_available() and releasing them again immediately after. This has several disadvantages: - Whenever we add something like interconnect handling to the OPP core we need to patch cpufreq-dt to request these resources early. - resources_available() is only run for CPU0, but other clusters may eventually depend on other resources that are not available yet. (See FIXME comment removed by this commit...) - All resources need to be looked up several times. Now that the OPP core can propagate -EPROBE_DEFER during initialization, it would be nice to avoid all that trouble and just propagate its error code when necessary. This commit refactors the cpufreq-dt driver to initialize private_data before registering the cpufreq driver. We do this by iterating over all possible CPUs and ensure that all resources are initialized: 1. dev_pm_opp_get_opp_table() ensures the OPP table is allocated and initialized with clock and interconnects. 2. dev_pm_opp_set_regulators() requests the regulators and assigns them to the OPP table. 3. We call dev_pm_opp_of_get_sharing_cpus() early so that we only initialize the OPP table once for each shared policy. With these changes, we actually end up saving a few lines of code, the resources are no longer looked up multiple times and everything should be much more robust. Signed-off-by: Stephan Gerhold <stephan@gerhold.net> [ Viresh: Use list_head structure for maintaining the list and minor changes ] Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
2020-07-27 16:30:47 +07:00
* OPP layer will be taking care of regulators now, but it needs to know
* the name of the regulator first.
*/
cpufreq: dt: Refactor initialization to handle probe deferral properly cpufreq-dt is currently unable to handle -EPROBE_DEFER properly because the error code is not propagated for the cpufreq_driver->init() callback. Instead, it attempts to avoid the situation by temporarily requesting all resources within resources_available() and releasing them again immediately after. This has several disadvantages: - Whenever we add something like interconnect handling to the OPP core we need to patch cpufreq-dt to request these resources early. - resources_available() is only run for CPU0, but other clusters may eventually depend on other resources that are not available yet. (See FIXME comment removed by this commit...) - All resources need to be looked up several times. Now that the OPP core can propagate -EPROBE_DEFER during initialization, it would be nice to avoid all that trouble and just propagate its error code when necessary. This commit refactors the cpufreq-dt driver to initialize private_data before registering the cpufreq driver. We do this by iterating over all possible CPUs and ensure that all resources are initialized: 1. dev_pm_opp_get_opp_table() ensures the OPP table is allocated and initialized with clock and interconnects. 2. dev_pm_opp_set_regulators() requests the regulators and assigns them to the OPP table. 3. We call dev_pm_opp_of_get_sharing_cpus() early so that we only initialize the OPP table once for each shared policy. With these changes, we actually end up saving a few lines of code, the resources are no longer looked up multiple times and everything should be much more robust. Signed-off-by: Stephan Gerhold <stephan@gerhold.net> [ Viresh: Use list_head structure for maintaining the list and minor changes ] Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
2020-07-27 16:30:47 +07:00
reg_name = find_supply_name(cpu_dev);
if (reg_name) {
priv->reg_opp_table = dev_pm_opp_set_regulators(cpu_dev,
&reg_name, 1);
if (IS_ERR(priv->reg_opp_table)) {
ret = PTR_ERR(priv->reg_opp_table);
if (ret != -EPROBE_DEFER)
dev_err(cpu_dev, "failed to set regulators: %d\n",
ret);
goto put_table;
}
}
/* Find OPP sharing information so we can fill pri->cpus here */
/* Get OPP-sharing information from "operating-points-v2" bindings */
ret = dev_pm_opp_of_get_sharing_cpus(cpu_dev, priv->cpus);
if (ret) {
if (ret != -ENOENT)
goto put_reg;
/*
* operating-points-v2 not supported, fallback to all CPUs share
* OPP for backward compatibility if the platform hasn't set
* sharing CPUs.
*/
if (dev_pm_opp_get_sharing_cpus(cpu_dev, priv->cpus)) {
cpumask_setall(priv->cpus);
/*
* OPP tables are initialized only for cpu, do it for
* others as well.
*/
ret = dev_pm_opp_set_sharing_cpus(cpu_dev, priv->cpus);
if (ret)
dev_err(cpu_dev, "%s: failed to mark OPPs as shared: %d\n",
__func__, ret);
}
}
list_add(&priv->node, &priv_list);
return 0;
put_reg:
if (priv->reg_opp_table)
dev_pm_opp_put_regulators(priv->reg_opp_table);
put_table:
dev_pm_opp_put_opp_table(priv->opp_table);
free_cpumask:
free_cpumask_var(priv->cpus);
return ret;
}
static void dt_cpufreq_release(void)
{
struct private_data *priv, *tmp;
list_for_each_entry_safe(priv, tmp, &priv_list, node) {
if (priv->reg_opp_table)
dev_pm_opp_put_regulators(priv->reg_opp_table);
dev_pm_opp_put_opp_table(priv->opp_table);
free_cpumask_var(priv->cpus);
list_del(&priv->node);
}
}
static int dt_cpufreq_probe(struct platform_device *pdev)
{
struct cpufreq_dt_platform_data *data = dev_get_platdata(&pdev->dev);
int ret, cpu;
/* Request resources early so we can return in case of -EPROBE_DEFER */
for_each_possible_cpu(cpu) {
ret = dt_cpufreq_early_init(&pdev->dev, cpu);
if (ret)
goto err;
}
if (data) {
if (data->have_governor_per_policy)
dt_cpufreq_driver.flags |= CPUFREQ_HAVE_GOVERNOR_PER_POLICY;
dt_cpufreq_driver.resume = data->resume;
if (data->suspend)
dt_cpufreq_driver.suspend = data->suspend;
if (data->get_intermediate) {
dt_cpufreq_driver.target_intermediate = data->target_intermediate;
dt_cpufreq_driver.get_intermediate = data->get_intermediate;
}
}
ret = cpufreq_register_driver(&dt_cpufreq_driver);
cpufreq: dt: Refactor initialization to handle probe deferral properly cpufreq-dt is currently unable to handle -EPROBE_DEFER properly because the error code is not propagated for the cpufreq_driver->init() callback. Instead, it attempts to avoid the situation by temporarily requesting all resources within resources_available() and releasing them again immediately after. This has several disadvantages: - Whenever we add something like interconnect handling to the OPP core we need to patch cpufreq-dt to request these resources early. - resources_available() is only run for CPU0, but other clusters may eventually depend on other resources that are not available yet. (See FIXME comment removed by this commit...) - All resources need to be looked up several times. Now that the OPP core can propagate -EPROBE_DEFER during initialization, it would be nice to avoid all that trouble and just propagate its error code when necessary. This commit refactors the cpufreq-dt driver to initialize private_data before registering the cpufreq driver. We do this by iterating over all possible CPUs and ensure that all resources are initialized: 1. dev_pm_opp_get_opp_table() ensures the OPP table is allocated and initialized with clock and interconnects. 2. dev_pm_opp_set_regulators() requests the regulators and assigns them to the OPP table. 3. We call dev_pm_opp_of_get_sharing_cpus() early so that we only initialize the OPP table once for each shared policy. With these changes, we actually end up saving a few lines of code, the resources are no longer looked up multiple times and everything should be much more robust. Signed-off-by: Stephan Gerhold <stephan@gerhold.net> [ Viresh: Use list_head structure for maintaining the list and minor changes ] Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
2020-07-27 16:30:47 +07:00
if (ret) {
dev_err(&pdev->dev, "failed register driver: %d\n", ret);
cpufreq: dt: Refactor initialization to handle probe deferral properly cpufreq-dt is currently unable to handle -EPROBE_DEFER properly because the error code is not propagated for the cpufreq_driver->init() callback. Instead, it attempts to avoid the situation by temporarily requesting all resources within resources_available() and releasing them again immediately after. This has several disadvantages: - Whenever we add something like interconnect handling to the OPP core we need to patch cpufreq-dt to request these resources early. - resources_available() is only run for CPU0, but other clusters may eventually depend on other resources that are not available yet. (See FIXME comment removed by this commit...) - All resources need to be looked up several times. Now that the OPP core can propagate -EPROBE_DEFER during initialization, it would be nice to avoid all that trouble and just propagate its error code when necessary. This commit refactors the cpufreq-dt driver to initialize private_data before registering the cpufreq driver. We do this by iterating over all possible CPUs and ensure that all resources are initialized: 1. dev_pm_opp_get_opp_table() ensures the OPP table is allocated and initialized with clock and interconnects. 2. dev_pm_opp_set_regulators() requests the regulators and assigns them to the OPP table. 3. We call dev_pm_opp_of_get_sharing_cpus() early so that we only initialize the OPP table once for each shared policy. With these changes, we actually end up saving a few lines of code, the resources are no longer looked up multiple times and everything should be much more robust. Signed-off-by: Stephan Gerhold <stephan@gerhold.net> [ Viresh: Use list_head structure for maintaining the list and minor changes ] Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
2020-07-27 16:30:47 +07:00
goto err;
}
cpufreq: dt: Refactor initialization to handle probe deferral properly cpufreq-dt is currently unable to handle -EPROBE_DEFER properly because the error code is not propagated for the cpufreq_driver->init() callback. Instead, it attempts to avoid the situation by temporarily requesting all resources within resources_available() and releasing them again immediately after. This has several disadvantages: - Whenever we add something like interconnect handling to the OPP core we need to patch cpufreq-dt to request these resources early. - resources_available() is only run for CPU0, but other clusters may eventually depend on other resources that are not available yet. (See FIXME comment removed by this commit...) - All resources need to be looked up several times. Now that the OPP core can propagate -EPROBE_DEFER during initialization, it would be nice to avoid all that trouble and just propagate its error code when necessary. This commit refactors the cpufreq-dt driver to initialize private_data before registering the cpufreq driver. We do this by iterating over all possible CPUs and ensure that all resources are initialized: 1. dev_pm_opp_get_opp_table() ensures the OPP table is allocated and initialized with clock and interconnects. 2. dev_pm_opp_set_regulators() requests the regulators and assigns them to the OPP table. 3. We call dev_pm_opp_of_get_sharing_cpus() early so that we only initialize the OPP table once for each shared policy. With these changes, we actually end up saving a few lines of code, the resources are no longer looked up multiple times and everything should be much more robust. Signed-off-by: Stephan Gerhold <stephan@gerhold.net> [ Viresh: Use list_head structure for maintaining the list and minor changes ] Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
2020-07-27 16:30:47 +07:00
return 0;
err:
dt_cpufreq_release();
return ret;
}
static int dt_cpufreq_remove(struct platform_device *pdev)
{
cpufreq_unregister_driver(&dt_cpufreq_driver);
cpufreq: dt: Refactor initialization to handle probe deferral properly cpufreq-dt is currently unable to handle -EPROBE_DEFER properly because the error code is not propagated for the cpufreq_driver->init() callback. Instead, it attempts to avoid the situation by temporarily requesting all resources within resources_available() and releasing them again immediately after. This has several disadvantages: - Whenever we add something like interconnect handling to the OPP core we need to patch cpufreq-dt to request these resources early. - resources_available() is only run for CPU0, but other clusters may eventually depend on other resources that are not available yet. (See FIXME comment removed by this commit...) - All resources need to be looked up several times. Now that the OPP core can propagate -EPROBE_DEFER during initialization, it would be nice to avoid all that trouble and just propagate its error code when necessary. This commit refactors the cpufreq-dt driver to initialize private_data before registering the cpufreq driver. We do this by iterating over all possible CPUs and ensure that all resources are initialized: 1. dev_pm_opp_get_opp_table() ensures the OPP table is allocated and initialized with clock and interconnects. 2. dev_pm_opp_set_regulators() requests the regulators and assigns them to the OPP table. 3. We call dev_pm_opp_of_get_sharing_cpus() early so that we only initialize the OPP table once for each shared policy. With these changes, we actually end up saving a few lines of code, the resources are no longer looked up multiple times and everything should be much more robust. Signed-off-by: Stephan Gerhold <stephan@gerhold.net> [ Viresh: Use list_head structure for maintaining the list and minor changes ] Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
2020-07-27 16:30:47 +07:00
dt_cpufreq_release();
return 0;
}
static struct platform_driver dt_cpufreq_platdrv = {
.driver = {
.name = "cpufreq-dt",
},
.probe = dt_cpufreq_probe,
.remove = dt_cpufreq_remove,
};
module_platform_driver(dt_cpufreq_platdrv);
MODULE_ALIAS("platform:cpufreq-dt");
MODULE_AUTHOR("Viresh Kumar <viresh.kumar@linaro.org>");
MODULE_AUTHOR("Shawn Guo <shawn.guo@linaro.org>");
MODULE_DESCRIPTION("Generic cpufreq driver");
MODULE_LICENSE("GPL");