2019-04-28 17:51:04 +08:00
|
|
|
// SPDX-License-Identifier: GPL-2.0
|
2012-08-16 19:41:40 +08:00
|
|
|
/*
|
|
|
|
* linux/drivers/thermal/cpu_cooling.c
|
|
|
|
*
|
|
|
|
* Copyright (C) 2012 Samsung Electronics Co., Ltd(http://www.samsung.com)
|
|
|
|
*
|
2019-04-28 17:51:03 +08:00
|
|
|
* Copyright (C) 2012-2018 Linaro Limited.
|
|
|
|
*
|
|
|
|
* Authors: Amit Daniel <amit.kachhap@linaro.org>
|
|
|
|
* Viresh Kumar <viresh.kumar@linaro.org>
|
2014-12-04 12:12:08 +08:00
|
|
|
*
|
2012-08-16 19:41:40 +08:00
|
|
|
*/
|
|
|
|
#include <linux/module.h>
|
|
|
|
#include <linux/thermal.h>
|
|
|
|
#include <linux/cpufreq.h>
|
|
|
|
#include <linux/err.h>
|
2016-12-22 01:47:05 +08:00
|
|
|
#include <linux/idr.h>
|
2015-02-27 03:00:29 +08:00
|
|
|
#include <linux/pm_opp.h>
|
2019-07-23 14:14:02 +08:00
|
|
|
#include <linux/pm_qos.h>
|
2012-08-16 19:41:40 +08:00
|
|
|
#include <linux/slab.h>
|
|
|
|
#include <linux/cpu.h>
|
|
|
|
#include <linux/cpu_cooling.h>
|
|
|
|
|
2015-03-03 01:17:20 +08:00
|
|
|
#include <trace/events/thermal.h>
|
|
|
|
|
2014-12-04 12:11:49 +08:00
|
|
|
/*
|
|
|
|
* Cooling state <-> CPUFreq frequency
|
|
|
|
*
|
|
|
|
* Cooling states are translated to frequencies throughout this driver and this
|
|
|
|
* is the relation between them.
|
|
|
|
*
|
|
|
|
* Highest cooling state corresponds to lowest possible frequency.
|
|
|
|
*
|
|
|
|
* i.e.
|
|
|
|
* level 0 --> 1st Max Freq
|
|
|
|
* level 1 --> 2nd Max Freq
|
|
|
|
* ...
|
|
|
|
*/
|
|
|
|
|
2015-02-27 03:00:29 +08:00
|
|
|
/**
|
2017-04-25 18:27:19 +08:00
|
|
|
* struct freq_table - frequency table along with power entries
|
2015-02-27 03:00:29 +08:00
|
|
|
* @frequency: frequency in KHz
|
|
|
|
* @power: power in mW
|
|
|
|
*
|
|
|
|
* This structure is built when the cooling device registers and helps
|
2017-04-25 18:27:19 +08:00
|
|
|
* in translating frequency to power and vice versa.
|
2015-02-27 03:00:29 +08:00
|
|
|
*/
|
2017-04-25 18:27:19 +08:00
|
|
|
struct freq_table {
|
2015-02-27 03:00:29 +08:00
|
|
|
u32 frequency;
|
|
|
|
u32 power;
|
|
|
|
};
|
|
|
|
|
2017-04-25 18:27:20 +08:00
|
|
|
/**
|
|
|
|
* struct time_in_idle - Idle time stats
|
|
|
|
* @time: previous reading of the absolute time that this cpu was idle
|
|
|
|
* @timestamp: wall time of the last invocation of get_cpu_idle_time_us()
|
|
|
|
*/
|
|
|
|
struct time_in_idle {
|
|
|
|
u64 time;
|
|
|
|
u64 timestamp;
|
|
|
|
};
|
|
|
|
|
2012-08-16 19:41:40 +08:00
|
|
|
/**
|
2013-04-18 01:11:56 +08:00
|
|
|
* struct cpufreq_cooling_device - data for cooling device with cpufreq
|
2012-08-16 19:41:40 +08:00
|
|
|
* @id: unique integer value corresponding to each cpufreq_cooling_device
|
|
|
|
* registered.
|
2017-04-25 18:27:24 +08:00
|
|
|
* @last_load: load measured by the latest call to cpufreq_get_requested_power()
|
2012-08-16 19:41:40 +08:00
|
|
|
* @cpufreq_state: integer value representing the current state of cpufreq
|
|
|
|
* cooling devices.
|
2014-12-04 12:12:02 +08:00
|
|
|
* @max_level: maximum cooling level. One less than total number of valid
|
|
|
|
* cpufreq frequencies.
|
2017-04-25 18:27:24 +08:00
|
|
|
* @freq_table: Freq table in descending order of frequencies
|
|
|
|
* @cdev: thermal_cooling_device pointer to keep track of the
|
|
|
|
* registered cooling device.
|
|
|
|
* @policy: cpufreq policy.
|
2014-12-16 00:55:52 +08:00
|
|
|
* @node: list_head to link all cpufreq_cooling_device together.
|
2017-04-25 18:27:20 +08:00
|
|
|
* @idle_time: idle time stats
|
2012-08-16 19:41:40 +08:00
|
|
|
*
|
2014-12-04 12:11:48 +08:00
|
|
|
* This structure is required for keeping information of each registered
|
|
|
|
* cpufreq_cooling_device.
|
2012-08-16 19:41:40 +08:00
|
|
|
*/
|
|
|
|
struct cpufreq_cooling_device {
|
|
|
|
int id;
|
2017-04-25 18:27:24 +08:00
|
|
|
u32 last_load;
|
2012-08-16 19:41:40 +08:00
|
|
|
unsigned int cpufreq_state;
|
2014-12-04 12:12:02 +08:00
|
|
|
unsigned int max_level;
|
2017-04-25 18:27:19 +08:00
|
|
|
struct freq_table *freq_table; /* In descending order */
|
2017-04-25 18:27:24 +08:00
|
|
|
struct cpufreq_policy *policy;
|
2014-11-07 21:42:29 +08:00
|
|
|
struct list_head node;
|
2017-04-25 18:27:20 +08:00
|
|
|
struct time_in_idle *idle_time;
|
cpufreq: Use per-policy frequency QoS
Replace the CPU device PM QoS used for the management of min and max
frequency constraints in cpufreq (and its users) with per-policy
frequency QoS to avoid problems with cpufreq policies covering
more then one CPU.
Namely, a cpufreq driver is registered with the subsys interface
which calls cpufreq_add_dev() for each CPU, starting from CPU0, so
currently the PM QoS notifiers are added to the first CPU in the
policy (i.e. CPU0 in the majority of cases).
In turn, when the cpufreq driver is unregistered, the subsys interface
doing that calls cpufreq_remove_dev() for each CPU, starting from CPU0,
and the PM QoS notifiers are only removed when cpufreq_remove_dev() is
called for the last CPU in the policy, say CPUx, which as a rule is
not CPU0 if the policy covers more than one CPU. Then, the PM QoS
notifiers cannot be removed, because CPUx does not have them, and
they are still there in the device PM QoS notifiers list of CPU0,
which prevents new PM QoS notifiers from being registered for CPU0
on the next attempt to register the cpufreq driver.
The same issue occurs when the first CPU in the policy goes offline
before unregistering the driver.
After this change it does not matter which CPU is the policy CPU at
the driver registration time and whether or not it is online all the
time, because the frequency QoS is per policy and not per CPU.
Fixes: 67d874c3b2c6 ("cpufreq: Register notifiers with the PM QoS framework")
Reported-by: Dmitry Osipenko <digetx@gmail.com>
Tested-by: Dmitry Osipenko <digetx@gmail.com>
Reported-by: Sudeep Holla <sudeep.holla@arm.com>
Tested-by: Sudeep Holla <sudeep.holla@arm.com>
Diagnosed-by: Viresh Kumar <viresh.kumar@linaro.org>
Link: https://lore.kernel.org/linux-pm/5ad2624194baa2f53acc1f1e627eb7684c577a19.1562210705.git.viresh.kumar@linaro.org/T/#md2d89e95906b8c91c15f582146173dce2e86e99f
Link: https://lore.kernel.org/linux-pm/20191017094612.6tbkwoq4harsjcqv@vireshk-i7/T/#m30d48cc23b9a80467fbaa16e30f90b3828a5a29b
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Acked-by: Viresh Kumar <viresh.kumar@linaro.org>
2019-10-16 18:47:06 +08:00
|
|
|
struct freq_qos_request qos_req;
|
2012-08-16 19:41:40 +08:00
|
|
|
};
|
|
|
|
|
2017-04-25 18:27:09 +08:00
|
|
|
static DEFINE_IDA(cpufreq_ida);
|
thermal: cpu_cooling: fix lockdep problems in cpu_cooling
A recent change to the cpu_cooling code introduced a AB-BA deadlock
scenario between the cpufreq_policy_notifier_list rwsem and the
cooling_cpufreq_lock. This is caused by cooling_cpufreq_lock being held
before the registration/removal of the notifier block (an operation
which takes the rwsem), and the notifier code itself which takes the
locks in the reverse order:
======================================================
[ INFO: possible circular locking dependency detected ]
3.18.0+ #1453 Not tainted
-------------------------------------------------------
rc.local/770 is trying to acquire lock:
(cooling_cpufreq_lock){+.+.+.}, at: [<c04abfc4>] cpufreq_thermal_notifier+0x34/0xfc
but task is already holding lock:
((cpufreq_policy_notifier_list).rwsem){++++.+}, at: [<c0042f04>] __blocking_notifier_call_chain+0x34/0x68
which lock already depends on the new lock.
the existing dependency chain (in reverse order) is:
-> #1 ((cpufreq_policy_notifier_list).rwsem){++++.+}:
[<c06bc3b0>] down_write+0x44/0x9c
[<c0043444>] blocking_notifier_chain_register+0x28/0xd8
[<c04ad610>] cpufreq_register_notifier+0x68/0x90
[<c04abe4c>] __cpufreq_cooling_register.part.1+0x120/0x180
[<c04abf44>] __cpufreq_cooling_register+0x98/0xa4
[<c04abf8c>] cpufreq_cooling_register+0x18/0x1c
[<bf0046f8>] imx_thermal_probe+0x1c0/0x470 [imx_thermal]
[<c037cef8>] platform_drv_probe+0x50/0xac
[<c037b710>] driver_probe_device+0x114/0x234
[<c037b8cc>] __driver_attach+0x9c/0xa0
[<c0379d68>] bus_for_each_dev+0x5c/0x90
[<c037b204>] driver_attach+0x24/0x28
[<c037ae7c>] bus_add_driver+0xe0/0x1d8
[<c037c0cc>] driver_register+0x80/0xfc
[<c037cd80>] __platform_driver_register+0x50/0x64
[<bf007018>] 0xbf007018
[<c0008a5c>] do_one_initcall+0x88/0x1d8
[<c0095da4>] load_module+0x1768/0x1ef8
[<c0096614>] SyS_init_module+0xe0/0xf4
[<c000ec00>] ret_fast_syscall+0x0/0x48
-> #0 (cooling_cpufreq_lock){+.+.+.}:
[<c00619f8>] lock_acquire+0xb0/0x124
[<c06ba3b4>] mutex_lock_nested+0x5c/0x3d8
[<c04abfc4>] cpufreq_thermal_notifier+0x34/0xfc
[<c0042bf4>] notifier_call_chain+0x4c/0x8c
[<c0042f20>] __blocking_notifier_call_chain+0x50/0x68
[<c0042f58>] blocking_notifier_call_chain+0x20/0x28
[<c04ae62c>] cpufreq_set_policy+0x7c/0x1d0
[<c04af3cc>] store_scaling_governor+0x74/0x9c
[<c04ad418>] store+0x90/0xc0
[<c0175384>] sysfs_kf_write+0x54/0x58
[<c01746b4>] kernfs_fop_write+0xdc/0x190
[<c010dcc0>] vfs_write+0xac/0x1b4
[<c010dfec>] SyS_write+0x44/0x90
[<c000ec00>] ret_fast_syscall+0x0/0x48
other info that might help us debug this:
Possible unsafe locking scenario:
CPU0 CPU1
---- ----
lock((cpufreq_policy_notifier_list).rwsem);
lock(cooling_cpufreq_lock);
lock((cpufreq_policy_notifier_list).rwsem);
lock(cooling_cpufreq_lock);
*** DEADLOCK ***
7 locks held by rc.local/770:
#0: (sb_writers#6){.+.+.+}, at: [<c010dda0>] vfs_write+0x18c/0x1b4
#1: (&of->mutex){+.+.+.}, at: [<c0174678>] kernfs_fop_write+0xa0/0x190
#2: (s_active#52){.+.+.+}, at: [<c0174680>] kernfs_fop_write+0xa8/0x190
#3: (cpu_hotplug.lock){++++++}, at: [<c0026a60>] get_online_cpus+0x34/0x90
#4: (cpufreq_rwsem){.+.+.+}, at: [<c04ad3e0>] store+0x58/0xc0
#5: (&policy->rwsem){+.+.+.}, at: [<c04ad3f8>] store+0x70/0xc0
#6: ((cpufreq_policy_notifier_list).rwsem){++++.+}, at: [<c0042f04>] __blocking_notifier_call_chain+0x34/0x68
stack backtrace:
CPU: 0 PID: 770 Comm: rc.local Not tainted 3.18.0+ #1453
Hardware name: Freescale i.MX6 Quad/DualLite (Device Tree)
Backtrace:
[<c00121c8>] (dump_backtrace) from [<c0012360>] (show_stack+0x18/0x1c)
r6:c0b85a80 r5:c0b75630 r4:00000000 r3:00000000
[<c0012348>] (show_stack) from [<c06b6c48>] (dump_stack+0x7c/0x98)
[<c06b6bcc>] (dump_stack) from [<c06b42a4>] (print_circular_bug+0x28c/0x2d8)
r4:c0b85a80 r3:d0071d40
[<c06b4018>] (print_circular_bug) from [<c00613b0>] (__lock_acquire+0x1acc/0x1bb0)
r10:c0b50660 r8:c09e6d80 r7:d0071d40 r6:c11d0f0c r5:00000007 r4:d0072240
[<c005f8e4>] (__lock_acquire) from [<c00619f8>] (lock_acquire+0xb0/0x124)
r10:00000000 r9:c04abfc4 r8:00000000 r7:00000000 r6:00000000 r5:c0a06f0c
r4:00000000
[<c0061948>] (lock_acquire) from [<c06ba3b4>] (mutex_lock_nested+0x5c/0x3d8)
r10:ec853800 r9:c0a06ed4 r8:d0071d40 r7:c0a06ed4 r6:c11d0f0c r5:00000000
r4:c04abfc4
[<c06ba358>] (mutex_lock_nested) from [<c04abfc4>] (cpufreq_thermal_notifier+0x34/0xfc)
r10:ec853800 r9:ec85380c r8:d00d7d3c r7:c0a06ed4 r6:d00d7d3c r5:00000000
r4:fffffffe
[<c04abf90>] (cpufreq_thermal_notifier) from [<c0042bf4>] (notifier_call_chain+0x4c/0x8c)
r7:00000000 r6:00000000 r5:00000000 r4:fffffffe
[<c0042ba8>] (notifier_call_chain) from [<c0042f20>] (__blocking_notifier_call_chain+0x50/0x68)
r8:c0a072a4 r7:00000000 r6:d00d7d3c r5:ffffffff r4:c0a06fc8 r3:ffffffff
[<c0042ed0>] (__blocking_notifier_call_chain) from [<c0042f58>] (blocking_notifier_call_chain+0x20/0x28)
r7:ec98b540 r6:c13ebc80 r5:ed76e600 r4:d00d7d3c
[<c0042f38>] (blocking_notifier_call_chain) from [<c04ae62c>] (cpufreq_set_policy+0x7c/0x1d0)
[<c04ae5b0>] (cpufreq_set_policy) from [<c04af3cc>] (store_scaling_governor+0x74/0x9c)
r7:ec98b540 r6:0000000c r5:ec98b540 r4:ed76e600
[<c04af358>] (store_scaling_governor) from [<c04ad418>] (store+0x90/0xc0)
r6:0000000c r5:ed76e6d4 r4:ed76e600
[<c04ad388>] (store) from [<c0175384>] (sysfs_kf_write+0x54/0x58)
r8:0000000c r7:d00d7f78 r6:ec98b540 r5:0000000c r4:ec853800 r3:0000000c
[<c0175330>] (sysfs_kf_write) from [<c01746b4>] (kernfs_fop_write+0xdc/0x190)
r6:ec98b540 r5:00000000 r4:00000000 r3:c0175330
[<c01745d8>] (kernfs_fop_write) from [<c010dcc0>] (vfs_write+0xac/0x1b4)
r10:0162aa70 r9:d00d6000 r8:0000000c r7:d00d7f78 r6:0162aa70 r5:0000000c
r4:eccde500
[<c010dc14>] (vfs_write) from [<c010dfec>] (SyS_write+0x44/0x90)
r10:0162aa70 r8:0000000c r7:eccde500 r6:eccde500 r5:00000000 r4:00000000
[<c010dfa8>] (SyS_write) from [<c000ec00>] (ret_fast_syscall+0x0/0x48)
r10:00000000 r8:c000edc4 r7:00000004 r6:000216cc r5:0000000c r4:0162aa70
Solve this by moving to finer grained locking - use one mutex to protect
the cpufreq_dev_list as a whole, and a separate lock to ensure correct
ordering of cpufreq notifier registration and removal.
cooling_list_lock is taken within cooling_cpufreq_lock on
(un)registration to preserve the behavior of the code, i.e. to
atomically add/remove to the list and (un)register the notifier.
Fixes: 2dcd851fe4b4 ("thermal: cpu_cooling: Update always cpufreq policy with
Reviewed-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Eduardo Valentin <edubezval@gmail.com>
2015-08-12 17:52:16 +08:00
|
|
|
static DEFINE_MUTEX(cooling_list_lock);
|
2017-04-25 18:27:10 +08:00
|
|
|
static LIST_HEAD(cpufreq_cdev_list);
|
2012-08-16 19:41:40 +08:00
|
|
|
|
|
|
|
/* Below code defines functions to be used for cpufreq as cooling device */
|
|
|
|
|
|
|
|
/**
|
2014-12-04 12:12:07 +08:00
|
|
|
* get_level: Find the level for a particular frequency
|
2017-04-25 18:27:10 +08:00
|
|
|
* @cpufreq_cdev: cpufreq_cdev for which the property is required
|
2014-12-04 12:12:07 +08:00
|
|
|
* @freq: Frequency
|
2013-04-18 01:12:00 +08:00
|
|
|
*
|
2017-04-25 18:27:21 +08:00
|
|
|
* Return: level corresponding to the frequency.
|
2012-08-16 19:41:40 +08:00
|
|
|
*/
|
2017-04-25 18:27:10 +08:00
|
|
|
static unsigned long get_level(struct cpufreq_cooling_device *cpufreq_cdev,
|
2014-12-04 12:12:07 +08:00
|
|
|
unsigned int freq)
|
2012-08-16 19:41:40 +08:00
|
|
|
{
|
2017-04-25 18:27:21 +08:00
|
|
|
struct freq_table *freq_table = cpufreq_cdev->freq_table;
|
2014-12-04 12:12:07 +08:00
|
|
|
unsigned long level;
|
2014-01-02 11:57:48 +08:00
|
|
|
|
2017-04-25 18:27:21 +08:00
|
|
|
for (level = 1; level <= cpufreq_cdev->max_level; level++)
|
|
|
|
if (freq > freq_table[level].frequency)
|
2014-12-04 12:12:07 +08:00
|
|
|
break;
|
2012-08-16 19:41:40 +08:00
|
|
|
|
2017-04-25 18:27:21 +08:00
|
|
|
return level - 1;
|
2013-02-08 13:09:32 +08:00
|
|
|
}
|
|
|
|
|
2015-02-27 03:00:29 +08:00
|
|
|
/**
|
2017-04-25 18:27:19 +08:00
|
|
|
* update_freq_table() - Update the freq table with power numbers
|
|
|
|
* @cpufreq_cdev: the cpufreq cooling device in which to update the table
|
2015-02-27 03:00:29 +08:00
|
|
|
* @capacitance: dynamic power coefficient for these cpus
|
|
|
|
*
|
2017-04-25 18:27:19 +08:00
|
|
|
* Update the freq table with power numbers. This table will be used in
|
|
|
|
* cpu_power_to_freq() and cpu_freq_to_power() to convert between power and
|
|
|
|
* frequency efficiently. Power is stored in mW, frequency in KHz. The
|
|
|
|
* resulting table is in descending order.
|
2015-02-27 03:00:29 +08:00
|
|
|
*
|
2015-08-18 02:21:42 +08:00
|
|
|
* Return: 0 on success, -EINVAL if there are no OPPs for any CPUs,
|
2017-04-25 18:27:19 +08:00
|
|
|
* or -ENOMEM if we run out of memory.
|
2015-02-27 03:00:29 +08:00
|
|
|
*/
|
2017-04-25 18:27:19 +08:00
|
|
|
static int update_freq_table(struct cpufreq_cooling_device *cpufreq_cdev,
|
|
|
|
u32 capacitance)
|
2015-02-27 03:00:29 +08:00
|
|
|
{
|
2017-04-25 18:27:19 +08:00
|
|
|
struct freq_table *freq_table = cpufreq_cdev->freq_table;
|
2015-02-27 03:00:29 +08:00
|
|
|
struct dev_pm_opp *opp;
|
|
|
|
struct device *dev = NULL;
|
2017-04-25 18:27:19 +08:00
|
|
|
int num_opps = 0, cpu = cpufreq_cdev->policy->cpu, i;
|
2015-02-27 03:00:29 +08:00
|
|
|
|
2017-04-25 18:27:17 +08:00
|
|
|
dev = get_cpu_device(cpu);
|
|
|
|
if (unlikely(!dev)) {
|
2019-04-28 17:51:05 +08:00
|
|
|
pr_warn("No cpu device for cpu %d\n", cpu);
|
2017-04-25 18:27:17 +08:00
|
|
|
return -ENODEV;
|
2015-02-27 03:00:29 +08:00
|
|
|
}
|
2012-08-16 19:41:40 +08:00
|
|
|
|
2017-04-25 18:27:17 +08:00
|
|
|
num_opps = dev_pm_opp_get_opp_count(dev);
|
|
|
|
if (num_opps < 0)
|
|
|
|
return num_opps;
|
|
|
|
|
2017-04-25 18:27:19 +08:00
|
|
|
/*
|
|
|
|
* The cpufreq table is also built from the OPP table and so the count
|
|
|
|
* should match.
|
|
|
|
*/
|
|
|
|
if (num_opps != cpufreq_cdev->max_level + 1) {
|
|
|
|
dev_warn(dev, "Number of OPPs not matching with max_levels\n");
|
2015-08-18 02:21:42 +08:00
|
|
|
return -EINVAL;
|
2017-04-25 18:27:19 +08:00
|
|
|
}
|
2012-08-16 19:41:40 +08:00
|
|
|
|
2017-04-25 18:27:19 +08:00
|
|
|
for (i = 0; i <= cpufreq_cdev->max_level; i++) {
|
|
|
|
unsigned long freq = freq_table[i].frequency * 1000;
|
|
|
|
u32 freq_mhz = freq_table[i].frequency / 1000;
|
2015-02-27 03:00:29 +08:00
|
|
|
u64 power;
|
2017-04-25 18:27:19 +08:00
|
|
|
u32 voltage_mv;
|
2015-02-27 03:00:29 +08:00
|
|
|
|
2017-04-25 18:27:19 +08:00
|
|
|
/*
|
|
|
|
* Find ceil frequency as 'freq' may be slightly lower than OPP
|
|
|
|
* freq due to truncation while converting to kHz.
|
|
|
|
*/
|
|
|
|
opp = dev_pm_opp_find_freq_ceil(dev, &freq);
|
|
|
|
if (IS_ERR(opp)) {
|
|
|
|
dev_err(dev, "failed to get opp for %lu frequency\n",
|
|
|
|
freq);
|
|
|
|
return -EINVAL;
|
2015-08-18 02:21:42 +08:00
|
|
|
}
|
|
|
|
|
2015-02-27 03:00:29 +08:00
|
|
|
voltage_mv = dev_pm_opp_get_voltage(opp) / 1000;
|
2017-01-23 12:41:47 +08:00
|
|
|
dev_pm_opp_put(opp);
|
2015-02-27 03:00:29 +08:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Do the multiplication with MHz and millivolt so as
|
|
|
|
* to not overflow.
|
|
|
|
*/
|
|
|
|
power = (u64)capacitance * freq_mhz * voltage_mv * voltage_mv;
|
|
|
|
do_div(power, 1000000000);
|
|
|
|
|
|
|
|
/* power is stored in mW */
|
2017-04-25 18:27:19 +08:00
|
|
|
freq_table[i].power = power;
|
2015-08-18 02:21:43 +08:00
|
|
|
}
|
2015-02-27 03:00:29 +08:00
|
|
|
|
2015-08-18 02:21:42 +08:00
|
|
|
return 0;
|
2015-02-27 03:00:29 +08:00
|
|
|
}
|
|
|
|
|
2017-04-25 18:27:10 +08:00
|
|
|
static u32 cpu_freq_to_power(struct cpufreq_cooling_device *cpufreq_cdev,
|
2015-02-27 03:00:29 +08:00
|
|
|
u32 freq)
|
|
|
|
{
|
|
|
|
int i;
|
2017-04-25 18:27:19 +08:00
|
|
|
struct freq_table *freq_table = cpufreq_cdev->freq_table;
|
2015-02-27 03:00:29 +08:00
|
|
|
|
2017-04-25 18:27:19 +08:00
|
|
|
for (i = 1; i <= cpufreq_cdev->max_level; i++)
|
|
|
|
if (freq > freq_table[i].frequency)
|
2015-02-27 03:00:29 +08:00
|
|
|
break;
|
|
|
|
|
2017-04-25 18:27:19 +08:00
|
|
|
return freq_table[i - 1].power;
|
2015-02-27 03:00:29 +08:00
|
|
|
}
|
|
|
|
|
2017-04-25 18:27:10 +08:00
|
|
|
static u32 cpu_power_to_freq(struct cpufreq_cooling_device *cpufreq_cdev,
|
2015-02-27 03:00:29 +08:00
|
|
|
u32 power)
|
|
|
|
{
|
|
|
|
int i;
|
2017-04-25 18:27:19 +08:00
|
|
|
struct freq_table *freq_table = cpufreq_cdev->freq_table;
|
2015-02-27 03:00:29 +08:00
|
|
|
|
2017-04-25 18:27:19 +08:00
|
|
|
for (i = 1; i <= cpufreq_cdev->max_level; i++)
|
|
|
|
if (power > freq_table[i].power)
|
2015-02-27 03:00:29 +08:00
|
|
|
break;
|
|
|
|
|
2017-04-25 18:27:19 +08:00
|
|
|
return freq_table[i - 1].frequency;
|
2015-02-27 03:00:29 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
|
|
|
* get_load() - get load for a cpu since last updated
|
2017-04-25 18:27:10 +08:00
|
|
|
* @cpufreq_cdev: &struct cpufreq_cooling_device for this cpu
|
2015-02-27 03:00:29 +08:00
|
|
|
* @cpu: cpu number
|
2017-04-25 18:27:18 +08:00
|
|
|
* @cpu_idx: index of the cpu in time_in_idle*
|
2015-02-27 03:00:29 +08:00
|
|
|
*
|
|
|
|
* Return: The average load of cpu @cpu in percentage since this
|
|
|
|
* function was last called.
|
|
|
|
*/
|
2017-04-25 18:27:10 +08:00
|
|
|
static u32 get_load(struct cpufreq_cooling_device *cpufreq_cdev, int cpu,
|
2016-02-11 20:00:51 +08:00
|
|
|
int cpu_idx)
|
2015-02-27 03:00:29 +08:00
|
|
|
{
|
|
|
|
u32 load;
|
|
|
|
u64 now, now_idle, delta_time, delta_idle;
|
2017-04-25 18:27:20 +08:00
|
|
|
struct time_in_idle *idle_time = &cpufreq_cdev->idle_time[cpu_idx];
|
2015-02-27 03:00:29 +08:00
|
|
|
|
|
|
|
now_idle = get_cpu_idle_time(cpu, &now, 0);
|
2017-04-25 18:27:20 +08:00
|
|
|
delta_idle = now_idle - idle_time->time;
|
|
|
|
delta_time = now - idle_time->timestamp;
|
2015-02-27 03:00:29 +08:00
|
|
|
|
|
|
|
if (delta_time <= delta_idle)
|
|
|
|
load = 0;
|
|
|
|
else
|
|
|
|
load = div64_u64(100 * (delta_time - delta_idle), delta_time);
|
|
|
|
|
2017-04-25 18:27:20 +08:00
|
|
|
idle_time->time = now_idle;
|
|
|
|
idle_time->timestamp = now;
|
2015-02-27 03:00:29 +08:00
|
|
|
|
|
|
|
return load;
|
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
|
|
|
* get_dynamic_power() - calculate the dynamic power
|
2017-04-25 18:27:10 +08:00
|
|
|
* @cpufreq_cdev: &cpufreq_cooling_device for this cdev
|
2015-02-27 03:00:29 +08:00
|
|
|
* @freq: current frequency
|
|
|
|
*
|
|
|
|
* Return: the dynamic power consumed by the cpus described by
|
2017-04-25 18:27:10 +08:00
|
|
|
* @cpufreq_cdev.
|
2015-02-27 03:00:29 +08:00
|
|
|
*/
|
2017-04-25 18:27:10 +08:00
|
|
|
static u32 get_dynamic_power(struct cpufreq_cooling_device *cpufreq_cdev,
|
2015-02-27 03:00:29 +08:00
|
|
|
unsigned long freq)
|
|
|
|
{
|
|
|
|
u32 raw_cpu_power;
|
|
|
|
|
2017-04-25 18:27:10 +08:00
|
|
|
raw_cpu_power = cpu_freq_to_power(cpufreq_cdev, freq);
|
|
|
|
return (raw_cpu_power * cpufreq_cdev->last_load) / 100;
|
2012-08-16 19:41:40 +08:00
|
|
|
}
|
|
|
|
|
2013-04-18 01:12:02 +08:00
|
|
|
/* cpufreq cooling device callback functions are defined below */
|
2012-08-16 19:41:40 +08:00
|
|
|
|
|
|
|
/**
|
|
|
|
* cpufreq_get_max_state - callback function to get the max cooling state.
|
|
|
|
* @cdev: thermal cooling device pointer.
|
|
|
|
* @state: fill this variable with the max cooling state.
|
2013-04-18 01:12:12 +08:00
|
|
|
*
|
|
|
|
* Callback for the thermal cooling device to return the cpufreq
|
|
|
|
* max cooling state.
|
|
|
|
*
|
|
|
|
* Return: 0 on success, an error code otherwise.
|
2012-08-16 19:41:40 +08:00
|
|
|
*/
|
|
|
|
static int cpufreq_get_max_state(struct thermal_cooling_device *cdev,
|
|
|
|
unsigned long *state)
|
|
|
|
{
|
2017-04-25 18:27:10 +08:00
|
|
|
struct cpufreq_cooling_device *cpufreq_cdev = cdev->devdata;
|
2012-10-31 00:48:58 +08:00
|
|
|
|
2017-04-25 18:27:10 +08:00
|
|
|
*state = cpufreq_cdev->max_level;
|
2014-12-04 12:12:02 +08:00
|
|
|
return 0;
|
2012-08-16 19:41:40 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
|
|
|
* cpufreq_get_cur_state - callback function to get the current cooling state.
|
|
|
|
* @cdev: thermal cooling device pointer.
|
|
|
|
* @state: fill this variable with the current cooling state.
|
2013-04-18 01:12:13 +08:00
|
|
|
*
|
|
|
|
* Callback for the thermal cooling device to return the cpufreq
|
|
|
|
* current cooling state.
|
|
|
|
*
|
|
|
|
* Return: 0 on success, an error code otherwise.
|
2012-08-16 19:41:40 +08:00
|
|
|
*/
|
|
|
|
static int cpufreq_get_cur_state(struct thermal_cooling_device *cdev,
|
|
|
|
unsigned long *state)
|
|
|
|
{
|
2017-04-25 18:27:10 +08:00
|
|
|
struct cpufreq_cooling_device *cpufreq_cdev = cdev->devdata;
|
2012-08-16 19:41:40 +08:00
|
|
|
|
2017-04-25 18:27:10 +08:00
|
|
|
*state = cpufreq_cdev->cpufreq_state;
|
2013-04-18 01:11:59 +08:00
|
|
|
|
2012-10-31 00:48:59 +08:00
|
|
|
return 0;
|
2012-08-16 19:41:40 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
|
|
|
* cpufreq_set_cur_state - callback function to set the current cooling state.
|
|
|
|
* @cdev: thermal cooling device pointer.
|
|
|
|
* @state: set this variable to the current cooling state.
|
2013-04-18 01:12:14 +08:00
|
|
|
*
|
|
|
|
* Callback for the thermal cooling device to change the cpufreq
|
|
|
|
* current cooling state.
|
|
|
|
*
|
|
|
|
* Return: 0 on success, an error code otherwise.
|
2012-08-16 19:41:40 +08:00
|
|
|
*/
|
|
|
|
static int cpufreq_set_cur_state(struct thermal_cooling_device *cdev,
|
|
|
|
unsigned long state)
|
|
|
|
{
|
2017-04-25 18:27:10 +08:00
|
|
|
struct cpufreq_cooling_device *cpufreq_cdev = cdev->devdata;
|
2014-12-04 12:12:07 +08:00
|
|
|
|
|
|
|
/* Request state should be less than max_level */
|
2017-04-25 18:27:10 +08:00
|
|
|
if (WARN_ON(state > cpufreq_cdev->max_level))
|
2014-12-04 12:12:07 +08:00
|
|
|
return -EINVAL;
|
2014-12-04 12:12:00 +08:00
|
|
|
|
|
|
|
/* Check if the old cooling action is same as new cooling action */
|
2017-04-25 18:27:10 +08:00
|
|
|
if (cpufreq_cdev->cpufreq_state == state)
|
2014-12-04 12:12:00 +08:00
|
|
|
return 0;
|
2012-08-16 19:41:40 +08:00
|
|
|
|
2017-04-25 18:27:10 +08:00
|
|
|
cpufreq_cdev->cpufreq_state = state;
|
2014-12-04 12:12:00 +08:00
|
|
|
|
cpufreq: Use per-policy frequency QoS
Replace the CPU device PM QoS used for the management of min and max
frequency constraints in cpufreq (and its users) with per-policy
frequency QoS to avoid problems with cpufreq policies covering
more then one CPU.
Namely, a cpufreq driver is registered with the subsys interface
which calls cpufreq_add_dev() for each CPU, starting from CPU0, so
currently the PM QoS notifiers are added to the first CPU in the
policy (i.e. CPU0 in the majority of cases).
In turn, when the cpufreq driver is unregistered, the subsys interface
doing that calls cpufreq_remove_dev() for each CPU, starting from CPU0,
and the PM QoS notifiers are only removed when cpufreq_remove_dev() is
called for the last CPU in the policy, say CPUx, which as a rule is
not CPU0 if the policy covers more than one CPU. Then, the PM QoS
notifiers cannot be removed, because CPUx does not have them, and
they are still there in the device PM QoS notifiers list of CPU0,
which prevents new PM QoS notifiers from being registered for CPU0
on the next attempt to register the cpufreq driver.
The same issue occurs when the first CPU in the policy goes offline
before unregistering the driver.
After this change it does not matter which CPU is the policy CPU at
the driver registration time and whether or not it is online all the
time, because the frequency QoS is per policy and not per CPU.
Fixes: 67d874c3b2c6 ("cpufreq: Register notifiers with the PM QoS framework")
Reported-by: Dmitry Osipenko <digetx@gmail.com>
Tested-by: Dmitry Osipenko <digetx@gmail.com>
Reported-by: Sudeep Holla <sudeep.holla@arm.com>
Tested-by: Sudeep Holla <sudeep.holla@arm.com>
Diagnosed-by: Viresh Kumar <viresh.kumar@linaro.org>
Link: https://lore.kernel.org/linux-pm/5ad2624194baa2f53acc1f1e627eb7684c577a19.1562210705.git.viresh.kumar@linaro.org/T/#md2d89e95906b8c91c15f582146173dce2e86e99f
Link: https://lore.kernel.org/linux-pm/20191017094612.6tbkwoq4harsjcqv@vireshk-i7/T/#m30d48cc23b9a80467fbaa16e30f90b3828a5a29b
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Acked-by: Viresh Kumar <viresh.kumar@linaro.org>
2019-10-16 18:47:06 +08:00
|
|
|
return freq_qos_update_request(&cpufreq_cdev->qos_req,
|
2019-07-23 14:14:02 +08:00
|
|
|
cpufreq_cdev->freq_table[state].frequency);
|
2012-08-16 19:41:40 +08:00
|
|
|
}
|
|
|
|
|
2015-02-27 03:00:29 +08:00
|
|
|
/**
|
|
|
|
* cpufreq_get_requested_power() - get the current power
|
|
|
|
* @cdev: &thermal_cooling_device pointer
|
|
|
|
* @tz: a valid thermal zone device pointer
|
|
|
|
* @power: pointer in which to store the resulting power
|
|
|
|
*
|
|
|
|
* Calculate the current power consumption of the cpus in milliwatts
|
|
|
|
* and store it in @power. This function should actually calculate
|
|
|
|
* the requested power, but it's hard to get the frequency that
|
|
|
|
* cpufreq would have assigned if there were no thermal limits.
|
|
|
|
* Instead, we calculate the current power on the assumption that the
|
|
|
|
* immediate future will look like the immediate past.
|
|
|
|
*
|
|
|
|
* We use the current frequency and the average load since this
|
|
|
|
* function was last called. In reality, there could have been
|
|
|
|
* multiple opps since this function was last called and that affects
|
|
|
|
* the load calculation. While it's not perfectly accurate, this
|
|
|
|
* simplification is good enough and works. REVISIT this, as more
|
|
|
|
* complex code may be needed if experiments show that it's not
|
|
|
|
* accurate enough.
|
|
|
|
*
|
|
|
|
* Return: 0 on success, -E* if getting the static power failed.
|
|
|
|
*/
|
|
|
|
static int cpufreq_get_requested_power(struct thermal_cooling_device *cdev,
|
|
|
|
struct thermal_zone_device *tz,
|
|
|
|
u32 *power)
|
|
|
|
{
|
|
|
|
unsigned long freq;
|
2017-12-05 13:32:46 +08:00
|
|
|
int i = 0, cpu;
|
|
|
|
u32 total_load = 0;
|
2017-04-25 18:27:10 +08:00
|
|
|
struct cpufreq_cooling_device *cpufreq_cdev = cdev->devdata;
|
2017-04-25 18:27:18 +08:00
|
|
|
struct cpufreq_policy *policy = cpufreq_cdev->policy;
|
2015-03-03 01:17:20 +08:00
|
|
|
u32 *load_cpu = NULL;
|
2015-02-27 03:00:29 +08:00
|
|
|
|
2017-04-25 18:27:18 +08:00
|
|
|
freq = cpufreq_quick_get(policy->cpu);
|
2015-02-27 03:00:29 +08:00
|
|
|
|
2015-03-03 01:17:20 +08:00
|
|
|
if (trace_thermal_power_cpu_get_power_enabled()) {
|
2017-04-25 18:27:18 +08:00
|
|
|
u32 ncpus = cpumask_weight(policy->related_cpus);
|
2015-03-03 01:17:20 +08:00
|
|
|
|
2015-08-19 14:22:19 +08:00
|
|
|
load_cpu = kcalloc(ncpus, sizeof(*load_cpu), GFP_KERNEL);
|
2015-03-03 01:17:20 +08:00
|
|
|
}
|
|
|
|
|
2017-04-25 18:27:18 +08:00
|
|
|
for_each_cpu(cpu, policy->related_cpus) {
|
2015-02-27 03:00:29 +08:00
|
|
|
u32 load;
|
|
|
|
|
|
|
|
if (cpu_online(cpu))
|
2017-04-25 18:27:10 +08:00
|
|
|
load = get_load(cpufreq_cdev, cpu, i);
|
2015-02-27 03:00:29 +08:00
|
|
|
else
|
|
|
|
load = 0;
|
|
|
|
|
|
|
|
total_load += load;
|
thermal: cpu_cooling: Actually trace CPU load in thermal_power_cpu_get_power
The CPU load values passed to the thermal_power_cpu_get_power
tracepoint are zero for all CPUs, unless, unless the
thermal_power_cpu_limit tracepoint is enabled too:
irq/41-rockchip-98 [000] .... 290.972410: thermal_power_cpu_get_power:
cpus=0000000f freq=1800000 load={{0x0,0x0,0x0,0x0}} dynamic_power=4815
vs
irq/41-rockchip-96 [000] .... 95.773585: thermal_power_cpu_get_power:
cpus=0000000f freq=1800000 load={{0x56,0x64,0x64,0x5e}} dynamic_power=4959
irq/41-rockchip-96 [000] .... 95.773596: thermal_power_cpu_limit:
cpus=0000000f freq=408000 cdev_state=10 power=416
There seems to be no good reason for omitting the CPU load information
depending on another tracepoint. My guess is that the intention was to
check whether thermal_power_cpu_get_power is (still) enabled, however
'load_cpu != NULL' already indicates that it was at least enabled when
cpufreq_get_requested_power() was entered, there seems little gain
from omitting the assignment if the tracepoint was just disabled, so
just remove the check.
Fixes: 6828a4711f99 ("thermal: add trace events to the power allocator governor")
Signed-off-by: Matthias Kaehlcke <mka@chromium.org>
Reviewed-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Acked-by: Javi Merino <javi.merino@kernel.org>
Acked-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Eduardo Valentin <edubezval@gmail.com>
2019-05-03 02:32:38 +08:00
|
|
|
if (load_cpu)
|
2015-03-03 01:17:20 +08:00
|
|
|
load_cpu[i] = load;
|
|
|
|
|
|
|
|
i++;
|
2015-02-27 03:00:29 +08:00
|
|
|
}
|
|
|
|
|
2017-04-25 18:27:10 +08:00
|
|
|
cpufreq_cdev->last_load = total_load;
|
2015-02-27 03:00:29 +08:00
|
|
|
|
2017-12-05 13:32:46 +08:00
|
|
|
*power = get_dynamic_power(cpufreq_cdev, freq);
|
2015-03-03 01:17:20 +08:00
|
|
|
|
|
|
|
if (load_cpu) {
|
2017-04-25 18:27:18 +08:00
|
|
|
trace_thermal_power_cpu_get_power(policy->related_cpus, freq,
|
2017-12-05 13:32:46 +08:00
|
|
|
load_cpu, i, *power);
|
2015-03-03 01:17:20 +08:00
|
|
|
|
2015-08-19 14:22:19 +08:00
|
|
|
kfree(load_cpu);
|
2015-03-03 01:17:20 +08:00
|
|
|
}
|
2015-02-27 03:00:29 +08:00
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
|
|
|
* cpufreq_state2power() - convert a cpu cdev state to power consumed
|
|
|
|
* @cdev: &thermal_cooling_device pointer
|
|
|
|
* @tz: a valid thermal zone device pointer
|
|
|
|
* @state: cooling device state to be converted
|
|
|
|
* @power: pointer in which to store the resulting power
|
|
|
|
*
|
|
|
|
* Convert cooling device state @state into power consumption in
|
|
|
|
* milliwatts assuming 100% load. Store the calculated power in
|
|
|
|
* @power.
|
|
|
|
*
|
|
|
|
* Return: 0 on success, -EINVAL if the cooling device state could not
|
|
|
|
* be converted into a frequency or other -E* if there was an error
|
|
|
|
* when calculating the static power.
|
|
|
|
*/
|
|
|
|
static int cpufreq_state2power(struct thermal_cooling_device *cdev,
|
|
|
|
struct thermal_zone_device *tz,
|
|
|
|
unsigned long state, u32 *power)
|
|
|
|
{
|
|
|
|
unsigned int freq, num_cpus;
|
2017-04-25 18:27:10 +08:00
|
|
|
struct cpufreq_cooling_device *cpufreq_cdev = cdev->devdata;
|
2015-02-27 03:00:29 +08:00
|
|
|
|
2017-04-25 18:27:23 +08:00
|
|
|
/* Request state should be less than max_level */
|
|
|
|
if (WARN_ON(state > cpufreq_cdev->max_level))
|
|
|
|
return -EINVAL;
|
|
|
|
|
2017-04-25 18:27:18 +08:00
|
|
|
num_cpus = cpumask_weight(cpufreq_cdev->policy->cpus);
|
2015-02-27 03:00:29 +08:00
|
|
|
|
2017-04-25 18:27:19 +08:00
|
|
|
freq = cpufreq_cdev->freq_table[state].frequency;
|
2017-12-05 13:32:46 +08:00
|
|
|
*power = cpu_freq_to_power(cpufreq_cdev, freq) * num_cpus;
|
2015-02-27 03:00:29 +08:00
|
|
|
|
2017-12-05 13:32:46 +08:00
|
|
|
return 0;
|
2015-02-27 03:00:29 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
|
|
|
* cpufreq_power2state() - convert power to a cooling device state
|
|
|
|
* @cdev: &thermal_cooling_device pointer
|
|
|
|
* @tz: a valid thermal zone device pointer
|
|
|
|
* @power: power in milliwatts to be converted
|
|
|
|
* @state: pointer in which to store the resulting state
|
|
|
|
*
|
|
|
|
* Calculate a cooling device state for the cpus described by @cdev
|
|
|
|
* that would allow them to consume at most @power mW and store it in
|
|
|
|
* @state. Note that this calculation depends on external factors
|
|
|
|
* such as the cpu load or the current static power. Calling this
|
|
|
|
* function with the same power as input can yield different cooling
|
|
|
|
* device states depending on those external factors.
|
|
|
|
*
|
|
|
|
* Return: 0 on success, -ENODEV if no cpus are online or -EINVAL if
|
|
|
|
* the calculated frequency could not be converted to a valid state.
|
|
|
|
* The latter should not happen unless the frequencies available to
|
|
|
|
* cpufreq have changed since the initialization of the cpu cooling
|
|
|
|
* device.
|
|
|
|
*/
|
|
|
|
static int cpufreq_power2state(struct thermal_cooling_device *cdev,
|
|
|
|
struct thermal_zone_device *tz, u32 power,
|
|
|
|
unsigned long *state)
|
|
|
|
{
|
2019-02-18 14:22:30 +08:00
|
|
|
unsigned int target_freq;
|
2017-12-05 13:32:46 +08:00
|
|
|
u32 last_load, normalised_power;
|
2017-04-25 18:27:10 +08:00
|
|
|
struct cpufreq_cooling_device *cpufreq_cdev = cdev->devdata;
|
2017-04-25 18:27:18 +08:00
|
|
|
struct cpufreq_policy *policy = cpufreq_cdev->policy;
|
2015-02-27 03:00:29 +08:00
|
|
|
|
2017-04-25 18:27:10 +08:00
|
|
|
last_load = cpufreq_cdev->last_load ?: 1;
|
2017-12-05 13:32:46 +08:00
|
|
|
normalised_power = (power * 100) / last_load;
|
2017-04-25 18:27:10 +08:00
|
|
|
target_freq = cpu_power_to_freq(cpufreq_cdev, normalised_power);
|
2015-02-27 03:00:29 +08:00
|
|
|
|
2017-04-25 18:27:12 +08:00
|
|
|
*state = get_level(cpufreq_cdev, target_freq);
|
2017-04-25 18:27:18 +08:00
|
|
|
trace_thermal_power_cpu_limit(policy->related_cpus, target_freq, *state,
|
|
|
|
power);
|
2015-02-27 03:00:29 +08:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2012-08-16 19:41:40 +08:00
|
|
|
/* Bind cpufreq callbacks to thermal cooling device ops */
|
2016-08-17 23:14:59 +08:00
|
|
|
|
2015-02-27 03:00:29 +08:00
|
|
|
static struct thermal_cooling_device_ops cpufreq_cooling_ops = {
|
2012-08-16 19:41:40 +08:00
|
|
|
.get_max_state = cpufreq_get_max_state,
|
|
|
|
.get_cur_state = cpufreq_get_cur_state,
|
|
|
|
.set_cur_state = cpufreq_set_cur_state,
|
|
|
|
};
|
|
|
|
|
2016-08-17 23:14:59 +08:00
|
|
|
static struct thermal_cooling_device_ops cpufreq_power_cooling_ops = {
|
|
|
|
.get_max_state = cpufreq_get_max_state,
|
|
|
|
.get_cur_state = cpufreq_get_cur_state,
|
|
|
|
.set_cur_state = cpufreq_set_cur_state,
|
|
|
|
.get_requested_power = cpufreq_get_requested_power,
|
|
|
|
.state2power = cpufreq_state2power,
|
|
|
|
.power2state = cpufreq_power2state,
|
|
|
|
};
|
|
|
|
|
2014-12-04 12:12:06 +08:00
|
|
|
static unsigned int find_next_max(struct cpufreq_frequency_table *table,
|
|
|
|
unsigned int prev_max)
|
|
|
|
{
|
|
|
|
struct cpufreq_frequency_table *pos;
|
|
|
|
unsigned int max = 0;
|
|
|
|
|
|
|
|
cpufreq_for_each_valid_entry(pos, table) {
|
|
|
|
if (pos->frequency > max && pos->frequency < prev_max)
|
|
|
|
max = pos->frequency;
|
|
|
|
}
|
|
|
|
|
|
|
|
return max;
|
|
|
|
}
|
|
|
|
|
2012-08-16 19:41:40 +08:00
|
|
|
/**
|
2013-09-13 07:26:45 +08:00
|
|
|
* __cpufreq_cooling_register - helper function to create cpufreq cooling device
|
|
|
|
* @np: a valid struct device_node to the cooling device device tree node
|
2017-04-25 18:27:14 +08:00
|
|
|
* @policy: cpufreq policy
|
2014-12-04 12:11:55 +08:00
|
|
|
* Normally this should be same as cpufreq policy->related_cpus.
|
2015-02-27 03:00:29 +08:00
|
|
|
* @capacitance: dynamic power coefficient for these cpus
|
2013-04-18 01:12:15 +08:00
|
|
|
*
|
|
|
|
* This interface function registers the cpufreq cooling device with the name
|
|
|
|
* "thermal-cpufreq-%x". This api can support multiple instances of cpufreq
|
2013-09-13 07:26:45 +08:00
|
|
|
* cooling devices. It also gives the opportunity to link the cooling device
|
|
|
|
* with a device tree node, in order to bind it via the thermal DT code.
|
2013-04-18 01:12:15 +08:00
|
|
|
*
|
|
|
|
* Return: a valid struct thermal_cooling_device pointer on success,
|
|
|
|
* on failure, it returns a corresponding ERR_PTR().
|
2012-08-16 19:41:40 +08:00
|
|
|
*/
|
2013-09-13 07:26:45 +08:00
|
|
|
static struct thermal_cooling_device *
|
|
|
|
__cpufreq_cooling_register(struct device_node *np,
|
2017-12-05 13:32:46 +08:00
|
|
|
struct cpufreq_policy *policy, u32 capacitance)
|
2012-08-16 19:41:40 +08:00
|
|
|
{
|
2017-04-25 18:27:11 +08:00
|
|
|
struct thermal_cooling_device *cdev;
|
2017-04-25 18:27:10 +08:00
|
|
|
struct cpufreq_cooling_device *cpufreq_cdev;
|
2012-08-16 19:41:40 +08:00
|
|
|
char dev_name[THERMAL_NAME_LENGTH];
|
2015-02-27 03:00:29 +08:00
|
|
|
unsigned int freq, i, num_cpus;
|
2019-07-23 14:14:02 +08:00
|
|
|
struct device *dev;
|
2014-12-04 12:11:55 +08:00
|
|
|
int ret;
|
2016-08-17 23:14:59 +08:00
|
|
|
struct thermal_cooling_device_ops *cooling_ops;
|
2019-07-23 14:14:02 +08:00
|
|
|
|
|
|
|
dev = get_cpu_device(policy->cpu);
|
|
|
|
if (unlikely(!dev)) {
|
|
|
|
pr_warn("No cpu device for cpu %d\n", policy->cpu);
|
|
|
|
return ERR_PTR(-ENODEV);
|
|
|
|
}
|
|
|
|
|
2012-08-16 19:41:40 +08:00
|
|
|
|
2017-04-25 18:27:14 +08:00
|
|
|
if (IS_ERR_OR_NULL(policy)) {
|
2017-10-24 15:50:39 +08:00
|
|
|
pr_err("%s: cpufreq policy isn't valid: %p\n", __func__, policy);
|
2017-04-25 18:27:14 +08:00
|
|
|
return ERR_PTR(-EINVAL);
|
2016-06-03 13:28:47 +08:00
|
|
|
}
|
|
|
|
|
2017-04-25 18:27:15 +08:00
|
|
|
i = cpufreq_table_count_valid_entries(policy);
|
|
|
|
if (!i) {
|
|
|
|
pr_debug("%s: CPUFreq table not found or has no valid entries\n",
|
|
|
|
__func__);
|
2017-04-25 18:27:14 +08:00
|
|
|
return ERR_PTR(-ENODEV);
|
2012-08-16 19:41:40 +08:00
|
|
|
}
|
2014-12-04 12:11:43 +08:00
|
|
|
|
2017-04-25 18:27:10 +08:00
|
|
|
cpufreq_cdev = kzalloc(sizeof(*cpufreq_cdev), GFP_KERNEL);
|
2017-04-25 18:27:14 +08:00
|
|
|
if (!cpufreq_cdev)
|
|
|
|
return ERR_PTR(-ENOMEM);
|
2012-08-16 19:41:40 +08:00
|
|
|
|
2017-04-25 18:27:16 +08:00
|
|
|
cpufreq_cdev->policy = policy;
|
2017-04-25 18:27:14 +08:00
|
|
|
num_cpus = cpumask_weight(policy->related_cpus);
|
2017-04-25 18:27:20 +08:00
|
|
|
cpufreq_cdev->idle_time = kcalloc(num_cpus,
|
|
|
|
sizeof(*cpufreq_cdev->idle_time),
|
|
|
|
GFP_KERNEL);
|
|
|
|
if (!cpufreq_cdev->idle_time) {
|
2017-04-25 18:27:11 +08:00
|
|
|
cdev = ERR_PTR(-ENOMEM);
|
2015-02-27 03:00:29 +08:00
|
|
|
goto free_cdev;
|
|
|
|
}
|
|
|
|
|
2017-04-25 18:27:15 +08:00
|
|
|
/* max_level is an index, not a counter */
|
|
|
|
cpufreq_cdev->max_level = i - 1;
|
2014-12-04 12:12:02 +08:00
|
|
|
|
2017-05-23 15:03:06 +08:00
|
|
|
cpufreq_cdev->freq_table = kmalloc_array(i,
|
|
|
|
sizeof(*cpufreq_cdev->freq_table),
|
|
|
|
GFP_KERNEL);
|
2017-04-25 18:27:10 +08:00
|
|
|
if (!cpufreq_cdev->freq_table) {
|
2017-04-25 18:27:11 +08:00
|
|
|
cdev = ERR_PTR(-ENOMEM);
|
2017-04-25 18:27:20 +08:00
|
|
|
goto free_idle_time;
|
2014-12-04 12:12:06 +08:00
|
|
|
}
|
|
|
|
|
2016-12-22 01:47:05 +08:00
|
|
|
ret = ida_simple_get(&cpufreq_ida, 0, 0, GFP_KERNEL);
|
|
|
|
if (ret < 0) {
|
2017-04-25 18:27:11 +08:00
|
|
|
cdev = ERR_PTR(ret);
|
2017-04-25 18:27:19 +08:00
|
|
|
goto free_table;
|
2012-08-16 19:41:40 +08:00
|
|
|
}
|
2017-04-25 18:27:10 +08:00
|
|
|
cpufreq_cdev->id = ret;
|
2012-08-16 19:41:40 +08:00
|
|
|
|
2017-04-25 18:27:19 +08:00
|
|
|
snprintf(dev_name, sizeof(dev_name), "thermal-cpufreq-%d",
|
|
|
|
cpufreq_cdev->id);
|
|
|
|
|
2014-12-04 12:12:06 +08:00
|
|
|
/* Fill freq-table in descending order of frequencies */
|
2017-04-25 18:27:10 +08:00
|
|
|
for (i = 0, freq = -1; i <= cpufreq_cdev->max_level; i++) {
|
2017-04-25 18:27:15 +08:00
|
|
|
freq = find_next_max(policy->freq_table, freq);
|
2017-04-25 18:27:19 +08:00
|
|
|
cpufreq_cdev->freq_table[i].frequency = freq;
|
2014-12-04 12:12:06 +08:00
|
|
|
|
|
|
|
/* Warn for duplicate entries */
|
|
|
|
if (!freq)
|
|
|
|
pr_warn("%s: table has duplicate entries\n", __func__);
|
|
|
|
else
|
|
|
|
pr_debug("%s: freq:%u KHz\n", __func__, freq);
|
2012-08-16 19:41:40 +08:00
|
|
|
}
|
2014-12-04 12:12:06 +08:00
|
|
|
|
2017-04-25 18:27:19 +08:00
|
|
|
if (capacitance) {
|
|
|
|
ret = update_freq_table(cpufreq_cdev, capacitance);
|
|
|
|
if (ret) {
|
|
|
|
cdev = ERR_PTR(ret);
|
|
|
|
goto remove_ida;
|
|
|
|
}
|
|
|
|
|
|
|
|
cooling_ops = &cpufreq_power_cooling_ops;
|
|
|
|
} else {
|
|
|
|
cooling_ops = &cpufreq_cooling_ops;
|
|
|
|
}
|
2016-05-31 18:32:02 +08:00
|
|
|
|
cpufreq: Use per-policy frequency QoS
Replace the CPU device PM QoS used for the management of min and max
frequency constraints in cpufreq (and its users) with per-policy
frequency QoS to avoid problems with cpufreq policies covering
more then one CPU.
Namely, a cpufreq driver is registered with the subsys interface
which calls cpufreq_add_dev() for each CPU, starting from CPU0, so
currently the PM QoS notifiers are added to the first CPU in the
policy (i.e. CPU0 in the majority of cases).
In turn, when the cpufreq driver is unregistered, the subsys interface
doing that calls cpufreq_remove_dev() for each CPU, starting from CPU0,
and the PM QoS notifiers are only removed when cpufreq_remove_dev() is
called for the last CPU in the policy, say CPUx, which as a rule is
not CPU0 if the policy covers more than one CPU. Then, the PM QoS
notifiers cannot be removed, because CPUx does not have them, and
they are still there in the device PM QoS notifiers list of CPU0,
which prevents new PM QoS notifiers from being registered for CPU0
on the next attempt to register the cpufreq driver.
The same issue occurs when the first CPU in the policy goes offline
before unregistering the driver.
After this change it does not matter which CPU is the policy CPU at
the driver registration time and whether or not it is online all the
time, because the frequency QoS is per policy and not per CPU.
Fixes: 67d874c3b2c6 ("cpufreq: Register notifiers with the PM QoS framework")
Reported-by: Dmitry Osipenko <digetx@gmail.com>
Tested-by: Dmitry Osipenko <digetx@gmail.com>
Reported-by: Sudeep Holla <sudeep.holla@arm.com>
Tested-by: Sudeep Holla <sudeep.holla@arm.com>
Diagnosed-by: Viresh Kumar <viresh.kumar@linaro.org>
Link: https://lore.kernel.org/linux-pm/5ad2624194baa2f53acc1f1e627eb7684c577a19.1562210705.git.viresh.kumar@linaro.org/T/#md2d89e95906b8c91c15f582146173dce2e86e99f
Link: https://lore.kernel.org/linux-pm/20191017094612.6tbkwoq4harsjcqv@vireshk-i7/T/#m30d48cc23b9a80467fbaa16e30f90b3828a5a29b
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Acked-by: Viresh Kumar <viresh.kumar@linaro.org>
2019-10-16 18:47:06 +08:00
|
|
|
ret = freq_qos_add_request(&policy->constraints,
|
|
|
|
&cpufreq_cdev->qos_req, FREQ_QOS_MAX,
|
|
|
|
cpufreq_cdev->freq_table[0].frequency);
|
2019-07-23 14:14:02 +08:00
|
|
|
if (ret < 0) {
|
|
|
|
pr_err("%s: Failed to add freq constraint (%d)\n", __func__,
|
|
|
|
ret);
|
|
|
|
cdev = ERR_PTR(ret);
|
|
|
|
goto remove_ida;
|
|
|
|
}
|
|
|
|
|
2017-04-25 18:27:11 +08:00
|
|
|
cdev = thermal_of_cooling_device_register(np, dev_name, cpufreq_cdev,
|
|
|
|
cooling_ops);
|
|
|
|
if (IS_ERR(cdev))
|
2019-07-23 14:14:02 +08:00
|
|
|
goto remove_qos_req;
|
2014-12-04 12:11:51 +08:00
|
|
|
|
thermal: cpu_cooling: fix lockdep problems in cpu_cooling
A recent change to the cpu_cooling code introduced a AB-BA deadlock
scenario between the cpufreq_policy_notifier_list rwsem and the
cooling_cpufreq_lock. This is caused by cooling_cpufreq_lock being held
before the registration/removal of the notifier block (an operation
which takes the rwsem), and the notifier code itself which takes the
locks in the reverse order:
======================================================
[ INFO: possible circular locking dependency detected ]
3.18.0+ #1453 Not tainted
-------------------------------------------------------
rc.local/770 is trying to acquire lock:
(cooling_cpufreq_lock){+.+.+.}, at: [<c04abfc4>] cpufreq_thermal_notifier+0x34/0xfc
but task is already holding lock:
((cpufreq_policy_notifier_list).rwsem){++++.+}, at: [<c0042f04>] __blocking_notifier_call_chain+0x34/0x68
which lock already depends on the new lock.
the existing dependency chain (in reverse order) is:
-> #1 ((cpufreq_policy_notifier_list).rwsem){++++.+}:
[<c06bc3b0>] down_write+0x44/0x9c
[<c0043444>] blocking_notifier_chain_register+0x28/0xd8
[<c04ad610>] cpufreq_register_notifier+0x68/0x90
[<c04abe4c>] __cpufreq_cooling_register.part.1+0x120/0x180
[<c04abf44>] __cpufreq_cooling_register+0x98/0xa4
[<c04abf8c>] cpufreq_cooling_register+0x18/0x1c
[<bf0046f8>] imx_thermal_probe+0x1c0/0x470 [imx_thermal]
[<c037cef8>] platform_drv_probe+0x50/0xac
[<c037b710>] driver_probe_device+0x114/0x234
[<c037b8cc>] __driver_attach+0x9c/0xa0
[<c0379d68>] bus_for_each_dev+0x5c/0x90
[<c037b204>] driver_attach+0x24/0x28
[<c037ae7c>] bus_add_driver+0xe0/0x1d8
[<c037c0cc>] driver_register+0x80/0xfc
[<c037cd80>] __platform_driver_register+0x50/0x64
[<bf007018>] 0xbf007018
[<c0008a5c>] do_one_initcall+0x88/0x1d8
[<c0095da4>] load_module+0x1768/0x1ef8
[<c0096614>] SyS_init_module+0xe0/0xf4
[<c000ec00>] ret_fast_syscall+0x0/0x48
-> #0 (cooling_cpufreq_lock){+.+.+.}:
[<c00619f8>] lock_acquire+0xb0/0x124
[<c06ba3b4>] mutex_lock_nested+0x5c/0x3d8
[<c04abfc4>] cpufreq_thermal_notifier+0x34/0xfc
[<c0042bf4>] notifier_call_chain+0x4c/0x8c
[<c0042f20>] __blocking_notifier_call_chain+0x50/0x68
[<c0042f58>] blocking_notifier_call_chain+0x20/0x28
[<c04ae62c>] cpufreq_set_policy+0x7c/0x1d0
[<c04af3cc>] store_scaling_governor+0x74/0x9c
[<c04ad418>] store+0x90/0xc0
[<c0175384>] sysfs_kf_write+0x54/0x58
[<c01746b4>] kernfs_fop_write+0xdc/0x190
[<c010dcc0>] vfs_write+0xac/0x1b4
[<c010dfec>] SyS_write+0x44/0x90
[<c000ec00>] ret_fast_syscall+0x0/0x48
other info that might help us debug this:
Possible unsafe locking scenario:
CPU0 CPU1
---- ----
lock((cpufreq_policy_notifier_list).rwsem);
lock(cooling_cpufreq_lock);
lock((cpufreq_policy_notifier_list).rwsem);
lock(cooling_cpufreq_lock);
*** DEADLOCK ***
7 locks held by rc.local/770:
#0: (sb_writers#6){.+.+.+}, at: [<c010dda0>] vfs_write+0x18c/0x1b4
#1: (&of->mutex){+.+.+.}, at: [<c0174678>] kernfs_fop_write+0xa0/0x190
#2: (s_active#52){.+.+.+}, at: [<c0174680>] kernfs_fop_write+0xa8/0x190
#3: (cpu_hotplug.lock){++++++}, at: [<c0026a60>] get_online_cpus+0x34/0x90
#4: (cpufreq_rwsem){.+.+.+}, at: [<c04ad3e0>] store+0x58/0xc0
#5: (&policy->rwsem){+.+.+.}, at: [<c04ad3f8>] store+0x70/0xc0
#6: ((cpufreq_policy_notifier_list).rwsem){++++.+}, at: [<c0042f04>] __blocking_notifier_call_chain+0x34/0x68
stack backtrace:
CPU: 0 PID: 770 Comm: rc.local Not tainted 3.18.0+ #1453
Hardware name: Freescale i.MX6 Quad/DualLite (Device Tree)
Backtrace:
[<c00121c8>] (dump_backtrace) from [<c0012360>] (show_stack+0x18/0x1c)
r6:c0b85a80 r5:c0b75630 r4:00000000 r3:00000000
[<c0012348>] (show_stack) from [<c06b6c48>] (dump_stack+0x7c/0x98)
[<c06b6bcc>] (dump_stack) from [<c06b42a4>] (print_circular_bug+0x28c/0x2d8)
r4:c0b85a80 r3:d0071d40
[<c06b4018>] (print_circular_bug) from [<c00613b0>] (__lock_acquire+0x1acc/0x1bb0)
r10:c0b50660 r8:c09e6d80 r7:d0071d40 r6:c11d0f0c r5:00000007 r4:d0072240
[<c005f8e4>] (__lock_acquire) from [<c00619f8>] (lock_acquire+0xb0/0x124)
r10:00000000 r9:c04abfc4 r8:00000000 r7:00000000 r6:00000000 r5:c0a06f0c
r4:00000000
[<c0061948>] (lock_acquire) from [<c06ba3b4>] (mutex_lock_nested+0x5c/0x3d8)
r10:ec853800 r9:c0a06ed4 r8:d0071d40 r7:c0a06ed4 r6:c11d0f0c r5:00000000
r4:c04abfc4
[<c06ba358>] (mutex_lock_nested) from [<c04abfc4>] (cpufreq_thermal_notifier+0x34/0xfc)
r10:ec853800 r9:ec85380c r8:d00d7d3c r7:c0a06ed4 r6:d00d7d3c r5:00000000
r4:fffffffe
[<c04abf90>] (cpufreq_thermal_notifier) from [<c0042bf4>] (notifier_call_chain+0x4c/0x8c)
r7:00000000 r6:00000000 r5:00000000 r4:fffffffe
[<c0042ba8>] (notifier_call_chain) from [<c0042f20>] (__blocking_notifier_call_chain+0x50/0x68)
r8:c0a072a4 r7:00000000 r6:d00d7d3c r5:ffffffff r4:c0a06fc8 r3:ffffffff
[<c0042ed0>] (__blocking_notifier_call_chain) from [<c0042f58>] (blocking_notifier_call_chain+0x20/0x28)
r7:ec98b540 r6:c13ebc80 r5:ed76e600 r4:d00d7d3c
[<c0042f38>] (blocking_notifier_call_chain) from [<c04ae62c>] (cpufreq_set_policy+0x7c/0x1d0)
[<c04ae5b0>] (cpufreq_set_policy) from [<c04af3cc>] (store_scaling_governor+0x74/0x9c)
r7:ec98b540 r6:0000000c r5:ec98b540 r4:ed76e600
[<c04af358>] (store_scaling_governor) from [<c04ad418>] (store+0x90/0xc0)
r6:0000000c r5:ed76e6d4 r4:ed76e600
[<c04ad388>] (store) from [<c0175384>] (sysfs_kf_write+0x54/0x58)
r8:0000000c r7:d00d7f78 r6:ec98b540 r5:0000000c r4:ec853800 r3:0000000c
[<c0175330>] (sysfs_kf_write) from [<c01746b4>] (kernfs_fop_write+0xdc/0x190)
r6:ec98b540 r5:00000000 r4:00000000 r3:c0175330
[<c01745d8>] (kernfs_fop_write) from [<c010dcc0>] (vfs_write+0xac/0x1b4)
r10:0162aa70 r9:d00d6000 r8:0000000c r7:d00d7f78 r6:0162aa70 r5:0000000c
r4:eccde500
[<c010dc14>] (vfs_write) from [<c010dfec>] (SyS_write+0x44/0x90)
r10:0162aa70 r8:0000000c r7:eccde500 r6:eccde500 r5:00000000 r4:00000000
[<c010dfa8>] (SyS_write) from [<c000ec00>] (ret_fast_syscall+0x0/0x48)
r10:00000000 r8:c000edc4 r7:00000004 r6:000216cc r5:0000000c r4:0162aa70
Solve this by moving to finer grained locking - use one mutex to protect
the cpufreq_dev_list as a whole, and a separate lock to ensure correct
ordering of cpufreq notifier registration and removal.
cooling_list_lock is taken within cooling_cpufreq_lock on
(un)registration to preserve the behavior of the code, i.e. to
atomically add/remove to the list and (un)register the notifier.
Fixes: 2dcd851fe4b4 ("thermal: cpu_cooling: Update always cpufreq policy with
Reviewed-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Eduardo Valentin <edubezval@gmail.com>
2015-08-12 17:52:16 +08:00
|
|
|
mutex_lock(&cooling_list_lock);
|
2017-04-25 18:27:10 +08:00
|
|
|
list_add(&cpufreq_cdev->node, &cpufreq_cdev_list);
|
2017-03-11 02:33:28 +08:00
|
|
|
mutex_unlock(&cooling_list_lock);
|
thermal: cpu_cooling: fix lockdep problems in cpu_cooling
A recent change to the cpu_cooling code introduced a AB-BA deadlock
scenario between the cpufreq_policy_notifier_list rwsem and the
cooling_cpufreq_lock. This is caused by cooling_cpufreq_lock being held
before the registration/removal of the notifier block (an operation
which takes the rwsem), and the notifier code itself which takes the
locks in the reverse order:
======================================================
[ INFO: possible circular locking dependency detected ]
3.18.0+ #1453 Not tainted
-------------------------------------------------------
rc.local/770 is trying to acquire lock:
(cooling_cpufreq_lock){+.+.+.}, at: [<c04abfc4>] cpufreq_thermal_notifier+0x34/0xfc
but task is already holding lock:
((cpufreq_policy_notifier_list).rwsem){++++.+}, at: [<c0042f04>] __blocking_notifier_call_chain+0x34/0x68
which lock already depends on the new lock.
the existing dependency chain (in reverse order) is:
-> #1 ((cpufreq_policy_notifier_list).rwsem){++++.+}:
[<c06bc3b0>] down_write+0x44/0x9c
[<c0043444>] blocking_notifier_chain_register+0x28/0xd8
[<c04ad610>] cpufreq_register_notifier+0x68/0x90
[<c04abe4c>] __cpufreq_cooling_register.part.1+0x120/0x180
[<c04abf44>] __cpufreq_cooling_register+0x98/0xa4
[<c04abf8c>] cpufreq_cooling_register+0x18/0x1c
[<bf0046f8>] imx_thermal_probe+0x1c0/0x470 [imx_thermal]
[<c037cef8>] platform_drv_probe+0x50/0xac
[<c037b710>] driver_probe_device+0x114/0x234
[<c037b8cc>] __driver_attach+0x9c/0xa0
[<c0379d68>] bus_for_each_dev+0x5c/0x90
[<c037b204>] driver_attach+0x24/0x28
[<c037ae7c>] bus_add_driver+0xe0/0x1d8
[<c037c0cc>] driver_register+0x80/0xfc
[<c037cd80>] __platform_driver_register+0x50/0x64
[<bf007018>] 0xbf007018
[<c0008a5c>] do_one_initcall+0x88/0x1d8
[<c0095da4>] load_module+0x1768/0x1ef8
[<c0096614>] SyS_init_module+0xe0/0xf4
[<c000ec00>] ret_fast_syscall+0x0/0x48
-> #0 (cooling_cpufreq_lock){+.+.+.}:
[<c00619f8>] lock_acquire+0xb0/0x124
[<c06ba3b4>] mutex_lock_nested+0x5c/0x3d8
[<c04abfc4>] cpufreq_thermal_notifier+0x34/0xfc
[<c0042bf4>] notifier_call_chain+0x4c/0x8c
[<c0042f20>] __blocking_notifier_call_chain+0x50/0x68
[<c0042f58>] blocking_notifier_call_chain+0x20/0x28
[<c04ae62c>] cpufreq_set_policy+0x7c/0x1d0
[<c04af3cc>] store_scaling_governor+0x74/0x9c
[<c04ad418>] store+0x90/0xc0
[<c0175384>] sysfs_kf_write+0x54/0x58
[<c01746b4>] kernfs_fop_write+0xdc/0x190
[<c010dcc0>] vfs_write+0xac/0x1b4
[<c010dfec>] SyS_write+0x44/0x90
[<c000ec00>] ret_fast_syscall+0x0/0x48
other info that might help us debug this:
Possible unsafe locking scenario:
CPU0 CPU1
---- ----
lock((cpufreq_policy_notifier_list).rwsem);
lock(cooling_cpufreq_lock);
lock((cpufreq_policy_notifier_list).rwsem);
lock(cooling_cpufreq_lock);
*** DEADLOCK ***
7 locks held by rc.local/770:
#0: (sb_writers#6){.+.+.+}, at: [<c010dda0>] vfs_write+0x18c/0x1b4
#1: (&of->mutex){+.+.+.}, at: [<c0174678>] kernfs_fop_write+0xa0/0x190
#2: (s_active#52){.+.+.+}, at: [<c0174680>] kernfs_fop_write+0xa8/0x190
#3: (cpu_hotplug.lock){++++++}, at: [<c0026a60>] get_online_cpus+0x34/0x90
#4: (cpufreq_rwsem){.+.+.+}, at: [<c04ad3e0>] store+0x58/0xc0
#5: (&policy->rwsem){+.+.+.}, at: [<c04ad3f8>] store+0x70/0xc0
#6: ((cpufreq_policy_notifier_list).rwsem){++++.+}, at: [<c0042f04>] __blocking_notifier_call_chain+0x34/0x68
stack backtrace:
CPU: 0 PID: 770 Comm: rc.local Not tainted 3.18.0+ #1453
Hardware name: Freescale i.MX6 Quad/DualLite (Device Tree)
Backtrace:
[<c00121c8>] (dump_backtrace) from [<c0012360>] (show_stack+0x18/0x1c)
r6:c0b85a80 r5:c0b75630 r4:00000000 r3:00000000
[<c0012348>] (show_stack) from [<c06b6c48>] (dump_stack+0x7c/0x98)
[<c06b6bcc>] (dump_stack) from [<c06b42a4>] (print_circular_bug+0x28c/0x2d8)
r4:c0b85a80 r3:d0071d40
[<c06b4018>] (print_circular_bug) from [<c00613b0>] (__lock_acquire+0x1acc/0x1bb0)
r10:c0b50660 r8:c09e6d80 r7:d0071d40 r6:c11d0f0c r5:00000007 r4:d0072240
[<c005f8e4>] (__lock_acquire) from [<c00619f8>] (lock_acquire+0xb0/0x124)
r10:00000000 r9:c04abfc4 r8:00000000 r7:00000000 r6:00000000 r5:c0a06f0c
r4:00000000
[<c0061948>] (lock_acquire) from [<c06ba3b4>] (mutex_lock_nested+0x5c/0x3d8)
r10:ec853800 r9:c0a06ed4 r8:d0071d40 r7:c0a06ed4 r6:c11d0f0c r5:00000000
r4:c04abfc4
[<c06ba358>] (mutex_lock_nested) from [<c04abfc4>] (cpufreq_thermal_notifier+0x34/0xfc)
r10:ec853800 r9:ec85380c r8:d00d7d3c r7:c0a06ed4 r6:d00d7d3c r5:00000000
r4:fffffffe
[<c04abf90>] (cpufreq_thermal_notifier) from [<c0042bf4>] (notifier_call_chain+0x4c/0x8c)
r7:00000000 r6:00000000 r5:00000000 r4:fffffffe
[<c0042ba8>] (notifier_call_chain) from [<c0042f20>] (__blocking_notifier_call_chain+0x50/0x68)
r8:c0a072a4 r7:00000000 r6:d00d7d3c r5:ffffffff r4:c0a06fc8 r3:ffffffff
[<c0042ed0>] (__blocking_notifier_call_chain) from [<c0042f58>] (blocking_notifier_call_chain+0x20/0x28)
r7:ec98b540 r6:c13ebc80 r5:ed76e600 r4:d00d7d3c
[<c0042f38>] (blocking_notifier_call_chain) from [<c04ae62c>] (cpufreq_set_policy+0x7c/0x1d0)
[<c04ae5b0>] (cpufreq_set_policy) from [<c04af3cc>] (store_scaling_governor+0x74/0x9c)
r7:ec98b540 r6:0000000c r5:ec98b540 r4:ed76e600
[<c04af358>] (store_scaling_governor) from [<c04ad418>] (store+0x90/0xc0)
r6:0000000c r5:ed76e6d4 r4:ed76e600
[<c04ad388>] (store) from [<c0175384>] (sysfs_kf_write+0x54/0x58)
r8:0000000c r7:d00d7f78 r6:ec98b540 r5:0000000c r4:ec853800 r3:0000000c
[<c0175330>] (sysfs_kf_write) from [<c01746b4>] (kernfs_fop_write+0xdc/0x190)
r6:ec98b540 r5:00000000 r4:00000000 r3:c0175330
[<c01745d8>] (kernfs_fop_write) from [<c010dcc0>] (vfs_write+0xac/0x1b4)
r10:0162aa70 r9:d00d6000 r8:0000000c r7:d00d7f78 r6:0162aa70 r5:0000000c
r4:eccde500
[<c010dc14>] (vfs_write) from [<c010dfec>] (SyS_write+0x44/0x90)
r10:0162aa70 r8:0000000c r7:eccde500 r6:eccde500 r5:00000000 r4:00000000
[<c010dfa8>] (SyS_write) from [<c000ec00>] (ret_fast_syscall+0x0/0x48)
r10:00000000 r8:c000edc4 r7:00000004 r6:000216cc r5:0000000c r4:0162aa70
Solve this by moving to finer grained locking - use one mutex to protect
the cpufreq_dev_list as a whole, and a separate lock to ensure correct
ordering of cpufreq notifier registration and removal.
cooling_list_lock is taken within cooling_cpufreq_lock on
(un)registration to preserve the behavior of the code, i.e. to
atomically add/remove to the list and (un)register the notifier.
Fixes: 2dcd851fe4b4 ("thermal: cpu_cooling: Update always cpufreq policy with
Reviewed-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Eduardo Valentin <edubezval@gmail.com>
2015-08-12 17:52:16 +08:00
|
|
|
|
2017-04-25 18:27:14 +08:00
|
|
|
return cdev;
|
2014-12-04 12:11:58 +08:00
|
|
|
|
2019-07-23 14:14:02 +08:00
|
|
|
remove_qos_req:
|
cpufreq: Use per-policy frequency QoS
Replace the CPU device PM QoS used for the management of min and max
frequency constraints in cpufreq (and its users) with per-policy
frequency QoS to avoid problems with cpufreq policies covering
more then one CPU.
Namely, a cpufreq driver is registered with the subsys interface
which calls cpufreq_add_dev() for each CPU, starting from CPU0, so
currently the PM QoS notifiers are added to the first CPU in the
policy (i.e. CPU0 in the majority of cases).
In turn, when the cpufreq driver is unregistered, the subsys interface
doing that calls cpufreq_remove_dev() for each CPU, starting from CPU0,
and the PM QoS notifiers are only removed when cpufreq_remove_dev() is
called for the last CPU in the policy, say CPUx, which as a rule is
not CPU0 if the policy covers more than one CPU. Then, the PM QoS
notifiers cannot be removed, because CPUx does not have them, and
they are still there in the device PM QoS notifiers list of CPU0,
which prevents new PM QoS notifiers from being registered for CPU0
on the next attempt to register the cpufreq driver.
The same issue occurs when the first CPU in the policy goes offline
before unregistering the driver.
After this change it does not matter which CPU is the policy CPU at
the driver registration time and whether or not it is online all the
time, because the frequency QoS is per policy and not per CPU.
Fixes: 67d874c3b2c6 ("cpufreq: Register notifiers with the PM QoS framework")
Reported-by: Dmitry Osipenko <digetx@gmail.com>
Tested-by: Dmitry Osipenko <digetx@gmail.com>
Reported-by: Sudeep Holla <sudeep.holla@arm.com>
Tested-by: Sudeep Holla <sudeep.holla@arm.com>
Diagnosed-by: Viresh Kumar <viresh.kumar@linaro.org>
Link: https://lore.kernel.org/linux-pm/5ad2624194baa2f53acc1f1e627eb7684c577a19.1562210705.git.viresh.kumar@linaro.org/T/#md2d89e95906b8c91c15f582146173dce2e86e99f
Link: https://lore.kernel.org/linux-pm/20191017094612.6tbkwoq4harsjcqv@vireshk-i7/T/#m30d48cc23b9a80467fbaa16e30f90b3828a5a29b
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Acked-by: Viresh Kumar <viresh.kumar@linaro.org>
2019-10-16 18:47:06 +08:00
|
|
|
freq_qos_remove_request(&cpufreq_cdev->qos_req);
|
2016-12-22 01:47:05 +08:00
|
|
|
remove_ida:
|
2017-04-25 18:27:10 +08:00
|
|
|
ida_simple_remove(&cpufreq_ida, cpufreq_cdev->id);
|
2014-12-04 12:12:06 +08:00
|
|
|
free_table:
|
2017-04-25 18:27:10 +08:00
|
|
|
kfree(cpufreq_cdev->freq_table);
|
2017-04-25 18:27:20 +08:00
|
|
|
free_idle_time:
|
|
|
|
kfree(cpufreq_cdev->idle_time);
|
2014-12-04 12:11:58 +08:00
|
|
|
free_cdev:
|
2017-04-25 18:27:10 +08:00
|
|
|
kfree(cpufreq_cdev);
|
2017-04-25 18:27:11 +08:00
|
|
|
return cdev;
|
2012-08-16 19:41:40 +08:00
|
|
|
}
|
2013-09-13 07:26:45 +08:00
|
|
|
|
|
|
|
/**
|
|
|
|
* cpufreq_cooling_register - function to create cpufreq cooling device.
|
2017-04-25 18:27:14 +08:00
|
|
|
* @policy: cpufreq policy
|
2013-09-13 07:26:45 +08:00
|
|
|
*
|
|
|
|
* This interface function registers the cpufreq cooling device with the name
|
|
|
|
* "thermal-cpufreq-%x". This api can support multiple instances of cpufreq
|
|
|
|
* cooling devices.
|
|
|
|
*
|
|
|
|
* Return: a valid struct thermal_cooling_device pointer on success,
|
|
|
|
* on failure, it returns a corresponding ERR_PTR().
|
|
|
|
*/
|
|
|
|
struct thermal_cooling_device *
|
2017-04-25 18:27:14 +08:00
|
|
|
cpufreq_cooling_register(struct cpufreq_policy *policy)
|
2013-09-13 07:26:45 +08:00
|
|
|
{
|
2017-12-05 13:32:46 +08:00
|
|
|
return __cpufreq_cooling_register(NULL, policy, 0);
|
2013-09-13 07:26:45 +08:00
|
|
|
}
|
2013-04-18 01:11:57 +08:00
|
|
|
EXPORT_SYMBOL_GPL(cpufreq_cooling_register);
|
2012-08-16 19:41:40 +08:00
|
|
|
|
2013-09-13 07:26:45 +08:00
|
|
|
/**
|
|
|
|
* of_cpufreq_cooling_register - function to create cpufreq cooling device.
|
2017-04-25 18:27:14 +08:00
|
|
|
* @policy: cpufreq policy
|
2013-09-13 07:26:45 +08:00
|
|
|
*
|
|
|
|
* This interface function registers the cpufreq cooling device with the name
|
|
|
|
* "thermal-cpufreq-%x". This api can support multiple instances of cpufreq
|
|
|
|
* cooling devices. Using this API, the cpufreq cooling device will be
|
|
|
|
* linked to the device tree node provided.
|
|
|
|
*
|
2015-02-27 03:00:29 +08:00
|
|
|
* Using this function, the cooling device will implement the power
|
|
|
|
* extensions by using a simple cpu power model. The cpus must have
|
|
|
|
* registered their OPPs using the OPP library.
|
|
|
|
*
|
2017-12-05 13:32:43 +08:00
|
|
|
* It also takes into account, if property present in policy CPU node, the
|
|
|
|
* static power consumed by the cpu.
|
2015-02-27 03:00:29 +08:00
|
|
|
*
|
|
|
|
* Return: a valid struct thermal_cooling_device pointer on success,
|
2017-12-05 13:32:43 +08:00
|
|
|
* and NULL on failure.
|
2015-02-27 03:00:29 +08:00
|
|
|
*/
|
|
|
|
struct thermal_cooling_device *
|
2017-12-05 13:32:45 +08:00
|
|
|
of_cpufreq_cooling_register(struct cpufreq_policy *policy)
|
2015-02-27 03:00:29 +08:00
|
|
|
{
|
2017-12-05 13:32:43 +08:00
|
|
|
struct device_node *np = of_get_cpu_node(policy->cpu, NULL);
|
|
|
|
struct thermal_cooling_device *cdev = NULL;
|
|
|
|
u32 capacitance = 0;
|
|
|
|
|
|
|
|
if (!np) {
|
|
|
|
pr_err("cpu_cooling: OF node not available for cpu%d\n",
|
|
|
|
policy->cpu);
|
|
|
|
return NULL;
|
|
|
|
}
|
2015-02-27 03:00:29 +08:00
|
|
|
|
2017-12-05 13:32:43 +08:00
|
|
|
if (of_find_property(np, "#cooling-cells", NULL)) {
|
|
|
|
of_property_read_u32(np, "dynamic-power-coefficient",
|
|
|
|
&capacitance);
|
|
|
|
|
2017-12-05 13:32:46 +08:00
|
|
|
cdev = __cpufreq_cooling_register(np, policy, capacitance);
|
2017-12-05 13:32:43 +08:00
|
|
|
if (IS_ERR(cdev)) {
|
2019-01-21 17:42:23 +08:00
|
|
|
pr_err("cpu_cooling: cpu%d failed to register as cooling device: %ld\n",
|
2017-12-05 13:32:43 +08:00
|
|
|
policy->cpu, PTR_ERR(cdev));
|
|
|
|
cdev = NULL;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
of_node_put(np);
|
|
|
|
return cdev;
|
2015-02-27 03:00:29 +08:00
|
|
|
}
|
2017-12-05 13:32:45 +08:00
|
|
|
EXPORT_SYMBOL_GPL(of_cpufreq_cooling_register);
|
2015-02-27 03:00:29 +08:00
|
|
|
|
2012-08-16 19:41:40 +08:00
|
|
|
/**
|
|
|
|
* cpufreq_cooling_unregister - function to remove cpufreq cooling device.
|
|
|
|
* @cdev: thermal cooling device pointer.
|
2013-04-18 01:12:16 +08:00
|
|
|
*
|
|
|
|
* This interface function unregisters the "thermal-cpufreq-%x" cooling device.
|
2012-08-16 19:41:40 +08:00
|
|
|
*/
|
|
|
|
void cpufreq_cooling_unregister(struct thermal_cooling_device *cdev)
|
|
|
|
{
|
2017-04-25 18:27:10 +08:00
|
|
|
struct cpufreq_cooling_device *cpufreq_cdev;
|
2012-08-16 19:41:40 +08:00
|
|
|
|
2013-08-15 22:54:46 +08:00
|
|
|
if (!cdev)
|
|
|
|
return;
|
|
|
|
|
2017-04-25 18:27:10 +08:00
|
|
|
cpufreq_cdev = cdev->devdata;
|
2012-08-16 19:41:40 +08:00
|
|
|
|
2016-12-22 01:47:05 +08:00
|
|
|
mutex_lock(&cooling_list_lock);
|
2017-04-25 18:27:10 +08:00
|
|
|
list_del(&cpufreq_cdev->node);
|
2017-03-11 02:33:28 +08:00
|
|
|
mutex_unlock(&cooling_list_lock);
|
|
|
|
|
2019-04-28 17:51:05 +08:00
|
|
|
thermal_cooling_device_unregister(cdev);
|
cpufreq: Use per-policy frequency QoS
Replace the CPU device PM QoS used for the management of min and max
frequency constraints in cpufreq (and its users) with per-policy
frequency QoS to avoid problems with cpufreq policies covering
more then one CPU.
Namely, a cpufreq driver is registered with the subsys interface
which calls cpufreq_add_dev() for each CPU, starting from CPU0, so
currently the PM QoS notifiers are added to the first CPU in the
policy (i.e. CPU0 in the majority of cases).
In turn, when the cpufreq driver is unregistered, the subsys interface
doing that calls cpufreq_remove_dev() for each CPU, starting from CPU0,
and the PM QoS notifiers are only removed when cpufreq_remove_dev() is
called for the last CPU in the policy, say CPUx, which as a rule is
not CPU0 if the policy covers more than one CPU. Then, the PM QoS
notifiers cannot be removed, because CPUx does not have them, and
they are still there in the device PM QoS notifiers list of CPU0,
which prevents new PM QoS notifiers from being registered for CPU0
on the next attempt to register the cpufreq driver.
The same issue occurs when the first CPU in the policy goes offline
before unregistering the driver.
After this change it does not matter which CPU is the policy CPU at
the driver registration time and whether or not it is online all the
time, because the frequency QoS is per policy and not per CPU.
Fixes: 67d874c3b2c6 ("cpufreq: Register notifiers with the PM QoS framework")
Reported-by: Dmitry Osipenko <digetx@gmail.com>
Tested-by: Dmitry Osipenko <digetx@gmail.com>
Reported-by: Sudeep Holla <sudeep.holla@arm.com>
Tested-by: Sudeep Holla <sudeep.holla@arm.com>
Diagnosed-by: Viresh Kumar <viresh.kumar@linaro.org>
Link: https://lore.kernel.org/linux-pm/5ad2624194baa2f53acc1f1e627eb7684c577a19.1562210705.git.viresh.kumar@linaro.org/T/#md2d89e95906b8c91c15f582146173dce2e86e99f
Link: https://lore.kernel.org/linux-pm/20191017094612.6tbkwoq4harsjcqv@vireshk-i7/T/#m30d48cc23b9a80467fbaa16e30f90b3828a5a29b
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Acked-by: Viresh Kumar <viresh.kumar@linaro.org>
2019-10-16 18:47:06 +08:00
|
|
|
freq_qos_remove_request(&cpufreq_cdev->qos_req);
|
2017-04-25 18:27:10 +08:00
|
|
|
ida_simple_remove(&cpufreq_ida, cpufreq_cdev->id);
|
2017-04-25 18:27:20 +08:00
|
|
|
kfree(cpufreq_cdev->idle_time);
|
2017-04-25 18:27:10 +08:00
|
|
|
kfree(cpufreq_cdev->freq_table);
|
|
|
|
kfree(cpufreq_cdev);
|
2012-08-16 19:41:40 +08:00
|
|
|
}
|
2013-04-18 01:11:57 +08:00
|
|
|
EXPORT_SYMBOL_GPL(cpufreq_cooling_unregister);
|