linux-stable/drivers/opp/opp.h

277 lines
9.4 KiB
C
Raw Normal View History

/* SPDX-License-Identifier: GPL-2.0-only */
/*
* Generic OPP Interface
*
* Copyright (C) 2009-2010 Texas Instruments Incorporated.
* Nishanth Menon
* Romit Dasgupta
* Kevin Hilman
*/
#ifndef __DRIVER_OPP_H__
#define __DRIVER_OPP_H__
#include <linux/device.h>
#include <linux/interconnect.h>
#include <linux/kernel.h>
#include <linux/kref.h>
#include <linux/list.h>
#include <linux/limits.h>
#include <linux/pm_opp.h>
#include <linux/notifier.h>
struct clk;
struct regulator;
PM / OPP: Protect updates to list_dev with mutex dev_opp_list_lock is used everywhere to protect device and OPP lists, but dev_pm_opp_set_sharing_cpus() is missed somehow. And instead we used rcu-lock, which wouldn't help here as we are adding a new list_dev. This also fixes a problem where we have called kzalloc(..., GFP_KERNEL) from within rcu-lock, which isn't allowed as kzalloc can sleep when called with GFP_KERNEL. With CONFIG_DEBUG_ATOMIC_SLEEP set, we get following lockdep-splat: include/linux/rcupdate.h:578 Illegal context switch in RCU read-side critical section! other info that might help us debug this: rcu_scheduler_active = 1, debug_locks = 0 5 locks held by swapper/0/1: #0: (&dev->mutex){......}, at: [<c02f68f4>] __driver_attach+0x48/0x98 #1: (&dev->mutex){......}, at: [<c02f6904>] __driver_attach+0x58/0x98 #2: (cpu_hotplug.lock){++++++}, at: [<c00249d0>] get_online_cpus+0x40/0xb0 #3: (subsys mutex#5){+.+.+.}, at: [<c02f4f8c>] subsys_interface_register+0x44/0xdc #4: (rcu_read_lock){......}, at: [<c0305c80>] dev_pm_opp_set_sharing_cpus+0x0/0x1e4 stack backtrace: CPU: 1 PID: 1 Comm: swapper/0 Tainted: G W 4.3.0-rc7-00047-g81f5932958a8 #59 Hardware name: SAMSUNG EXYNOS (Flattened Device Tree) [<c0016874>] (unwind_backtrace) from [<c001355c>] (show_stack+0x10/0x14) [<c001355c>] (show_stack) from [<c022553c>] (dump_stack+0x94/0xbc) [<c022553c>] (dump_stack) from [<c004904c>] (___might_sleep+0x24c/0x298) [<c004904c>] (___might_sleep) from [<c00f07e4>] (kmem_cache_alloc+0xe8/0x164) [<c00f07e4>] (kmem_cache_alloc) from [<c0305354>] (_add_list_dev+0x30/0x58) [<c0305354>] (_add_list_dev) from [<c0305d50>] (dev_pm_opp_set_sharing_cpus+0xd0/0x1e4) [<c0305d50>] (dev_pm_opp_set_sharing_cpus) from [<c040eda4>] (cpufreq_init+0x4cc/0x62c) [<c040eda4>] (cpufreq_init) from [<c040a964>] (cpufreq_online+0xbc/0x73c) [<c040a964>] (cpufreq_online) from [<c02f4fe0>] (subsys_interface_register+0x98/0xdc) [<c02f4fe0>] (subsys_interface_register) from [<c040a640>] (cpufreq_register_driver+0x110/0x17c) [<c040a640>] (cpufreq_register_driver) from [<c040ef64>] (dt_cpufreq_probe+0x60/0x8c) [<c040ef64>] (dt_cpufreq_probe) from [<c02f8084>] (platform_drv_probe+0x44/0xa4) [<c02f8084>] (platform_drv_probe) from [<c02f67c0>] (driver_probe_device+0x208/0x2f4) [<c02f67c0>] (driver_probe_device) from [<c02f6940>] (__driver_attach+0x94/0x98) [<c02f6940>] (__driver_attach) from [<c02f4c1c>] (bus_for_each_dev+0x68/0x9c) Reported-by: Michael Turquette <mturquette@baylibre.com> Reviewed-by: Stephen Boyd <sboyd@codeaurora.org> Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org> Cc: 4.3 <stable@vger.kernel.org> # 4.3 Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
2015-11-05 08:51:19 +00:00
/* Lock to allow exclusive modification to the device and opp lists */
extern struct mutex opp_table_lock;
PM / OPP: Protect updates to list_dev with mutex dev_opp_list_lock is used everywhere to protect device and OPP lists, but dev_pm_opp_set_sharing_cpus() is missed somehow. And instead we used rcu-lock, which wouldn't help here as we are adding a new list_dev. This also fixes a problem where we have called kzalloc(..., GFP_KERNEL) from within rcu-lock, which isn't allowed as kzalloc can sleep when called with GFP_KERNEL. With CONFIG_DEBUG_ATOMIC_SLEEP set, we get following lockdep-splat: include/linux/rcupdate.h:578 Illegal context switch in RCU read-side critical section! other info that might help us debug this: rcu_scheduler_active = 1, debug_locks = 0 5 locks held by swapper/0/1: #0: (&dev->mutex){......}, at: [<c02f68f4>] __driver_attach+0x48/0x98 #1: (&dev->mutex){......}, at: [<c02f6904>] __driver_attach+0x58/0x98 #2: (cpu_hotplug.lock){++++++}, at: [<c00249d0>] get_online_cpus+0x40/0xb0 #3: (subsys mutex#5){+.+.+.}, at: [<c02f4f8c>] subsys_interface_register+0x44/0xdc #4: (rcu_read_lock){......}, at: [<c0305c80>] dev_pm_opp_set_sharing_cpus+0x0/0x1e4 stack backtrace: CPU: 1 PID: 1 Comm: swapper/0 Tainted: G W 4.3.0-rc7-00047-g81f5932958a8 #59 Hardware name: SAMSUNG EXYNOS (Flattened Device Tree) [<c0016874>] (unwind_backtrace) from [<c001355c>] (show_stack+0x10/0x14) [<c001355c>] (show_stack) from [<c022553c>] (dump_stack+0x94/0xbc) [<c022553c>] (dump_stack) from [<c004904c>] (___might_sleep+0x24c/0x298) [<c004904c>] (___might_sleep) from [<c00f07e4>] (kmem_cache_alloc+0xe8/0x164) [<c00f07e4>] (kmem_cache_alloc) from [<c0305354>] (_add_list_dev+0x30/0x58) [<c0305354>] (_add_list_dev) from [<c0305d50>] (dev_pm_opp_set_sharing_cpus+0xd0/0x1e4) [<c0305d50>] (dev_pm_opp_set_sharing_cpus) from [<c040eda4>] (cpufreq_init+0x4cc/0x62c) [<c040eda4>] (cpufreq_init) from [<c040a964>] (cpufreq_online+0xbc/0x73c) [<c040a964>] (cpufreq_online) from [<c02f4fe0>] (subsys_interface_register+0x98/0xdc) [<c02f4fe0>] (subsys_interface_register) from [<c040a640>] (cpufreq_register_driver+0x110/0x17c) [<c040a640>] (cpufreq_register_driver) from [<c040ef64>] (dt_cpufreq_probe+0x60/0x8c) [<c040ef64>] (dt_cpufreq_probe) from [<c02f8084>] (platform_drv_probe+0x44/0xa4) [<c02f8084>] (platform_drv_probe) from [<c02f67c0>] (driver_probe_device+0x208/0x2f4) [<c02f67c0>] (driver_probe_device) from [<c02f6940>] (__driver_attach+0x94/0x98) [<c02f6940>] (__driver_attach) from [<c02f4c1c>] (bus_for_each_dev+0x68/0x9c) Reported-by: Michael Turquette <mturquette@baylibre.com> Reviewed-by: Stephen Boyd <sboyd@codeaurora.org> Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org> Cc: 4.3 <stable@vger.kernel.org> # 4.3 Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
2015-11-05 08:51:19 +00:00
extern struct list_head opp_tables, lazy_opp_tables;
/*
* Internal data structure organization with the OPP layer library is as
* follows:
* opp_tables (root)
* |- device 1 (represents voltage domain 1)
* | |- opp 1 (availability, freq, voltage)
* | |- opp 2 ..
* ... ...
* | `- opp n ..
* |- device 2 (represents the next voltage domain)
* ...
* `- device m (represents mth voltage domain)
* device 1, 2.. are represented by opp_table structure while each opp
* is represented by the opp structure.
*/
/**
* struct dev_pm_opp - Generic OPP description structure
* @node: opp table node. The nodes are maintained throughout the lifetime
* of boot. It is expected only an optimal set of OPPs are
* added to the library by the SoC framework.
* IMPORTANT: the opp nodes should be maintained in increasing
* order.
* @kref: for reference count of the OPP.
* @available: true/false - marks if this OPP as available or not
* @dynamic: not-created from static DT entries.
* @turbo: true if turbo (boost) OPP
* @suspend: true if suspend OPP
opp: Don't drop extra references to OPPs accidentally We are required to call dev_pm_opp_put() from outside of the opp_table->lock as debugfs removal needs to happen lock-less to avoid circular dependency issues. commit cf1fac943c63 ("opp: Reduce the size of critical section in _opp_kref_release()") tried to fix that introducing a new routine _opp_get_next() which keeps returning OPPs that can be freed by the callers and this routine shall be called without holding the opp_table->lock. Though the commit overlooked the fact that the OPPs can be referenced by other users as well and this routine will end up dropping references which were taken by other users and hence freeing the OPPs prematurely. In effect, other users of the OPPs will end up having invalid pointers at hand. We didn't see any crash reports earlier as the exact situation never happened, though it is certainly possible. We need a way to mark which OPPs are no longer referenced by the OPP core, so we don't drop extra references to them accidentally. This commit adds another OPP flag, "removed", which is used to track this. And now we should never end up dropping extra references to the OPPs. Cc: v5.11+ <stable@vger.kernel.org> # v5.11+ Fixes: cf1fac943c63 ("opp: Reduce the size of critical section in _opp_kref_release()") Signed-off-by: Beata Michalska <beata.michalska@arm.com> [ Viresh: Almost rewrote entire patch, added new "removed" field, rewrote commit log and added the correct Fixes tag. ] Co-developed-by: Viresh Kumar <viresh.kumar@linaro.org> Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
2021-03-04 15:07:34 +00:00
* @removed: flag indicating that OPP's reference is dropped by OPP core.
* @pstate: Device's power domain's performance state.
* @rate: Frequency in hertz
* @level: Performance level
* @supplies: Power supplies voltage/current values
* @bandwidth: Interconnect bandwidth values
* @clock_latency_ns: Latency (in nanoseconds) of switching to this OPP's
* frequency from any other OPP's frequency.
* @required_opps: List of OPPs that are required by this OPP.
* @opp_table: points back to the opp_table struct this opp belongs to
* @np: OPP's device node.
* @dentry: debugfs dentry pointer (per opp)
*
* This structure stores the OPP information for a given device.
*/
struct dev_pm_opp {
struct list_head node;
struct kref kref;
bool available;
bool dynamic;
bool turbo;
bool suspend;
opp: Don't drop extra references to OPPs accidentally We are required to call dev_pm_opp_put() from outside of the opp_table->lock as debugfs removal needs to happen lock-less to avoid circular dependency issues. commit cf1fac943c63 ("opp: Reduce the size of critical section in _opp_kref_release()") tried to fix that introducing a new routine _opp_get_next() which keeps returning OPPs that can be freed by the callers and this routine shall be called without holding the opp_table->lock. Though the commit overlooked the fact that the OPPs can be referenced by other users as well and this routine will end up dropping references which were taken by other users and hence freeing the OPPs prematurely. In effect, other users of the OPPs will end up having invalid pointers at hand. We didn't see any crash reports earlier as the exact situation never happened, though it is certainly possible. We need a way to mark which OPPs are no longer referenced by the OPP core, so we don't drop extra references to them accidentally. This commit adds another OPP flag, "removed", which is used to track this. And now we should never end up dropping extra references to the OPPs. Cc: v5.11+ <stable@vger.kernel.org> # v5.11+ Fixes: cf1fac943c63 ("opp: Reduce the size of critical section in _opp_kref_release()") Signed-off-by: Beata Michalska <beata.michalska@arm.com> [ Viresh: Almost rewrote entire patch, added new "removed" field, rewrote commit log and added the correct Fixes tag. ] Co-developed-by: Viresh Kumar <viresh.kumar@linaro.org> Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
2021-03-04 15:07:34 +00:00
bool removed;
unsigned int pstate;
unsigned long rate;
unsigned int level;
struct dev_pm_opp_supply *supplies;
struct dev_pm_opp_icc_bw *bandwidth;
unsigned long clock_latency_ns;
struct dev_pm_opp **required_opps;
struct opp_table *opp_table;
struct device_node *np;
#ifdef CONFIG_DEBUG_FS
struct dentry *dentry;
#endif
};
/**
* struct opp_device - devices managed by 'struct opp_table'
* @node: list node
* @dev: device to which the struct object belongs
* @dentry: debugfs dentry pointer (per device)
*
* This is an internal data structure maintaining the devices that are managed
* by 'struct opp_table'.
*/
struct opp_device {
struct list_head node;
const struct device *dev;
#ifdef CONFIG_DEBUG_FS
struct dentry *dentry;
#endif
};
enum opp_table_access {
OPP_TABLE_ACCESS_UNKNOWN = 0,
OPP_TABLE_ACCESS_EXCLUSIVE = 1,
OPP_TABLE_ACCESS_SHARED = 2,
};
/**
* struct opp_table - Device opp structure
* @node: table node - contains the devices with OPPs that
* have been registered. Nodes once added are not modified in this
* table.
* @head: notifier head to notify the OPP availability changes.
* @dev_list: list of devices that share these OPPs
* @opp_list: table of opps
* @kref: for reference count of the table.
* @lock: mutex protecting the opp_list and dev_list.
* @np: struct device_node pointer for opp's DT node.
* @clock_latency_ns_max: Max clock latency in nanoseconds.
* @parsed_static_opps: Count of devices for which OPPs are initialized from DT.
* @shared_opp: OPP is shared between multiple devices.
* @current_rate: Currently configured frequency.
* @current_opp: Currently configured OPP for the table.
* @suspend_opp: Pointer to OPP to be used during device suspend.
OPP: Add dev_pm_opp_{set|put}_genpd_virt_dev() helper Multiple generic power domains for a consumer device are supported with the help of virtual devices, which are created for each consumer device - genpd pair. These are the device structures which are attached to the power domain and are required by the OPP core to set the performance state of the genpd. The helpers added by this commit are required to be called once for each of these virtual devices. These are required only if multiple domains are available for a device, otherwise the actual device structure will be used instead by the OPP core. The new helpers also support the complex cases where the consumer device wouldn't always require all the domains. For example, a camera may require only one power domain during normal operations but two during high resolution operations. The consumer driver can call dev_pm_opp_put_genpd_virt_dev(high_resolution_genpd_virt_dev) if it is currently operating in the normal mode and doesn't have any performance requirements from the genpd which manages high resolution power requirements. The consumer driver can later call dev_pm_opp_set_genpd_virt_dev(high_resolution_genpd_virt_dev) once it switches back to the high resolution mode. The new helpers differ from other OPP set/put helpers as the new ones can be called with OPPs initialized for the table as we may need to call them on the fly because of the complex case explained above. For this reason it is possible that the genpd virt_dev structure may be used in parallel while the new helpers are running and a new mutex is added to protect against that. We didn't use the existing opp_table->lock mutex as that is widely used in the OPP core and we will need this lock in the dev_pm_opp_set_rate() helper while changing OPP and we need to make sure there is not much contention while doing that as that's the hotpath. Reviewed-by: Ulf Hansson <ulf.hansson@linaro.org> Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
2018-06-26 10:59:34 +00:00
* @genpd_virt_dev_lock: Mutex protecting the genpd virtual device pointers.
* @genpd_virt_devs: List of virtual devices for multiple genpd support.
* @required_opp_tables: List of device OPP tables that are required by OPPs in
* this table.
* @required_opp_count: Number of required devices.
* @supported_hw: Array of version number to support.
* @supported_hw_count: Number of elements in supported_hw array.
* @prop_name: A name to postfix to many DT properties, while parsing them.
* @clk: Device's clock handle
* @regulators: Supply regulators
* @regulator_count: Number of power supply regulators. Its value can be -1
* (uninitialized), 0 (no opp-microvolt property) or > 0 (has opp-microvolt
* property).
* @paths: Interconnect path handles
* @path_count: Number of interconnect paths
* @enabled: Set to true if the device's resources are enabled/configured.
* @genpd_performance_state: Device's power domain support performance state.
* @is_genpd: Marks if the OPP table belongs to a genpd.
* @set_opp: Platform specific set_opp callback
* @sod_supplies: Set opp data supplies
* @set_opp_data: Data to be passed to set_opp callback
* @dentry: debugfs dentry pointer of the real device directory (not links).
* @dentry_name: Name of the real dentry.
*
* @voltage_tolerance_v1: In percentage, for v1 bindings only.
*
* This is an internal data structure maintaining the link to opps attached to
* a device. This structure is not meant to be shared to users as it is
* meant for book keeping and private to OPP library.
*/
struct opp_table {
struct list_head node, lazy;
struct blocking_notifier_head head;
struct list_head dev_list;
struct list_head opp_list;
struct kref kref;
struct mutex lock;
struct device_node *np;
unsigned long clock_latency_ns_max;
/* For backward compatibility with v1 bindings */
unsigned int voltage_tolerance_v1;
unsigned int parsed_static_opps;
enum opp_table_access shared_opp;
unsigned long current_rate;
struct dev_pm_opp *current_opp;
struct dev_pm_opp *suspend_opp;
OPP: Add dev_pm_opp_{set|put}_genpd_virt_dev() helper Multiple generic power domains for a consumer device are supported with the help of virtual devices, which are created for each consumer device - genpd pair. These are the device structures which are attached to the power domain and are required by the OPP core to set the performance state of the genpd. The helpers added by this commit are required to be called once for each of these virtual devices. These are required only if multiple domains are available for a device, otherwise the actual device structure will be used instead by the OPP core. The new helpers also support the complex cases where the consumer device wouldn't always require all the domains. For example, a camera may require only one power domain during normal operations but two during high resolution operations. The consumer driver can call dev_pm_opp_put_genpd_virt_dev(high_resolution_genpd_virt_dev) if it is currently operating in the normal mode and doesn't have any performance requirements from the genpd which manages high resolution power requirements. The consumer driver can later call dev_pm_opp_set_genpd_virt_dev(high_resolution_genpd_virt_dev) once it switches back to the high resolution mode. The new helpers differ from other OPP set/put helpers as the new ones can be called with OPPs initialized for the table as we may need to call them on the fly because of the complex case explained above. For this reason it is possible that the genpd virt_dev structure may be used in parallel while the new helpers are running and a new mutex is added to protect against that. We didn't use the existing opp_table->lock mutex as that is widely used in the OPP core and we will need this lock in the dev_pm_opp_set_rate() helper while changing OPP and we need to make sure there is not much contention while doing that as that's the hotpath. Reviewed-by: Ulf Hansson <ulf.hansson@linaro.org> Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
2018-06-26 10:59:34 +00:00
struct mutex genpd_virt_dev_lock;
struct device **genpd_virt_devs;
struct opp_table **required_opp_tables;
unsigned int required_opp_count;
unsigned int *supported_hw;
unsigned int supported_hw_count;
const char *prop_name;
struct clk *clk;
struct regulator **regulators;
int regulator_count;
struct icc_path **paths;
unsigned int path_count;
bool enabled;
bool genpd_performance_state;
bool is_genpd;
int (*set_opp)(struct dev_pm_set_opp_data *data);
struct dev_pm_opp_supply *sod_supplies;
struct dev_pm_set_opp_data *set_opp_data;
#ifdef CONFIG_DEBUG_FS
struct dentry *dentry;
char dentry_name[NAME_MAX];
#endif
};
/* Routines internal to opp core */
void dev_pm_opp_get(struct dev_pm_opp *opp);
bool _opp_remove_all_static(struct opp_table *opp_table);
void _get_opp_table_kref(struct opp_table *opp_table);
int _get_opp_count(struct opp_table *opp_table);
struct opp_table *_find_opp_table(struct device *dev);
struct opp_device *_add_opp_dev(const struct device *dev, struct opp_table *opp_table);
struct dev_pm_opp *_opp_allocate(struct opp_table *opp_table);
void _opp_free(struct dev_pm_opp *opp);
int _opp_compare_key(struct dev_pm_opp *opp1, struct dev_pm_opp *opp2);
int _opp_add(struct device *dev, struct dev_pm_opp *new_opp, struct opp_table *opp_table, bool rate_not_available);
int _opp_add_v1(struct opp_table *opp_table, struct device *dev, unsigned long freq, long u_volt, bool dynamic);
void _dev_pm_opp_cpumask_remove_table(const struct cpumask *cpumask, int last_cpu);
struct opp_table *_add_opp_table_indexed(struct device *dev, int index, bool getclk);
void _put_opp_list_kref(struct opp_table *opp_table);
void _required_opps_available(struct dev_pm_opp *opp, int count);
static inline bool lazy_linking_pending(struct opp_table *opp_table)
{
return unlikely(!list_empty(&opp_table->lazy));
}
#ifdef CONFIG_OF
void _of_init_opp_table(struct opp_table *opp_table, struct device *dev, int index);
void _of_clear_opp_table(struct opp_table *opp_table);
struct opp_table *_managed_opp(struct device *dev, int index);
void _of_opp_free_required_opps(struct opp_table *opp_table,
struct dev_pm_opp *opp);
#else
static inline void _of_init_opp_table(struct opp_table *opp_table, struct device *dev, int index) {}
static inline void _of_clear_opp_table(struct opp_table *opp_table) {}
static inline struct opp_table *_managed_opp(struct device *dev, int index) { return NULL; }
static inline void _of_opp_free_required_opps(struct opp_table *opp_table,
struct dev_pm_opp *opp) {}
#endif
#ifdef CONFIG_DEBUG_FS
void opp_debug_remove_one(struct dev_pm_opp *opp);
void opp_debug_create_one(struct dev_pm_opp *opp, struct opp_table *opp_table);
void opp_debug_register(struct opp_device *opp_dev, struct opp_table *opp_table);
void opp_debug_unregister(struct opp_device *opp_dev, struct opp_table *opp_table);
#else
static inline void opp_debug_remove_one(struct dev_pm_opp *opp) {}
static inline void opp_debug_create_one(struct dev_pm_opp *opp,
struct opp_table *opp_table) { }
static inline void opp_debug_register(struct opp_device *opp_dev,
struct opp_table *opp_table) { }
static inline void opp_debug_unregister(struct opp_device *opp_dev,
struct opp_table *opp_table)
{ }
#endif /* DEBUG_FS */
#endif /* __DRIVER_OPP_H__ */