WSL2-Linux-Kernel/drivers/pinctrl/pinconf.c

821 строка
20 KiB
C
Исходник Обычный вид История

pinctrl: add a pin config interface This add per-pin and per-group pin config interfaces for biasing, driving and other such electronic properties. The details of passed configurations are passed in an opaque unsigned long which may be dereferences to integer types, structs or lists on either side of the configuration interface. ChangeLog v1->v2: - Clear split of terminology: we now have pin controllers, and those may support two interfaces using vtables: pin multiplexing and pin configuration. - Break out pin configuration to its own C file, controllers may implement only config without mux, and vice versa, so keep each sub-functionality of pin controllers separate. Introduce CONFIG_PINCONF in Kconfig. - Implement some core logic around pin configuration in the pinconf.c file. - Remove UNKNOWN config states, these were just surplus baggage. - Remove FLOAT config state - HIGH_IMPEDANCE should be enough for everyone. - PIN_CONFIG_POWER_SOURCE added to handle switching the power supply for the pin logic between different sources - Explicit DISABLE config enums to turn schmitt-trigger, wakeup etc OFF. - Update documentation to reflect all the recent reasoning. ChangeLog v2->v3: - Twist API around to pass around arrays of config tuples instead of (param, value) pairs everywhere. - Explicit drive strength semantics for push/pull and similar drive modes, this shall be the number of drive stages vs nominal load impedance, which should match the actual electronics used in push/pull CMOS or TTY totempoles. - Drop load capacitance configuration - I probably don't know what I'm doing here so leave it out. - Drop PIN_CONFIG_INPUT_SCHMITT_OFF, instead the argument zero to PIN_CONFIG_INPUT_SCHMITT turns schmitt trigger off. - Drop PIN_CONFIG_NORMAL_POWER_MODE and have a well defined argument to PIN_CONFIG_LOW_POWER_MODE to get out of it instead. - Drop PIN_CONFIG_WAKEUP_ENABLE/DISABLE and just use PIN_CONFIG_WAKEUP with defined value zero to turn wakeup off. - Add PIN_CONFIG_INPUT_DEBOUNCE for configuring debounce time on input lines. - Fix a bug when we tried to configure pins for pin controllers without pinconf support. - Initialized debugfs properly so it works. - Initialize the mutex properly and lock around config tampering sections. - Check the return value from get_initial_config() properly. ChangeLog v3->v4: - Export the pin_config_get(), pin_config_set() and pin_config_group() functions. - Drop the entire concept of just getting initial config and keeping track of pin states internally, instead ask the pins what state they are in. Previous idea was plain wrong, if the device cannot keep track of its state, the driver should do it. - Drop the generic configuration layout, it seems this impose too much restriction on some pin controllers, so let them do things the way they want and split off support for generic config as an optional add-on. ChangeLog v4->v5: - Introduce two symmetric driver calls for group configuration, .pin_config_group_[get|set] and corresponding external calls. - Remove generic semantic meanings of return values from config calls, these belong in the generic config patch. Just pass the return value through instead. - Add a debugfs entry "pinconf-groups" to read status from group configuration only, also slam in a per-group debug callback in the pinconf_ops so custom drivers can display something meaningful for their pins. - Fix some dangling newline. - Drop dangling #else clause. - Update documentation to match the above. ChangeLog v5->v6: - Change to using a pin name as parameter for the [get|set]_config() functions, as suggested by Stephen Warren. This is more natural as names will be what a developer has access to in written documentation etc. ChangeLog v6->v7: - Refactor out by-pin and by-name get/set functions, only expose the by-name functions externally, expose the by-pin functions internally. - Show supported pin control functionality in the debugfs pinctrl-devices file. Acked-by: Stephen Warren <swarren@nvidia.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2011-10-19 20:14:33 +04:00
/*
* Core driver for the pin config portions of the pin control subsystem
*
* Copyright (C) 2011 ST-Ericsson SA
* Written on behalf of Linaro for ST-Ericsson
*
* Author: Linus Walleij <linus.walleij@linaro.org>
*
* License terms: GNU General Public License (GPL) version 2
*/
#define pr_fmt(fmt) "pinconfig core: " fmt
#include <linux/kernel.h>
#include <linux/module.h>
#include <linux/init.h>
#include <linux/device.h>
#include <linux/slab.h>
#include <linux/debugfs.h>
#include <linux/seq_file.h>
#include <linux/uaccess.h>
pinctrl: add a pin config interface This add per-pin and per-group pin config interfaces for biasing, driving and other such electronic properties. The details of passed configurations are passed in an opaque unsigned long which may be dereferences to integer types, structs or lists on either side of the configuration interface. ChangeLog v1->v2: - Clear split of terminology: we now have pin controllers, and those may support two interfaces using vtables: pin multiplexing and pin configuration. - Break out pin configuration to its own C file, controllers may implement only config without mux, and vice versa, so keep each sub-functionality of pin controllers separate. Introduce CONFIG_PINCONF in Kconfig. - Implement some core logic around pin configuration in the pinconf.c file. - Remove UNKNOWN config states, these were just surplus baggage. - Remove FLOAT config state - HIGH_IMPEDANCE should be enough for everyone. - PIN_CONFIG_POWER_SOURCE added to handle switching the power supply for the pin logic between different sources - Explicit DISABLE config enums to turn schmitt-trigger, wakeup etc OFF. - Update documentation to reflect all the recent reasoning. ChangeLog v2->v3: - Twist API around to pass around arrays of config tuples instead of (param, value) pairs everywhere. - Explicit drive strength semantics for push/pull and similar drive modes, this shall be the number of drive stages vs nominal load impedance, which should match the actual electronics used in push/pull CMOS or TTY totempoles. - Drop load capacitance configuration - I probably don't know what I'm doing here so leave it out. - Drop PIN_CONFIG_INPUT_SCHMITT_OFF, instead the argument zero to PIN_CONFIG_INPUT_SCHMITT turns schmitt trigger off. - Drop PIN_CONFIG_NORMAL_POWER_MODE and have a well defined argument to PIN_CONFIG_LOW_POWER_MODE to get out of it instead. - Drop PIN_CONFIG_WAKEUP_ENABLE/DISABLE and just use PIN_CONFIG_WAKEUP with defined value zero to turn wakeup off. - Add PIN_CONFIG_INPUT_DEBOUNCE for configuring debounce time on input lines. - Fix a bug when we tried to configure pins for pin controllers without pinconf support. - Initialized debugfs properly so it works. - Initialize the mutex properly and lock around config tampering sections. - Check the return value from get_initial_config() properly. ChangeLog v3->v4: - Export the pin_config_get(), pin_config_set() and pin_config_group() functions. - Drop the entire concept of just getting initial config and keeping track of pin states internally, instead ask the pins what state they are in. Previous idea was plain wrong, if the device cannot keep track of its state, the driver should do it. - Drop the generic configuration layout, it seems this impose too much restriction on some pin controllers, so let them do things the way they want and split off support for generic config as an optional add-on. ChangeLog v4->v5: - Introduce two symmetric driver calls for group configuration, .pin_config_group_[get|set] and corresponding external calls. - Remove generic semantic meanings of return values from config calls, these belong in the generic config patch. Just pass the return value through instead. - Add a debugfs entry "pinconf-groups" to read status from group configuration only, also slam in a per-group debug callback in the pinconf_ops so custom drivers can display something meaningful for their pins. - Fix some dangling newline. - Drop dangling #else clause. - Update documentation to match the above. ChangeLog v5->v6: - Change to using a pin name as parameter for the [get|set]_config() functions, as suggested by Stephen Warren. This is more natural as names will be what a developer has access to in written documentation etc. ChangeLog v6->v7: - Refactor out by-pin and by-name get/set functions, only expose the by-name functions externally, expose the by-pin functions internally. - Show supported pin control functionality in the debugfs pinctrl-devices file. Acked-by: Stephen Warren <swarren@nvidia.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2011-10-19 20:14:33 +04:00
#include <linux/pinctrl/machine.h>
#include <linux/pinctrl/pinctrl.h>
#include <linux/pinctrl/pinconf.h>
#include "core.h"
#include "pinconf.h"
int pinconf_check_ops(struct pinctrl_dev *pctldev)
{
const struct pinconf_ops *ops = pctldev->desc->confops;
/* We must be able to read out pin status */
if (!ops->pin_config_get && !ops->pin_config_group_get) {
dev_err(pctldev->dev,
"pinconf must be able to read out pin status\n");
return -EINVAL;
}
/* We have to be able to config the pins in SOME way */
if (!ops->pin_config_set && !ops->pin_config_group_set) {
dev_err(pctldev->dev,
"pinconf has to be able to set a pins config\n");
return -EINVAL;
}
return 0;
}
int pinconf_validate_map(struct pinctrl_map const *map, int i)
{
if (!map->data.configs.group_or_pin) {
pr_err("failed to register map %s (%d): no group/pin given\n",
map->name, i);
return -EINVAL;
}
if (!map->data.configs.num_configs ||
!map->data.configs.configs) {
pr_err("failed to register map %s (%d): no configs given\n",
map->name, i);
return -EINVAL;
}
return 0;
}
int pin_config_get_for_pin(struct pinctrl_dev *pctldev, unsigned pin,
pinctrl: add a pin config interface This add per-pin and per-group pin config interfaces for biasing, driving and other such electronic properties. The details of passed configurations are passed in an opaque unsigned long which may be dereferences to integer types, structs or lists on either side of the configuration interface. ChangeLog v1->v2: - Clear split of terminology: we now have pin controllers, and those may support two interfaces using vtables: pin multiplexing and pin configuration. - Break out pin configuration to its own C file, controllers may implement only config without mux, and vice versa, so keep each sub-functionality of pin controllers separate. Introduce CONFIG_PINCONF in Kconfig. - Implement some core logic around pin configuration in the pinconf.c file. - Remove UNKNOWN config states, these were just surplus baggage. - Remove FLOAT config state - HIGH_IMPEDANCE should be enough for everyone. - PIN_CONFIG_POWER_SOURCE added to handle switching the power supply for the pin logic between different sources - Explicit DISABLE config enums to turn schmitt-trigger, wakeup etc OFF. - Update documentation to reflect all the recent reasoning. ChangeLog v2->v3: - Twist API around to pass around arrays of config tuples instead of (param, value) pairs everywhere. - Explicit drive strength semantics for push/pull and similar drive modes, this shall be the number of drive stages vs nominal load impedance, which should match the actual electronics used in push/pull CMOS or TTY totempoles. - Drop load capacitance configuration - I probably don't know what I'm doing here so leave it out. - Drop PIN_CONFIG_INPUT_SCHMITT_OFF, instead the argument zero to PIN_CONFIG_INPUT_SCHMITT turns schmitt trigger off. - Drop PIN_CONFIG_NORMAL_POWER_MODE and have a well defined argument to PIN_CONFIG_LOW_POWER_MODE to get out of it instead. - Drop PIN_CONFIG_WAKEUP_ENABLE/DISABLE and just use PIN_CONFIG_WAKEUP with defined value zero to turn wakeup off. - Add PIN_CONFIG_INPUT_DEBOUNCE for configuring debounce time on input lines. - Fix a bug when we tried to configure pins for pin controllers without pinconf support. - Initialized debugfs properly so it works. - Initialize the mutex properly and lock around config tampering sections. - Check the return value from get_initial_config() properly. ChangeLog v3->v4: - Export the pin_config_get(), pin_config_set() and pin_config_group() functions. - Drop the entire concept of just getting initial config and keeping track of pin states internally, instead ask the pins what state they are in. Previous idea was plain wrong, if the device cannot keep track of its state, the driver should do it. - Drop the generic configuration layout, it seems this impose too much restriction on some pin controllers, so let them do things the way they want and split off support for generic config as an optional add-on. ChangeLog v4->v5: - Introduce two symmetric driver calls for group configuration, .pin_config_group_[get|set] and corresponding external calls. - Remove generic semantic meanings of return values from config calls, these belong in the generic config patch. Just pass the return value through instead. - Add a debugfs entry "pinconf-groups" to read status from group configuration only, also slam in a per-group debug callback in the pinconf_ops so custom drivers can display something meaningful for their pins. - Fix some dangling newline. - Drop dangling #else clause. - Update documentation to match the above. ChangeLog v5->v6: - Change to using a pin name as parameter for the [get|set]_config() functions, as suggested by Stephen Warren. This is more natural as names will be what a developer has access to in written documentation etc. ChangeLog v6->v7: - Refactor out by-pin and by-name get/set functions, only expose the by-name functions externally, expose the by-pin functions internally. - Show supported pin control functionality in the debugfs pinctrl-devices file. Acked-by: Stephen Warren <swarren@nvidia.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2011-10-19 20:14:33 +04:00
unsigned long *config)
{
const struct pinconf_ops *ops = pctldev->desc->confops;
if (!ops || !ops->pin_config_get) {
dev_err(pctldev->dev, "cannot get pin configuration, missing "
pinctrl: add a pin config interface This add per-pin and per-group pin config interfaces for biasing, driving and other such electronic properties. The details of passed configurations are passed in an opaque unsigned long which may be dereferences to integer types, structs or lists on either side of the configuration interface. ChangeLog v1->v2: - Clear split of terminology: we now have pin controllers, and those may support two interfaces using vtables: pin multiplexing and pin configuration. - Break out pin configuration to its own C file, controllers may implement only config without mux, and vice versa, so keep each sub-functionality of pin controllers separate. Introduce CONFIG_PINCONF in Kconfig. - Implement some core logic around pin configuration in the pinconf.c file. - Remove UNKNOWN config states, these were just surplus baggage. - Remove FLOAT config state - HIGH_IMPEDANCE should be enough for everyone. - PIN_CONFIG_POWER_SOURCE added to handle switching the power supply for the pin logic between different sources - Explicit DISABLE config enums to turn schmitt-trigger, wakeup etc OFF. - Update documentation to reflect all the recent reasoning. ChangeLog v2->v3: - Twist API around to pass around arrays of config tuples instead of (param, value) pairs everywhere. - Explicit drive strength semantics for push/pull and similar drive modes, this shall be the number of drive stages vs nominal load impedance, which should match the actual electronics used in push/pull CMOS or TTY totempoles. - Drop load capacitance configuration - I probably don't know what I'm doing here so leave it out. - Drop PIN_CONFIG_INPUT_SCHMITT_OFF, instead the argument zero to PIN_CONFIG_INPUT_SCHMITT turns schmitt trigger off. - Drop PIN_CONFIG_NORMAL_POWER_MODE and have a well defined argument to PIN_CONFIG_LOW_POWER_MODE to get out of it instead. - Drop PIN_CONFIG_WAKEUP_ENABLE/DISABLE and just use PIN_CONFIG_WAKEUP with defined value zero to turn wakeup off. - Add PIN_CONFIG_INPUT_DEBOUNCE for configuring debounce time on input lines. - Fix a bug when we tried to configure pins for pin controllers without pinconf support. - Initialized debugfs properly so it works. - Initialize the mutex properly and lock around config tampering sections. - Check the return value from get_initial_config() properly. ChangeLog v3->v4: - Export the pin_config_get(), pin_config_set() and pin_config_group() functions. - Drop the entire concept of just getting initial config and keeping track of pin states internally, instead ask the pins what state they are in. Previous idea was plain wrong, if the device cannot keep track of its state, the driver should do it. - Drop the generic configuration layout, it seems this impose too much restriction on some pin controllers, so let them do things the way they want and split off support for generic config as an optional add-on. ChangeLog v4->v5: - Introduce two symmetric driver calls for group configuration, .pin_config_group_[get|set] and corresponding external calls. - Remove generic semantic meanings of return values from config calls, these belong in the generic config patch. Just pass the return value through instead. - Add a debugfs entry "pinconf-groups" to read status from group configuration only, also slam in a per-group debug callback in the pinconf_ops so custom drivers can display something meaningful for their pins. - Fix some dangling newline. - Drop dangling #else clause. - Update documentation to match the above. ChangeLog v5->v6: - Change to using a pin name as parameter for the [get|set]_config() functions, as suggested by Stephen Warren. This is more natural as names will be what a developer has access to in written documentation etc. ChangeLog v6->v7: - Refactor out by-pin and by-name get/set functions, only expose the by-name functions externally, expose the by-pin functions internally. - Show supported pin control functionality in the debugfs pinctrl-devices file. Acked-by: Stephen Warren <swarren@nvidia.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2011-10-19 20:14:33 +04:00
"pin_config_get() function in driver\n");
return -EINVAL;
}
return ops->pin_config_get(pctldev, pin, config);
}
/**
* pin_config_get() - get the configuration of a single pin parameter
* @dev_name: name of the pin controller device for this pin
pinctrl: add a pin config interface This add per-pin and per-group pin config interfaces for biasing, driving and other such electronic properties. The details of passed configurations are passed in an opaque unsigned long which may be dereferences to integer types, structs or lists on either side of the configuration interface. ChangeLog v1->v2: - Clear split of terminology: we now have pin controllers, and those may support two interfaces using vtables: pin multiplexing and pin configuration. - Break out pin configuration to its own C file, controllers may implement only config without mux, and vice versa, so keep each sub-functionality of pin controllers separate. Introduce CONFIG_PINCONF in Kconfig. - Implement some core logic around pin configuration in the pinconf.c file. - Remove UNKNOWN config states, these were just surplus baggage. - Remove FLOAT config state - HIGH_IMPEDANCE should be enough for everyone. - PIN_CONFIG_POWER_SOURCE added to handle switching the power supply for the pin logic between different sources - Explicit DISABLE config enums to turn schmitt-trigger, wakeup etc OFF. - Update documentation to reflect all the recent reasoning. ChangeLog v2->v3: - Twist API around to pass around arrays of config tuples instead of (param, value) pairs everywhere. - Explicit drive strength semantics for push/pull and similar drive modes, this shall be the number of drive stages vs nominal load impedance, which should match the actual electronics used in push/pull CMOS or TTY totempoles. - Drop load capacitance configuration - I probably don't know what I'm doing here so leave it out. - Drop PIN_CONFIG_INPUT_SCHMITT_OFF, instead the argument zero to PIN_CONFIG_INPUT_SCHMITT turns schmitt trigger off. - Drop PIN_CONFIG_NORMAL_POWER_MODE and have a well defined argument to PIN_CONFIG_LOW_POWER_MODE to get out of it instead. - Drop PIN_CONFIG_WAKEUP_ENABLE/DISABLE and just use PIN_CONFIG_WAKEUP with defined value zero to turn wakeup off. - Add PIN_CONFIG_INPUT_DEBOUNCE for configuring debounce time on input lines. - Fix a bug when we tried to configure pins for pin controllers without pinconf support. - Initialized debugfs properly so it works. - Initialize the mutex properly and lock around config tampering sections. - Check the return value from get_initial_config() properly. ChangeLog v3->v4: - Export the pin_config_get(), pin_config_set() and pin_config_group() functions. - Drop the entire concept of just getting initial config and keeping track of pin states internally, instead ask the pins what state they are in. Previous idea was plain wrong, if the device cannot keep track of its state, the driver should do it. - Drop the generic configuration layout, it seems this impose too much restriction on some pin controllers, so let them do things the way they want and split off support for generic config as an optional add-on. ChangeLog v4->v5: - Introduce two symmetric driver calls for group configuration, .pin_config_group_[get|set] and corresponding external calls. - Remove generic semantic meanings of return values from config calls, these belong in the generic config patch. Just pass the return value through instead. - Add a debugfs entry "pinconf-groups" to read status from group configuration only, also slam in a per-group debug callback in the pinconf_ops so custom drivers can display something meaningful for their pins. - Fix some dangling newline. - Drop dangling #else clause. - Update documentation to match the above. ChangeLog v5->v6: - Change to using a pin name as parameter for the [get|set]_config() functions, as suggested by Stephen Warren. This is more natural as names will be what a developer has access to in written documentation etc. ChangeLog v6->v7: - Refactor out by-pin and by-name get/set functions, only expose the by-name functions externally, expose the by-pin functions internally. - Show supported pin control functionality in the debugfs pinctrl-devices file. Acked-by: Stephen Warren <swarren@nvidia.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2011-10-19 20:14:33 +04:00
* @name: name of the pin to get the config for
* @config: the config pointed to by this argument will be filled in with the
* current pin state, it can be used directly by drivers as a numeral, or
* it can be dereferenced to any struct.
*/
int pin_config_get(const char *dev_name, const char *name,
pinctrl: add a pin config interface This add per-pin and per-group pin config interfaces for biasing, driving and other such electronic properties. The details of passed configurations are passed in an opaque unsigned long which may be dereferences to integer types, structs or lists on either side of the configuration interface. ChangeLog v1->v2: - Clear split of terminology: we now have pin controllers, and those may support two interfaces using vtables: pin multiplexing and pin configuration. - Break out pin configuration to its own C file, controllers may implement only config without mux, and vice versa, so keep each sub-functionality of pin controllers separate. Introduce CONFIG_PINCONF in Kconfig. - Implement some core logic around pin configuration in the pinconf.c file. - Remove UNKNOWN config states, these were just surplus baggage. - Remove FLOAT config state - HIGH_IMPEDANCE should be enough for everyone. - PIN_CONFIG_POWER_SOURCE added to handle switching the power supply for the pin logic between different sources - Explicit DISABLE config enums to turn schmitt-trigger, wakeup etc OFF. - Update documentation to reflect all the recent reasoning. ChangeLog v2->v3: - Twist API around to pass around arrays of config tuples instead of (param, value) pairs everywhere. - Explicit drive strength semantics for push/pull and similar drive modes, this shall be the number of drive stages vs nominal load impedance, which should match the actual electronics used in push/pull CMOS or TTY totempoles. - Drop load capacitance configuration - I probably don't know what I'm doing here so leave it out. - Drop PIN_CONFIG_INPUT_SCHMITT_OFF, instead the argument zero to PIN_CONFIG_INPUT_SCHMITT turns schmitt trigger off. - Drop PIN_CONFIG_NORMAL_POWER_MODE and have a well defined argument to PIN_CONFIG_LOW_POWER_MODE to get out of it instead. - Drop PIN_CONFIG_WAKEUP_ENABLE/DISABLE and just use PIN_CONFIG_WAKEUP with defined value zero to turn wakeup off. - Add PIN_CONFIG_INPUT_DEBOUNCE for configuring debounce time on input lines. - Fix a bug when we tried to configure pins for pin controllers without pinconf support. - Initialized debugfs properly so it works. - Initialize the mutex properly and lock around config tampering sections. - Check the return value from get_initial_config() properly. ChangeLog v3->v4: - Export the pin_config_get(), pin_config_set() and pin_config_group() functions. - Drop the entire concept of just getting initial config and keeping track of pin states internally, instead ask the pins what state they are in. Previous idea was plain wrong, if the device cannot keep track of its state, the driver should do it. - Drop the generic configuration layout, it seems this impose too much restriction on some pin controllers, so let them do things the way they want and split off support for generic config as an optional add-on. ChangeLog v4->v5: - Introduce two symmetric driver calls for group configuration, .pin_config_group_[get|set] and corresponding external calls. - Remove generic semantic meanings of return values from config calls, these belong in the generic config patch. Just pass the return value through instead. - Add a debugfs entry "pinconf-groups" to read status from group configuration only, also slam in a per-group debug callback in the pinconf_ops so custom drivers can display something meaningful for their pins. - Fix some dangling newline. - Drop dangling #else clause. - Update documentation to match the above. ChangeLog v5->v6: - Change to using a pin name as parameter for the [get|set]_config() functions, as suggested by Stephen Warren. This is more natural as names will be what a developer has access to in written documentation etc. ChangeLog v6->v7: - Refactor out by-pin and by-name get/set functions, only expose the by-name functions externally, expose the by-pin functions internally. - Show supported pin control functionality in the debugfs pinctrl-devices file. Acked-by: Stephen Warren <swarren@nvidia.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2011-10-19 20:14:33 +04:00
unsigned long *config)
{
struct pinctrl_dev *pctldev;
pinctrl: add a pin config interface This add per-pin and per-group pin config interfaces for biasing, driving and other such electronic properties. The details of passed configurations are passed in an opaque unsigned long which may be dereferences to integer types, structs or lists on either side of the configuration interface. ChangeLog v1->v2: - Clear split of terminology: we now have pin controllers, and those may support two interfaces using vtables: pin multiplexing and pin configuration. - Break out pin configuration to its own C file, controllers may implement only config without mux, and vice versa, so keep each sub-functionality of pin controllers separate. Introduce CONFIG_PINCONF in Kconfig. - Implement some core logic around pin configuration in the pinconf.c file. - Remove UNKNOWN config states, these were just surplus baggage. - Remove FLOAT config state - HIGH_IMPEDANCE should be enough for everyone. - PIN_CONFIG_POWER_SOURCE added to handle switching the power supply for the pin logic between different sources - Explicit DISABLE config enums to turn schmitt-trigger, wakeup etc OFF. - Update documentation to reflect all the recent reasoning. ChangeLog v2->v3: - Twist API around to pass around arrays of config tuples instead of (param, value) pairs everywhere. - Explicit drive strength semantics for push/pull and similar drive modes, this shall be the number of drive stages vs nominal load impedance, which should match the actual electronics used in push/pull CMOS or TTY totempoles. - Drop load capacitance configuration - I probably don't know what I'm doing here so leave it out. - Drop PIN_CONFIG_INPUT_SCHMITT_OFF, instead the argument zero to PIN_CONFIG_INPUT_SCHMITT turns schmitt trigger off. - Drop PIN_CONFIG_NORMAL_POWER_MODE and have a well defined argument to PIN_CONFIG_LOW_POWER_MODE to get out of it instead. - Drop PIN_CONFIG_WAKEUP_ENABLE/DISABLE and just use PIN_CONFIG_WAKEUP with defined value zero to turn wakeup off. - Add PIN_CONFIG_INPUT_DEBOUNCE for configuring debounce time on input lines. - Fix a bug when we tried to configure pins for pin controllers without pinconf support. - Initialized debugfs properly so it works. - Initialize the mutex properly and lock around config tampering sections. - Check the return value from get_initial_config() properly. ChangeLog v3->v4: - Export the pin_config_get(), pin_config_set() and pin_config_group() functions. - Drop the entire concept of just getting initial config and keeping track of pin states internally, instead ask the pins what state they are in. Previous idea was plain wrong, if the device cannot keep track of its state, the driver should do it. - Drop the generic configuration layout, it seems this impose too much restriction on some pin controllers, so let them do things the way they want and split off support for generic config as an optional add-on. ChangeLog v4->v5: - Introduce two symmetric driver calls for group configuration, .pin_config_group_[get|set] and corresponding external calls. - Remove generic semantic meanings of return values from config calls, these belong in the generic config patch. Just pass the return value through instead. - Add a debugfs entry "pinconf-groups" to read status from group configuration only, also slam in a per-group debug callback in the pinconf_ops so custom drivers can display something meaningful for their pins. - Fix some dangling newline. - Drop dangling #else clause. - Update documentation to match the above. ChangeLog v5->v6: - Change to using a pin name as parameter for the [get|set]_config() functions, as suggested by Stephen Warren. This is more natural as names will be what a developer has access to in written documentation etc. ChangeLog v6->v7: - Refactor out by-pin and by-name get/set functions, only expose the by-name functions externally, expose the by-pin functions internally. - Show supported pin control functionality in the debugfs pinctrl-devices file. Acked-by: Stephen Warren <swarren@nvidia.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2011-10-19 20:14:33 +04:00
int pin;
pctldev = get_pinctrl_dev_from_devname(dev_name);
pinctrl: fix and simplify locking There are many problems with the current pinctrl locking: struct pinctrl_dev's gpio_ranges_lock isn't effective; pinctrl_match_gpio_range() only holds this lock while searching for a gpio range, but the found range is return and manipulated after releading the lock. This could allow pinctrl_remove_gpio_range() for that range while it is in use, and the caller may very well delete the range after removing it, causing pinctrl code to touch the now-free range object. Solving this requires the introduction of a higher-level lock, at least a lock per pin controller, which both gpio range registration and pinctrl_get()/put() will acquire. There is missing locking on HW programming; pin controllers may pack the configuration for different pins/groups/config options/... into one register, and hence have to read-modify-write the register. This needs to be protected, but currently isn't. Related, a future change will add a "complete" op to the pin controller drivers, the idea being that each state's programming will be programmed into the pinctrl driver followed by the "complete" call, which may e.g. flush a register cache to HW. For this to work, it must not be possible to interleave the pinctrl driver calls for different devices. As above, solving this requires the introduction of a higher-level lock, at least a lock per pin controller, which will be held for the duration of any pinctrl_enable()/disable() call. However, each pinctrl mapping table entry may affect a different pin controller if necessary. Hence, with a per-pin-controller lock, almost any pinctrl API may need to acquire multiple locks, one per controller. To avoid deadlock, these would need to be acquired in the same order in all cases. This is extremely difficult to implement in the case of pinctrl_get(), which doesn't know which pin controllers to lock until it has parsed the entire mapping table, since it contains somewhat arbitrary data. The simplest solution here is to introduce a single lock that covers all pin controllers at once. This will be acquired by all pinctrl APIs. This then makes struct pinctrl's mutex irrelevant, since that single lock will always be held whenever this mutex is currently held. Signed-off-by: Stephen Warren <swarren@nvidia.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2012-03-03 00:05:44 +04:00
if (!pctldev) {
pin = -EINVAL;
pinctrl: move subsystem mutex to pinctrl_dev struct This mutex avoids deadlock in case of use of multiple pin controllers. Before this modification, by using a global mutex, deadlock appeared when, for example, a call to pinctrl_pins_show() locked the pinctrl_mutex, called the ops->pin_dbg_show of a particular pin controller. If this pin controller needs I2C access to retrieve configuration information and I2C driver is using pinctrl to drive its pins, a call to pinctrl_select_state() try to lock again pinctrl_mutex which leads to a deadlock. Notice that the mutex grab from the two direction functions was moved into pinctrl_gpio_direction(). For several cases, we can't replace pinctrl_mutex by pctldev->mutex, because at this stage, pctldev is not accessible : - pinctrl_get()/pinctrl_put() - pinctrl_register_maps() So add respectively pinctrl_list_mutex and pinctrl_maps_mutex in order to protect pinctrl_list and pinctrl_maps list instead. Reintroduce pinctrldev_list_mutex in find_pinctrl_by_of_node(), pinctrl_find_and_add_gpio_range() pinctrl_request_gpio(), pinctrl_free_gpio(), pinctrl_gpio_direction(), pinctrl_devices_show(), pinctrl_register() and pinctrl_unregister() to protect pinctrldev_list. Changes v2->v3: - Fix a missing EXPORT_SYMBOL_GPL() for pinctrl_select_state(). Changes v1->v2: - pinctrl_select_state_locked() is removed, all lock mechanism is located inside pinctrl_select_state(). When parsing the state->setting list, take the per-pin-controller driver lock. (Patrice). - Introduce pinctrldev_list_mutex to protect pinctrldev_list in all functions which parse or modify pictrldev_list. (Patrice). - move find_pinctrl_by_of_node() from pinctrl/devicetree.c to pinctrl/core.c in order to protect pinctrldev_list. (Patrice). - Sink mutex:es into some functions and remove some _locked variants down to where the lists are actually accessed to make things simpler. (Linus) - Drop *all* mutexes completely from pinctrl_lookup_state() and pinctrl_select_state() - no relevant mutex was taken and it was unclear what this was protecting against. (Linus) Reported by : Seraphin Bonnaffe <seraphin.bonnaffe@stericsson.com> Signed-off-by: Patrice Chotard <patrice.chotard@st.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2013-04-11 13:01:27 +04:00
return pin;
pinctrl: fix and simplify locking There are many problems with the current pinctrl locking: struct pinctrl_dev's gpio_ranges_lock isn't effective; pinctrl_match_gpio_range() only holds this lock while searching for a gpio range, but the found range is return and manipulated after releading the lock. This could allow pinctrl_remove_gpio_range() for that range while it is in use, and the caller may very well delete the range after removing it, causing pinctrl code to touch the now-free range object. Solving this requires the introduction of a higher-level lock, at least a lock per pin controller, which both gpio range registration and pinctrl_get()/put() will acquire. There is missing locking on HW programming; pin controllers may pack the configuration for different pins/groups/config options/... into one register, and hence have to read-modify-write the register. This needs to be protected, but currently isn't. Related, a future change will add a "complete" op to the pin controller drivers, the idea being that each state's programming will be programmed into the pinctrl driver followed by the "complete" call, which may e.g. flush a register cache to HW. For this to work, it must not be possible to interleave the pinctrl driver calls for different devices. As above, solving this requires the introduction of a higher-level lock, at least a lock per pin controller, which will be held for the duration of any pinctrl_enable()/disable() call. However, each pinctrl mapping table entry may affect a different pin controller if necessary. Hence, with a per-pin-controller lock, almost any pinctrl API may need to acquire multiple locks, one per controller. To avoid deadlock, these would need to be acquired in the same order in all cases. This is extremely difficult to implement in the case of pinctrl_get(), which doesn't know which pin controllers to lock until it has parsed the entire mapping table, since it contains somewhat arbitrary data. The simplest solution here is to introduce a single lock that covers all pin controllers at once. This will be acquired by all pinctrl APIs. This then makes struct pinctrl's mutex irrelevant, since that single lock will always be held whenever this mutex is currently held. Signed-off-by: Stephen Warren <swarren@nvidia.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2012-03-03 00:05:44 +04:00
}
pinctrl: move subsystem mutex to pinctrl_dev struct This mutex avoids deadlock in case of use of multiple pin controllers. Before this modification, by using a global mutex, deadlock appeared when, for example, a call to pinctrl_pins_show() locked the pinctrl_mutex, called the ops->pin_dbg_show of a particular pin controller. If this pin controller needs I2C access to retrieve configuration information and I2C driver is using pinctrl to drive its pins, a call to pinctrl_select_state() try to lock again pinctrl_mutex which leads to a deadlock. Notice that the mutex grab from the two direction functions was moved into pinctrl_gpio_direction(). For several cases, we can't replace pinctrl_mutex by pctldev->mutex, because at this stage, pctldev is not accessible : - pinctrl_get()/pinctrl_put() - pinctrl_register_maps() So add respectively pinctrl_list_mutex and pinctrl_maps_mutex in order to protect pinctrl_list and pinctrl_maps list instead. Reintroduce pinctrldev_list_mutex in find_pinctrl_by_of_node(), pinctrl_find_and_add_gpio_range() pinctrl_request_gpio(), pinctrl_free_gpio(), pinctrl_gpio_direction(), pinctrl_devices_show(), pinctrl_register() and pinctrl_unregister() to protect pinctrldev_list. Changes v2->v3: - Fix a missing EXPORT_SYMBOL_GPL() for pinctrl_select_state(). Changes v1->v2: - pinctrl_select_state_locked() is removed, all lock mechanism is located inside pinctrl_select_state(). When parsing the state->setting list, take the per-pin-controller driver lock. (Patrice). - Introduce pinctrldev_list_mutex to protect pinctrldev_list in all functions which parse or modify pictrldev_list. (Patrice). - move find_pinctrl_by_of_node() from pinctrl/devicetree.c to pinctrl/core.c in order to protect pinctrldev_list. (Patrice). - Sink mutex:es into some functions and remove some _locked variants down to where the lists are actually accessed to make things simpler. (Linus) - Drop *all* mutexes completely from pinctrl_lookup_state() and pinctrl_select_state() - no relevant mutex was taken and it was unclear what this was protecting against. (Linus) Reported by : Seraphin Bonnaffe <seraphin.bonnaffe@stericsson.com> Signed-off-by: Patrice Chotard <patrice.chotard@st.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2013-04-11 13:01:27 +04:00
mutex_lock(&pctldev->mutex);
pinctrl: add a pin config interface This add per-pin and per-group pin config interfaces for biasing, driving and other such electronic properties. The details of passed configurations are passed in an opaque unsigned long which may be dereferences to integer types, structs or lists on either side of the configuration interface. ChangeLog v1->v2: - Clear split of terminology: we now have pin controllers, and those may support two interfaces using vtables: pin multiplexing and pin configuration. - Break out pin configuration to its own C file, controllers may implement only config without mux, and vice versa, so keep each sub-functionality of pin controllers separate. Introduce CONFIG_PINCONF in Kconfig. - Implement some core logic around pin configuration in the pinconf.c file. - Remove UNKNOWN config states, these were just surplus baggage. - Remove FLOAT config state - HIGH_IMPEDANCE should be enough for everyone. - PIN_CONFIG_POWER_SOURCE added to handle switching the power supply for the pin logic between different sources - Explicit DISABLE config enums to turn schmitt-trigger, wakeup etc OFF. - Update documentation to reflect all the recent reasoning. ChangeLog v2->v3: - Twist API around to pass around arrays of config tuples instead of (param, value) pairs everywhere. - Explicit drive strength semantics for push/pull and similar drive modes, this shall be the number of drive stages vs nominal load impedance, which should match the actual electronics used in push/pull CMOS or TTY totempoles. - Drop load capacitance configuration - I probably don't know what I'm doing here so leave it out. - Drop PIN_CONFIG_INPUT_SCHMITT_OFF, instead the argument zero to PIN_CONFIG_INPUT_SCHMITT turns schmitt trigger off. - Drop PIN_CONFIG_NORMAL_POWER_MODE and have a well defined argument to PIN_CONFIG_LOW_POWER_MODE to get out of it instead. - Drop PIN_CONFIG_WAKEUP_ENABLE/DISABLE and just use PIN_CONFIG_WAKEUP with defined value zero to turn wakeup off. - Add PIN_CONFIG_INPUT_DEBOUNCE for configuring debounce time on input lines. - Fix a bug when we tried to configure pins for pin controllers without pinconf support. - Initialized debugfs properly so it works. - Initialize the mutex properly and lock around config tampering sections. - Check the return value from get_initial_config() properly. ChangeLog v3->v4: - Export the pin_config_get(), pin_config_set() and pin_config_group() functions. - Drop the entire concept of just getting initial config and keeping track of pin states internally, instead ask the pins what state they are in. Previous idea was plain wrong, if the device cannot keep track of its state, the driver should do it. - Drop the generic configuration layout, it seems this impose too much restriction on some pin controllers, so let them do things the way they want and split off support for generic config as an optional add-on. ChangeLog v4->v5: - Introduce two symmetric driver calls for group configuration, .pin_config_group_[get|set] and corresponding external calls. - Remove generic semantic meanings of return values from config calls, these belong in the generic config patch. Just pass the return value through instead. - Add a debugfs entry "pinconf-groups" to read status from group configuration only, also slam in a per-group debug callback in the pinconf_ops so custom drivers can display something meaningful for their pins. - Fix some dangling newline. - Drop dangling #else clause. - Update documentation to match the above. ChangeLog v5->v6: - Change to using a pin name as parameter for the [get|set]_config() functions, as suggested by Stephen Warren. This is more natural as names will be what a developer has access to in written documentation etc. ChangeLog v6->v7: - Refactor out by-pin and by-name get/set functions, only expose the by-name functions externally, expose the by-pin functions internally. - Show supported pin control functionality in the debugfs pinctrl-devices file. Acked-by: Stephen Warren <swarren@nvidia.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2011-10-19 20:14:33 +04:00
pin = pin_get_from_name(pctldev, name);
if (pin < 0)
pinctrl: fix and simplify locking There are many problems with the current pinctrl locking: struct pinctrl_dev's gpio_ranges_lock isn't effective; pinctrl_match_gpio_range() only holds this lock while searching for a gpio range, but the found range is return and manipulated after releading the lock. This could allow pinctrl_remove_gpio_range() for that range while it is in use, and the caller may very well delete the range after removing it, causing pinctrl code to touch the now-free range object. Solving this requires the introduction of a higher-level lock, at least a lock per pin controller, which both gpio range registration and pinctrl_get()/put() will acquire. There is missing locking on HW programming; pin controllers may pack the configuration for different pins/groups/config options/... into one register, and hence have to read-modify-write the register. This needs to be protected, but currently isn't. Related, a future change will add a "complete" op to the pin controller drivers, the idea being that each state's programming will be programmed into the pinctrl driver followed by the "complete" call, which may e.g. flush a register cache to HW. For this to work, it must not be possible to interleave the pinctrl driver calls for different devices. As above, solving this requires the introduction of a higher-level lock, at least a lock per pin controller, which will be held for the duration of any pinctrl_enable()/disable() call. However, each pinctrl mapping table entry may affect a different pin controller if necessary. Hence, with a per-pin-controller lock, almost any pinctrl API may need to acquire multiple locks, one per controller. To avoid deadlock, these would need to be acquired in the same order in all cases. This is extremely difficult to implement in the case of pinctrl_get(), which doesn't know which pin controllers to lock until it has parsed the entire mapping table, since it contains somewhat arbitrary data. The simplest solution here is to introduce a single lock that covers all pin controllers at once. This will be acquired by all pinctrl APIs. This then makes struct pinctrl's mutex irrelevant, since that single lock will always be held whenever this mutex is currently held. Signed-off-by: Stephen Warren <swarren@nvidia.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2012-03-03 00:05:44 +04:00
goto unlock;
pinctrl: add a pin config interface This add per-pin and per-group pin config interfaces for biasing, driving and other such electronic properties. The details of passed configurations are passed in an opaque unsigned long which may be dereferences to integer types, structs or lists on either side of the configuration interface. ChangeLog v1->v2: - Clear split of terminology: we now have pin controllers, and those may support two interfaces using vtables: pin multiplexing and pin configuration. - Break out pin configuration to its own C file, controllers may implement only config without mux, and vice versa, so keep each sub-functionality of pin controllers separate. Introduce CONFIG_PINCONF in Kconfig. - Implement some core logic around pin configuration in the pinconf.c file. - Remove UNKNOWN config states, these were just surplus baggage. - Remove FLOAT config state - HIGH_IMPEDANCE should be enough for everyone. - PIN_CONFIG_POWER_SOURCE added to handle switching the power supply for the pin logic between different sources - Explicit DISABLE config enums to turn schmitt-trigger, wakeup etc OFF. - Update documentation to reflect all the recent reasoning. ChangeLog v2->v3: - Twist API around to pass around arrays of config tuples instead of (param, value) pairs everywhere. - Explicit drive strength semantics for push/pull and similar drive modes, this shall be the number of drive stages vs nominal load impedance, which should match the actual electronics used in push/pull CMOS or TTY totempoles. - Drop load capacitance configuration - I probably don't know what I'm doing here so leave it out. - Drop PIN_CONFIG_INPUT_SCHMITT_OFF, instead the argument zero to PIN_CONFIG_INPUT_SCHMITT turns schmitt trigger off. - Drop PIN_CONFIG_NORMAL_POWER_MODE and have a well defined argument to PIN_CONFIG_LOW_POWER_MODE to get out of it instead. - Drop PIN_CONFIG_WAKEUP_ENABLE/DISABLE and just use PIN_CONFIG_WAKEUP with defined value zero to turn wakeup off. - Add PIN_CONFIG_INPUT_DEBOUNCE for configuring debounce time on input lines. - Fix a bug when we tried to configure pins for pin controllers without pinconf support. - Initialized debugfs properly so it works. - Initialize the mutex properly and lock around config tampering sections. - Check the return value from get_initial_config() properly. ChangeLog v3->v4: - Export the pin_config_get(), pin_config_set() and pin_config_group() functions. - Drop the entire concept of just getting initial config and keeping track of pin states internally, instead ask the pins what state they are in. Previous idea was plain wrong, if the device cannot keep track of its state, the driver should do it. - Drop the generic configuration layout, it seems this impose too much restriction on some pin controllers, so let them do things the way they want and split off support for generic config as an optional add-on. ChangeLog v4->v5: - Introduce two symmetric driver calls for group configuration, .pin_config_group_[get|set] and corresponding external calls. - Remove generic semantic meanings of return values from config calls, these belong in the generic config patch. Just pass the return value through instead. - Add a debugfs entry "pinconf-groups" to read status from group configuration only, also slam in a per-group debug callback in the pinconf_ops so custom drivers can display something meaningful for their pins. - Fix some dangling newline. - Drop dangling #else clause. - Update documentation to match the above. ChangeLog v5->v6: - Change to using a pin name as parameter for the [get|set]_config() functions, as suggested by Stephen Warren. This is more natural as names will be what a developer has access to in written documentation etc. ChangeLog v6->v7: - Refactor out by-pin and by-name get/set functions, only expose the by-name functions externally, expose the by-pin functions internally. - Show supported pin control functionality in the debugfs pinctrl-devices file. Acked-by: Stephen Warren <swarren@nvidia.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2011-10-19 20:14:33 +04:00
pinctrl: fix and simplify locking There are many problems with the current pinctrl locking: struct pinctrl_dev's gpio_ranges_lock isn't effective; pinctrl_match_gpio_range() only holds this lock while searching for a gpio range, but the found range is return and manipulated after releading the lock. This could allow pinctrl_remove_gpio_range() for that range while it is in use, and the caller may very well delete the range after removing it, causing pinctrl code to touch the now-free range object. Solving this requires the introduction of a higher-level lock, at least a lock per pin controller, which both gpio range registration and pinctrl_get()/put() will acquire. There is missing locking on HW programming; pin controllers may pack the configuration for different pins/groups/config options/... into one register, and hence have to read-modify-write the register. This needs to be protected, but currently isn't. Related, a future change will add a "complete" op to the pin controller drivers, the idea being that each state's programming will be programmed into the pinctrl driver followed by the "complete" call, which may e.g. flush a register cache to HW. For this to work, it must not be possible to interleave the pinctrl driver calls for different devices. As above, solving this requires the introduction of a higher-level lock, at least a lock per pin controller, which will be held for the duration of any pinctrl_enable()/disable() call. However, each pinctrl mapping table entry may affect a different pin controller if necessary. Hence, with a per-pin-controller lock, almost any pinctrl API may need to acquire multiple locks, one per controller. To avoid deadlock, these would need to be acquired in the same order in all cases. This is extremely difficult to implement in the case of pinctrl_get(), which doesn't know which pin controllers to lock until it has parsed the entire mapping table, since it contains somewhat arbitrary data. The simplest solution here is to introduce a single lock that covers all pin controllers at once. This will be acquired by all pinctrl APIs. This then makes struct pinctrl's mutex irrelevant, since that single lock will always be held whenever this mutex is currently held. Signed-off-by: Stephen Warren <swarren@nvidia.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2012-03-03 00:05:44 +04:00
pin = pin_config_get_for_pin(pctldev, pin, config);
unlock:
pinctrl: move subsystem mutex to pinctrl_dev struct This mutex avoids deadlock in case of use of multiple pin controllers. Before this modification, by using a global mutex, deadlock appeared when, for example, a call to pinctrl_pins_show() locked the pinctrl_mutex, called the ops->pin_dbg_show of a particular pin controller. If this pin controller needs I2C access to retrieve configuration information and I2C driver is using pinctrl to drive its pins, a call to pinctrl_select_state() try to lock again pinctrl_mutex which leads to a deadlock. Notice that the mutex grab from the two direction functions was moved into pinctrl_gpio_direction(). For several cases, we can't replace pinctrl_mutex by pctldev->mutex, because at this stage, pctldev is not accessible : - pinctrl_get()/pinctrl_put() - pinctrl_register_maps() So add respectively pinctrl_list_mutex and pinctrl_maps_mutex in order to protect pinctrl_list and pinctrl_maps list instead. Reintroduce pinctrldev_list_mutex in find_pinctrl_by_of_node(), pinctrl_find_and_add_gpio_range() pinctrl_request_gpio(), pinctrl_free_gpio(), pinctrl_gpio_direction(), pinctrl_devices_show(), pinctrl_register() and pinctrl_unregister() to protect pinctrldev_list. Changes v2->v3: - Fix a missing EXPORT_SYMBOL_GPL() for pinctrl_select_state(). Changes v1->v2: - pinctrl_select_state_locked() is removed, all lock mechanism is located inside pinctrl_select_state(). When parsing the state->setting list, take the per-pin-controller driver lock. (Patrice). - Introduce pinctrldev_list_mutex to protect pinctrldev_list in all functions which parse or modify pictrldev_list. (Patrice). - move find_pinctrl_by_of_node() from pinctrl/devicetree.c to pinctrl/core.c in order to protect pinctrldev_list. (Patrice). - Sink mutex:es into some functions and remove some _locked variants down to where the lists are actually accessed to make things simpler. (Linus) - Drop *all* mutexes completely from pinctrl_lookup_state() and pinctrl_select_state() - no relevant mutex was taken and it was unclear what this was protecting against. (Linus) Reported by : Seraphin Bonnaffe <seraphin.bonnaffe@stericsson.com> Signed-off-by: Patrice Chotard <patrice.chotard@st.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2013-04-11 13:01:27 +04:00
mutex_unlock(&pctldev->mutex);
pinctrl: fix and simplify locking There are many problems with the current pinctrl locking: struct pinctrl_dev's gpio_ranges_lock isn't effective; pinctrl_match_gpio_range() only holds this lock while searching for a gpio range, but the found range is return and manipulated after releading the lock. This could allow pinctrl_remove_gpio_range() for that range while it is in use, and the caller may very well delete the range after removing it, causing pinctrl code to touch the now-free range object. Solving this requires the introduction of a higher-level lock, at least a lock per pin controller, which both gpio range registration and pinctrl_get()/put() will acquire. There is missing locking on HW programming; pin controllers may pack the configuration for different pins/groups/config options/... into one register, and hence have to read-modify-write the register. This needs to be protected, but currently isn't. Related, a future change will add a "complete" op to the pin controller drivers, the idea being that each state's programming will be programmed into the pinctrl driver followed by the "complete" call, which may e.g. flush a register cache to HW. For this to work, it must not be possible to interleave the pinctrl driver calls for different devices. As above, solving this requires the introduction of a higher-level lock, at least a lock per pin controller, which will be held for the duration of any pinctrl_enable()/disable() call. However, each pinctrl mapping table entry may affect a different pin controller if necessary. Hence, with a per-pin-controller lock, almost any pinctrl API may need to acquire multiple locks, one per controller. To avoid deadlock, these would need to be acquired in the same order in all cases. This is extremely difficult to implement in the case of pinctrl_get(), which doesn't know which pin controllers to lock until it has parsed the entire mapping table, since it contains somewhat arbitrary data. The simplest solution here is to introduce a single lock that covers all pin controllers at once. This will be acquired by all pinctrl APIs. This then makes struct pinctrl's mutex irrelevant, since that single lock will always be held whenever this mutex is currently held. Signed-off-by: Stephen Warren <swarren@nvidia.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2012-03-03 00:05:44 +04:00
return pin;
pinctrl: add a pin config interface This add per-pin and per-group pin config interfaces for biasing, driving and other such electronic properties. The details of passed configurations are passed in an opaque unsigned long which may be dereferences to integer types, structs or lists on either side of the configuration interface. ChangeLog v1->v2: - Clear split of terminology: we now have pin controllers, and those may support two interfaces using vtables: pin multiplexing and pin configuration. - Break out pin configuration to its own C file, controllers may implement only config without mux, and vice versa, so keep each sub-functionality of pin controllers separate. Introduce CONFIG_PINCONF in Kconfig. - Implement some core logic around pin configuration in the pinconf.c file. - Remove UNKNOWN config states, these were just surplus baggage. - Remove FLOAT config state - HIGH_IMPEDANCE should be enough for everyone. - PIN_CONFIG_POWER_SOURCE added to handle switching the power supply for the pin logic between different sources - Explicit DISABLE config enums to turn schmitt-trigger, wakeup etc OFF. - Update documentation to reflect all the recent reasoning. ChangeLog v2->v3: - Twist API around to pass around arrays of config tuples instead of (param, value) pairs everywhere. - Explicit drive strength semantics for push/pull and similar drive modes, this shall be the number of drive stages vs nominal load impedance, which should match the actual electronics used in push/pull CMOS or TTY totempoles. - Drop load capacitance configuration - I probably don't know what I'm doing here so leave it out. - Drop PIN_CONFIG_INPUT_SCHMITT_OFF, instead the argument zero to PIN_CONFIG_INPUT_SCHMITT turns schmitt trigger off. - Drop PIN_CONFIG_NORMAL_POWER_MODE and have a well defined argument to PIN_CONFIG_LOW_POWER_MODE to get out of it instead. - Drop PIN_CONFIG_WAKEUP_ENABLE/DISABLE and just use PIN_CONFIG_WAKEUP with defined value zero to turn wakeup off. - Add PIN_CONFIG_INPUT_DEBOUNCE for configuring debounce time on input lines. - Fix a bug when we tried to configure pins for pin controllers without pinconf support. - Initialized debugfs properly so it works. - Initialize the mutex properly and lock around config tampering sections. - Check the return value from get_initial_config() properly. ChangeLog v3->v4: - Export the pin_config_get(), pin_config_set() and pin_config_group() functions. - Drop the entire concept of just getting initial config and keeping track of pin states internally, instead ask the pins what state they are in. Previous idea was plain wrong, if the device cannot keep track of its state, the driver should do it. - Drop the generic configuration layout, it seems this impose too much restriction on some pin controllers, so let them do things the way they want and split off support for generic config as an optional add-on. ChangeLog v4->v5: - Introduce two symmetric driver calls for group configuration, .pin_config_group_[get|set] and corresponding external calls. - Remove generic semantic meanings of return values from config calls, these belong in the generic config patch. Just pass the return value through instead. - Add a debugfs entry "pinconf-groups" to read status from group configuration only, also slam in a per-group debug callback in the pinconf_ops so custom drivers can display something meaningful for their pins. - Fix some dangling newline. - Drop dangling #else clause. - Update documentation to match the above. ChangeLog v5->v6: - Change to using a pin name as parameter for the [get|set]_config() functions, as suggested by Stephen Warren. This is more natural as names will be what a developer has access to in written documentation etc. ChangeLog v6->v7: - Refactor out by-pin and by-name get/set functions, only expose the by-name functions externally, expose the by-pin functions internally. - Show supported pin control functionality in the debugfs pinctrl-devices file. Acked-by: Stephen Warren <swarren@nvidia.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2011-10-19 20:14:33 +04:00
}
EXPORT_SYMBOL(pin_config_get);
static int pin_config_set_for_pin(struct pinctrl_dev *pctldev, unsigned pin,
pinctrl: add a pin config interface This add per-pin and per-group pin config interfaces for biasing, driving and other such electronic properties. The details of passed configurations are passed in an opaque unsigned long which may be dereferences to integer types, structs or lists on either side of the configuration interface. ChangeLog v1->v2: - Clear split of terminology: we now have pin controllers, and those may support two interfaces using vtables: pin multiplexing and pin configuration. - Break out pin configuration to its own C file, controllers may implement only config without mux, and vice versa, so keep each sub-functionality of pin controllers separate. Introduce CONFIG_PINCONF in Kconfig. - Implement some core logic around pin configuration in the pinconf.c file. - Remove UNKNOWN config states, these were just surplus baggage. - Remove FLOAT config state - HIGH_IMPEDANCE should be enough for everyone. - PIN_CONFIG_POWER_SOURCE added to handle switching the power supply for the pin logic between different sources - Explicit DISABLE config enums to turn schmitt-trigger, wakeup etc OFF. - Update documentation to reflect all the recent reasoning. ChangeLog v2->v3: - Twist API around to pass around arrays of config tuples instead of (param, value) pairs everywhere. - Explicit drive strength semantics for push/pull and similar drive modes, this shall be the number of drive stages vs nominal load impedance, which should match the actual electronics used in push/pull CMOS or TTY totempoles. - Drop load capacitance configuration - I probably don't know what I'm doing here so leave it out. - Drop PIN_CONFIG_INPUT_SCHMITT_OFF, instead the argument zero to PIN_CONFIG_INPUT_SCHMITT turns schmitt trigger off. - Drop PIN_CONFIG_NORMAL_POWER_MODE and have a well defined argument to PIN_CONFIG_LOW_POWER_MODE to get out of it instead. - Drop PIN_CONFIG_WAKEUP_ENABLE/DISABLE and just use PIN_CONFIG_WAKEUP with defined value zero to turn wakeup off. - Add PIN_CONFIG_INPUT_DEBOUNCE for configuring debounce time on input lines. - Fix a bug when we tried to configure pins for pin controllers without pinconf support. - Initialized debugfs properly so it works. - Initialize the mutex properly and lock around config tampering sections. - Check the return value from get_initial_config() properly. ChangeLog v3->v4: - Export the pin_config_get(), pin_config_set() and pin_config_group() functions. - Drop the entire concept of just getting initial config and keeping track of pin states internally, instead ask the pins what state they are in. Previous idea was plain wrong, if the device cannot keep track of its state, the driver should do it. - Drop the generic configuration layout, it seems this impose too much restriction on some pin controllers, so let them do things the way they want and split off support for generic config as an optional add-on. ChangeLog v4->v5: - Introduce two symmetric driver calls for group configuration, .pin_config_group_[get|set] and corresponding external calls. - Remove generic semantic meanings of return values from config calls, these belong in the generic config patch. Just pass the return value through instead. - Add a debugfs entry "pinconf-groups" to read status from group configuration only, also slam in a per-group debug callback in the pinconf_ops so custom drivers can display something meaningful for their pins. - Fix some dangling newline. - Drop dangling #else clause. - Update documentation to match the above. ChangeLog v5->v6: - Change to using a pin name as parameter for the [get|set]_config() functions, as suggested by Stephen Warren. This is more natural as names will be what a developer has access to in written documentation etc. ChangeLog v6->v7: - Refactor out by-pin and by-name get/set functions, only expose the by-name functions externally, expose the by-pin functions internally. - Show supported pin control functionality in the debugfs pinctrl-devices file. Acked-by: Stephen Warren <swarren@nvidia.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2011-10-19 20:14:33 +04:00
unsigned long config)
{
const struct pinconf_ops *ops = pctldev->desc->confops;
int ret;
if (!ops || !ops->pin_config_set) {
dev_err(pctldev->dev, "cannot configure pin, missing "
pinctrl: add a pin config interface This add per-pin and per-group pin config interfaces for biasing, driving and other such electronic properties. The details of passed configurations are passed in an opaque unsigned long which may be dereferences to integer types, structs or lists on either side of the configuration interface. ChangeLog v1->v2: - Clear split of terminology: we now have pin controllers, and those may support two interfaces using vtables: pin multiplexing and pin configuration. - Break out pin configuration to its own C file, controllers may implement only config without mux, and vice versa, so keep each sub-functionality of pin controllers separate. Introduce CONFIG_PINCONF in Kconfig. - Implement some core logic around pin configuration in the pinconf.c file. - Remove UNKNOWN config states, these were just surplus baggage. - Remove FLOAT config state - HIGH_IMPEDANCE should be enough for everyone. - PIN_CONFIG_POWER_SOURCE added to handle switching the power supply for the pin logic between different sources - Explicit DISABLE config enums to turn schmitt-trigger, wakeup etc OFF. - Update documentation to reflect all the recent reasoning. ChangeLog v2->v3: - Twist API around to pass around arrays of config tuples instead of (param, value) pairs everywhere. - Explicit drive strength semantics for push/pull and similar drive modes, this shall be the number of drive stages vs nominal load impedance, which should match the actual electronics used in push/pull CMOS or TTY totempoles. - Drop load capacitance configuration - I probably don't know what I'm doing here so leave it out. - Drop PIN_CONFIG_INPUT_SCHMITT_OFF, instead the argument zero to PIN_CONFIG_INPUT_SCHMITT turns schmitt trigger off. - Drop PIN_CONFIG_NORMAL_POWER_MODE and have a well defined argument to PIN_CONFIG_LOW_POWER_MODE to get out of it instead. - Drop PIN_CONFIG_WAKEUP_ENABLE/DISABLE and just use PIN_CONFIG_WAKEUP with defined value zero to turn wakeup off. - Add PIN_CONFIG_INPUT_DEBOUNCE for configuring debounce time on input lines. - Fix a bug when we tried to configure pins for pin controllers without pinconf support. - Initialized debugfs properly so it works. - Initialize the mutex properly and lock around config tampering sections. - Check the return value from get_initial_config() properly. ChangeLog v3->v4: - Export the pin_config_get(), pin_config_set() and pin_config_group() functions. - Drop the entire concept of just getting initial config and keeping track of pin states internally, instead ask the pins what state they are in. Previous idea was plain wrong, if the device cannot keep track of its state, the driver should do it. - Drop the generic configuration layout, it seems this impose too much restriction on some pin controllers, so let them do things the way they want and split off support for generic config as an optional add-on. ChangeLog v4->v5: - Introduce two symmetric driver calls for group configuration, .pin_config_group_[get|set] and corresponding external calls. - Remove generic semantic meanings of return values from config calls, these belong in the generic config patch. Just pass the return value through instead. - Add a debugfs entry "pinconf-groups" to read status from group configuration only, also slam in a per-group debug callback in the pinconf_ops so custom drivers can display something meaningful for their pins. - Fix some dangling newline. - Drop dangling #else clause. - Update documentation to match the above. ChangeLog v5->v6: - Change to using a pin name as parameter for the [get|set]_config() functions, as suggested by Stephen Warren. This is more natural as names will be what a developer has access to in written documentation etc. ChangeLog v6->v7: - Refactor out by-pin and by-name get/set functions, only expose the by-name functions externally, expose the by-pin functions internally. - Show supported pin control functionality in the debugfs pinctrl-devices file. Acked-by: Stephen Warren <swarren@nvidia.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2011-10-19 20:14:33 +04:00
"config function in driver\n");
return -EINVAL;
}
ret = ops->pin_config_set(pctldev, pin, config);
if (ret) {
dev_err(pctldev->dev,
pinctrl: add a pin config interface This add per-pin and per-group pin config interfaces for biasing, driving and other such electronic properties. The details of passed configurations are passed in an opaque unsigned long which may be dereferences to integer types, structs or lists on either side of the configuration interface. ChangeLog v1->v2: - Clear split of terminology: we now have pin controllers, and those may support two interfaces using vtables: pin multiplexing and pin configuration. - Break out pin configuration to its own C file, controllers may implement only config without mux, and vice versa, so keep each sub-functionality of pin controllers separate. Introduce CONFIG_PINCONF in Kconfig. - Implement some core logic around pin configuration in the pinconf.c file. - Remove UNKNOWN config states, these were just surplus baggage. - Remove FLOAT config state - HIGH_IMPEDANCE should be enough for everyone. - PIN_CONFIG_POWER_SOURCE added to handle switching the power supply for the pin logic between different sources - Explicit DISABLE config enums to turn schmitt-trigger, wakeup etc OFF. - Update documentation to reflect all the recent reasoning. ChangeLog v2->v3: - Twist API around to pass around arrays of config tuples instead of (param, value) pairs everywhere. - Explicit drive strength semantics for push/pull and similar drive modes, this shall be the number of drive stages vs nominal load impedance, which should match the actual electronics used in push/pull CMOS or TTY totempoles. - Drop load capacitance configuration - I probably don't know what I'm doing here so leave it out. - Drop PIN_CONFIG_INPUT_SCHMITT_OFF, instead the argument zero to PIN_CONFIG_INPUT_SCHMITT turns schmitt trigger off. - Drop PIN_CONFIG_NORMAL_POWER_MODE and have a well defined argument to PIN_CONFIG_LOW_POWER_MODE to get out of it instead. - Drop PIN_CONFIG_WAKEUP_ENABLE/DISABLE and just use PIN_CONFIG_WAKEUP with defined value zero to turn wakeup off. - Add PIN_CONFIG_INPUT_DEBOUNCE for configuring debounce time on input lines. - Fix a bug when we tried to configure pins for pin controllers without pinconf support. - Initialized debugfs properly so it works. - Initialize the mutex properly and lock around config tampering sections. - Check the return value from get_initial_config() properly. ChangeLog v3->v4: - Export the pin_config_get(), pin_config_set() and pin_config_group() functions. - Drop the entire concept of just getting initial config and keeping track of pin states internally, instead ask the pins what state they are in. Previous idea was plain wrong, if the device cannot keep track of its state, the driver should do it. - Drop the generic configuration layout, it seems this impose too much restriction on some pin controllers, so let them do things the way they want and split off support for generic config as an optional add-on. ChangeLog v4->v5: - Introduce two symmetric driver calls for group configuration, .pin_config_group_[get|set] and corresponding external calls. - Remove generic semantic meanings of return values from config calls, these belong in the generic config patch. Just pass the return value through instead. - Add a debugfs entry "pinconf-groups" to read status from group configuration only, also slam in a per-group debug callback in the pinconf_ops so custom drivers can display something meaningful for their pins. - Fix some dangling newline. - Drop dangling #else clause. - Update documentation to match the above. ChangeLog v5->v6: - Change to using a pin name as parameter for the [get|set]_config() functions, as suggested by Stephen Warren. This is more natural as names will be what a developer has access to in written documentation etc. ChangeLog v6->v7: - Refactor out by-pin and by-name get/set functions, only expose the by-name functions externally, expose the by-pin functions internally. - Show supported pin control functionality in the debugfs pinctrl-devices file. Acked-by: Stephen Warren <swarren@nvidia.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2011-10-19 20:14:33 +04:00
"unable to set pin configuration on pin %d\n", pin);
return ret;
}
return 0;
}
/**
* pin_config_set() - set the configuration of a single pin parameter
* @dev_name: name of pin controller device for this pin
pinctrl: add a pin config interface This add per-pin and per-group pin config interfaces for biasing, driving and other such electronic properties. The details of passed configurations are passed in an opaque unsigned long which may be dereferences to integer types, structs or lists on either side of the configuration interface. ChangeLog v1->v2: - Clear split of terminology: we now have pin controllers, and those may support two interfaces using vtables: pin multiplexing and pin configuration. - Break out pin configuration to its own C file, controllers may implement only config without mux, and vice versa, so keep each sub-functionality of pin controllers separate. Introduce CONFIG_PINCONF in Kconfig. - Implement some core logic around pin configuration in the pinconf.c file. - Remove UNKNOWN config states, these were just surplus baggage. - Remove FLOAT config state - HIGH_IMPEDANCE should be enough for everyone. - PIN_CONFIG_POWER_SOURCE added to handle switching the power supply for the pin logic between different sources - Explicit DISABLE config enums to turn schmitt-trigger, wakeup etc OFF. - Update documentation to reflect all the recent reasoning. ChangeLog v2->v3: - Twist API around to pass around arrays of config tuples instead of (param, value) pairs everywhere. - Explicit drive strength semantics for push/pull and similar drive modes, this shall be the number of drive stages vs nominal load impedance, which should match the actual electronics used in push/pull CMOS or TTY totempoles. - Drop load capacitance configuration - I probably don't know what I'm doing here so leave it out. - Drop PIN_CONFIG_INPUT_SCHMITT_OFF, instead the argument zero to PIN_CONFIG_INPUT_SCHMITT turns schmitt trigger off. - Drop PIN_CONFIG_NORMAL_POWER_MODE and have a well defined argument to PIN_CONFIG_LOW_POWER_MODE to get out of it instead. - Drop PIN_CONFIG_WAKEUP_ENABLE/DISABLE and just use PIN_CONFIG_WAKEUP with defined value zero to turn wakeup off. - Add PIN_CONFIG_INPUT_DEBOUNCE for configuring debounce time on input lines. - Fix a bug when we tried to configure pins for pin controllers without pinconf support. - Initialized debugfs properly so it works. - Initialize the mutex properly and lock around config tampering sections. - Check the return value from get_initial_config() properly. ChangeLog v3->v4: - Export the pin_config_get(), pin_config_set() and pin_config_group() functions. - Drop the entire concept of just getting initial config and keeping track of pin states internally, instead ask the pins what state they are in. Previous idea was plain wrong, if the device cannot keep track of its state, the driver should do it. - Drop the generic configuration layout, it seems this impose too much restriction on some pin controllers, so let them do things the way they want and split off support for generic config as an optional add-on. ChangeLog v4->v5: - Introduce two symmetric driver calls for group configuration, .pin_config_group_[get|set] and corresponding external calls. - Remove generic semantic meanings of return values from config calls, these belong in the generic config patch. Just pass the return value through instead. - Add a debugfs entry "pinconf-groups" to read status from group configuration only, also slam in a per-group debug callback in the pinconf_ops so custom drivers can display something meaningful for their pins. - Fix some dangling newline. - Drop dangling #else clause. - Update documentation to match the above. ChangeLog v5->v6: - Change to using a pin name as parameter for the [get|set]_config() functions, as suggested by Stephen Warren. This is more natural as names will be what a developer has access to in written documentation etc. ChangeLog v6->v7: - Refactor out by-pin and by-name get/set functions, only expose the by-name functions externally, expose the by-pin functions internally. - Show supported pin control functionality in the debugfs pinctrl-devices file. Acked-by: Stephen Warren <swarren@nvidia.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2011-10-19 20:14:33 +04:00
* @name: name of the pin to set the config for
* @config: the config in this argument will contain the desired pin state, it
* can be used directly by drivers as a numeral, or it can be dereferenced
* to any struct.
*/
int pin_config_set(const char *dev_name, const char *name,
pinctrl: add a pin config interface This add per-pin and per-group pin config interfaces for biasing, driving and other such electronic properties. The details of passed configurations are passed in an opaque unsigned long which may be dereferences to integer types, structs or lists on either side of the configuration interface. ChangeLog v1->v2: - Clear split of terminology: we now have pin controllers, and those may support two interfaces using vtables: pin multiplexing and pin configuration. - Break out pin configuration to its own C file, controllers may implement only config without mux, and vice versa, so keep each sub-functionality of pin controllers separate. Introduce CONFIG_PINCONF in Kconfig. - Implement some core logic around pin configuration in the pinconf.c file. - Remove UNKNOWN config states, these were just surplus baggage. - Remove FLOAT config state - HIGH_IMPEDANCE should be enough for everyone. - PIN_CONFIG_POWER_SOURCE added to handle switching the power supply for the pin logic between different sources - Explicit DISABLE config enums to turn schmitt-trigger, wakeup etc OFF. - Update documentation to reflect all the recent reasoning. ChangeLog v2->v3: - Twist API around to pass around arrays of config tuples instead of (param, value) pairs everywhere. - Explicit drive strength semantics for push/pull and similar drive modes, this shall be the number of drive stages vs nominal load impedance, which should match the actual electronics used in push/pull CMOS or TTY totempoles. - Drop load capacitance configuration - I probably don't know what I'm doing here so leave it out. - Drop PIN_CONFIG_INPUT_SCHMITT_OFF, instead the argument zero to PIN_CONFIG_INPUT_SCHMITT turns schmitt trigger off. - Drop PIN_CONFIG_NORMAL_POWER_MODE and have a well defined argument to PIN_CONFIG_LOW_POWER_MODE to get out of it instead. - Drop PIN_CONFIG_WAKEUP_ENABLE/DISABLE and just use PIN_CONFIG_WAKEUP with defined value zero to turn wakeup off. - Add PIN_CONFIG_INPUT_DEBOUNCE for configuring debounce time on input lines. - Fix a bug when we tried to configure pins for pin controllers without pinconf support. - Initialized debugfs properly so it works. - Initialize the mutex properly and lock around config tampering sections. - Check the return value from get_initial_config() properly. ChangeLog v3->v4: - Export the pin_config_get(), pin_config_set() and pin_config_group() functions. - Drop the entire concept of just getting initial config and keeping track of pin states internally, instead ask the pins what state they are in. Previous idea was plain wrong, if the device cannot keep track of its state, the driver should do it. - Drop the generic configuration layout, it seems this impose too much restriction on some pin controllers, so let them do things the way they want and split off support for generic config as an optional add-on. ChangeLog v4->v5: - Introduce two symmetric driver calls for group configuration, .pin_config_group_[get|set] and corresponding external calls. - Remove generic semantic meanings of return values from config calls, these belong in the generic config patch. Just pass the return value through instead. - Add a debugfs entry "pinconf-groups" to read status from group configuration only, also slam in a per-group debug callback in the pinconf_ops so custom drivers can display something meaningful for their pins. - Fix some dangling newline. - Drop dangling #else clause. - Update documentation to match the above. ChangeLog v5->v6: - Change to using a pin name as parameter for the [get|set]_config() functions, as suggested by Stephen Warren. This is more natural as names will be what a developer has access to in written documentation etc. ChangeLog v6->v7: - Refactor out by-pin and by-name get/set functions, only expose the by-name functions externally, expose the by-pin functions internally. - Show supported pin control functionality in the debugfs pinctrl-devices file. Acked-by: Stephen Warren <swarren@nvidia.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2011-10-19 20:14:33 +04:00
unsigned long config)
{
struct pinctrl_dev *pctldev;
pinctrl: fix and simplify locking There are many problems with the current pinctrl locking: struct pinctrl_dev's gpio_ranges_lock isn't effective; pinctrl_match_gpio_range() only holds this lock while searching for a gpio range, but the found range is return and manipulated after releading the lock. This could allow pinctrl_remove_gpio_range() for that range while it is in use, and the caller may very well delete the range after removing it, causing pinctrl code to touch the now-free range object. Solving this requires the introduction of a higher-level lock, at least a lock per pin controller, which both gpio range registration and pinctrl_get()/put() will acquire. There is missing locking on HW programming; pin controllers may pack the configuration for different pins/groups/config options/... into one register, and hence have to read-modify-write the register. This needs to be protected, but currently isn't. Related, a future change will add a "complete" op to the pin controller drivers, the idea being that each state's programming will be programmed into the pinctrl driver followed by the "complete" call, which may e.g. flush a register cache to HW. For this to work, it must not be possible to interleave the pinctrl driver calls for different devices. As above, solving this requires the introduction of a higher-level lock, at least a lock per pin controller, which will be held for the duration of any pinctrl_enable()/disable() call. However, each pinctrl mapping table entry may affect a different pin controller if necessary. Hence, with a per-pin-controller lock, almost any pinctrl API may need to acquire multiple locks, one per controller. To avoid deadlock, these would need to be acquired in the same order in all cases. This is extremely difficult to implement in the case of pinctrl_get(), which doesn't know which pin controllers to lock until it has parsed the entire mapping table, since it contains somewhat arbitrary data. The simplest solution here is to introduce a single lock that covers all pin controllers at once. This will be acquired by all pinctrl APIs. This then makes struct pinctrl's mutex irrelevant, since that single lock will always be held whenever this mutex is currently held. Signed-off-by: Stephen Warren <swarren@nvidia.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2012-03-03 00:05:44 +04:00
int pin, ret;
pctldev = get_pinctrl_dev_from_devname(dev_name);
pinctrl: fix and simplify locking There are many problems with the current pinctrl locking: struct pinctrl_dev's gpio_ranges_lock isn't effective; pinctrl_match_gpio_range() only holds this lock while searching for a gpio range, but the found range is return and manipulated after releading the lock. This could allow pinctrl_remove_gpio_range() for that range while it is in use, and the caller may very well delete the range after removing it, causing pinctrl code to touch the now-free range object. Solving this requires the introduction of a higher-level lock, at least a lock per pin controller, which both gpio range registration and pinctrl_get()/put() will acquire. There is missing locking on HW programming; pin controllers may pack the configuration for different pins/groups/config options/... into one register, and hence have to read-modify-write the register. This needs to be protected, but currently isn't. Related, a future change will add a "complete" op to the pin controller drivers, the idea being that each state's programming will be programmed into the pinctrl driver followed by the "complete" call, which may e.g. flush a register cache to HW. For this to work, it must not be possible to interleave the pinctrl driver calls for different devices. As above, solving this requires the introduction of a higher-level lock, at least a lock per pin controller, which will be held for the duration of any pinctrl_enable()/disable() call. However, each pinctrl mapping table entry may affect a different pin controller if necessary. Hence, with a per-pin-controller lock, almost any pinctrl API may need to acquire multiple locks, one per controller. To avoid deadlock, these would need to be acquired in the same order in all cases. This is extremely difficult to implement in the case of pinctrl_get(), which doesn't know which pin controllers to lock until it has parsed the entire mapping table, since it contains somewhat arbitrary data. The simplest solution here is to introduce a single lock that covers all pin controllers at once. This will be acquired by all pinctrl APIs. This then makes struct pinctrl's mutex irrelevant, since that single lock will always be held whenever this mutex is currently held. Signed-off-by: Stephen Warren <swarren@nvidia.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2012-03-03 00:05:44 +04:00
if (!pctldev) {
ret = -EINVAL;
pinctrl: move subsystem mutex to pinctrl_dev struct This mutex avoids deadlock in case of use of multiple pin controllers. Before this modification, by using a global mutex, deadlock appeared when, for example, a call to pinctrl_pins_show() locked the pinctrl_mutex, called the ops->pin_dbg_show of a particular pin controller. If this pin controller needs I2C access to retrieve configuration information and I2C driver is using pinctrl to drive its pins, a call to pinctrl_select_state() try to lock again pinctrl_mutex which leads to a deadlock. Notice that the mutex grab from the two direction functions was moved into pinctrl_gpio_direction(). For several cases, we can't replace pinctrl_mutex by pctldev->mutex, because at this stage, pctldev is not accessible : - pinctrl_get()/pinctrl_put() - pinctrl_register_maps() So add respectively pinctrl_list_mutex and pinctrl_maps_mutex in order to protect pinctrl_list and pinctrl_maps list instead. Reintroduce pinctrldev_list_mutex in find_pinctrl_by_of_node(), pinctrl_find_and_add_gpio_range() pinctrl_request_gpio(), pinctrl_free_gpio(), pinctrl_gpio_direction(), pinctrl_devices_show(), pinctrl_register() and pinctrl_unregister() to protect pinctrldev_list. Changes v2->v3: - Fix a missing EXPORT_SYMBOL_GPL() for pinctrl_select_state(). Changes v1->v2: - pinctrl_select_state_locked() is removed, all lock mechanism is located inside pinctrl_select_state(). When parsing the state->setting list, take the per-pin-controller driver lock. (Patrice). - Introduce pinctrldev_list_mutex to protect pinctrldev_list in all functions which parse or modify pictrldev_list. (Patrice). - move find_pinctrl_by_of_node() from pinctrl/devicetree.c to pinctrl/core.c in order to protect pinctrldev_list. (Patrice). - Sink mutex:es into some functions and remove some _locked variants down to where the lists are actually accessed to make things simpler. (Linus) - Drop *all* mutexes completely from pinctrl_lookup_state() and pinctrl_select_state() - no relevant mutex was taken and it was unclear what this was protecting against. (Linus) Reported by : Seraphin Bonnaffe <seraphin.bonnaffe@stericsson.com> Signed-off-by: Patrice Chotard <patrice.chotard@st.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2013-04-11 13:01:27 +04:00
return ret;
pinctrl: fix and simplify locking There are many problems with the current pinctrl locking: struct pinctrl_dev's gpio_ranges_lock isn't effective; pinctrl_match_gpio_range() only holds this lock while searching for a gpio range, but the found range is return and manipulated after releading the lock. This could allow pinctrl_remove_gpio_range() for that range while it is in use, and the caller may very well delete the range after removing it, causing pinctrl code to touch the now-free range object. Solving this requires the introduction of a higher-level lock, at least a lock per pin controller, which both gpio range registration and pinctrl_get()/put() will acquire. There is missing locking on HW programming; pin controllers may pack the configuration for different pins/groups/config options/... into one register, and hence have to read-modify-write the register. This needs to be protected, but currently isn't. Related, a future change will add a "complete" op to the pin controller drivers, the idea being that each state's programming will be programmed into the pinctrl driver followed by the "complete" call, which may e.g. flush a register cache to HW. For this to work, it must not be possible to interleave the pinctrl driver calls for different devices. As above, solving this requires the introduction of a higher-level lock, at least a lock per pin controller, which will be held for the duration of any pinctrl_enable()/disable() call. However, each pinctrl mapping table entry may affect a different pin controller if necessary. Hence, with a per-pin-controller lock, almost any pinctrl API may need to acquire multiple locks, one per controller. To avoid deadlock, these would need to be acquired in the same order in all cases. This is extremely difficult to implement in the case of pinctrl_get(), which doesn't know which pin controllers to lock until it has parsed the entire mapping table, since it contains somewhat arbitrary data. The simplest solution here is to introduce a single lock that covers all pin controllers at once. This will be acquired by all pinctrl APIs. This then makes struct pinctrl's mutex irrelevant, since that single lock will always be held whenever this mutex is currently held. Signed-off-by: Stephen Warren <swarren@nvidia.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2012-03-03 00:05:44 +04:00
}
pinctrl: move subsystem mutex to pinctrl_dev struct This mutex avoids deadlock in case of use of multiple pin controllers. Before this modification, by using a global mutex, deadlock appeared when, for example, a call to pinctrl_pins_show() locked the pinctrl_mutex, called the ops->pin_dbg_show of a particular pin controller. If this pin controller needs I2C access to retrieve configuration information and I2C driver is using pinctrl to drive its pins, a call to pinctrl_select_state() try to lock again pinctrl_mutex which leads to a deadlock. Notice that the mutex grab from the two direction functions was moved into pinctrl_gpio_direction(). For several cases, we can't replace pinctrl_mutex by pctldev->mutex, because at this stage, pctldev is not accessible : - pinctrl_get()/pinctrl_put() - pinctrl_register_maps() So add respectively pinctrl_list_mutex and pinctrl_maps_mutex in order to protect pinctrl_list and pinctrl_maps list instead. Reintroduce pinctrldev_list_mutex in find_pinctrl_by_of_node(), pinctrl_find_and_add_gpio_range() pinctrl_request_gpio(), pinctrl_free_gpio(), pinctrl_gpio_direction(), pinctrl_devices_show(), pinctrl_register() and pinctrl_unregister() to protect pinctrldev_list. Changes v2->v3: - Fix a missing EXPORT_SYMBOL_GPL() for pinctrl_select_state(). Changes v1->v2: - pinctrl_select_state_locked() is removed, all lock mechanism is located inside pinctrl_select_state(). When parsing the state->setting list, take the per-pin-controller driver lock. (Patrice). - Introduce pinctrldev_list_mutex to protect pinctrldev_list in all functions which parse or modify pictrldev_list. (Patrice). - move find_pinctrl_by_of_node() from pinctrl/devicetree.c to pinctrl/core.c in order to protect pinctrldev_list. (Patrice). - Sink mutex:es into some functions and remove some _locked variants down to where the lists are actually accessed to make things simpler. (Linus) - Drop *all* mutexes completely from pinctrl_lookup_state() and pinctrl_select_state() - no relevant mutex was taken and it was unclear what this was protecting against. (Linus) Reported by : Seraphin Bonnaffe <seraphin.bonnaffe@stericsson.com> Signed-off-by: Patrice Chotard <patrice.chotard@st.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2013-04-11 13:01:27 +04:00
mutex_lock(&pctldev->mutex);
pinctrl: add a pin config interface This add per-pin and per-group pin config interfaces for biasing, driving and other such electronic properties. The details of passed configurations are passed in an opaque unsigned long which may be dereferences to integer types, structs or lists on either side of the configuration interface. ChangeLog v1->v2: - Clear split of terminology: we now have pin controllers, and those may support two interfaces using vtables: pin multiplexing and pin configuration. - Break out pin configuration to its own C file, controllers may implement only config without mux, and vice versa, so keep each sub-functionality of pin controllers separate. Introduce CONFIG_PINCONF in Kconfig. - Implement some core logic around pin configuration in the pinconf.c file. - Remove UNKNOWN config states, these were just surplus baggage. - Remove FLOAT config state - HIGH_IMPEDANCE should be enough for everyone. - PIN_CONFIG_POWER_SOURCE added to handle switching the power supply for the pin logic between different sources - Explicit DISABLE config enums to turn schmitt-trigger, wakeup etc OFF. - Update documentation to reflect all the recent reasoning. ChangeLog v2->v3: - Twist API around to pass around arrays of config tuples instead of (param, value) pairs everywhere. - Explicit drive strength semantics for push/pull and similar drive modes, this shall be the number of drive stages vs nominal load impedance, which should match the actual electronics used in push/pull CMOS or TTY totempoles. - Drop load capacitance configuration - I probably don't know what I'm doing here so leave it out. - Drop PIN_CONFIG_INPUT_SCHMITT_OFF, instead the argument zero to PIN_CONFIG_INPUT_SCHMITT turns schmitt trigger off. - Drop PIN_CONFIG_NORMAL_POWER_MODE and have a well defined argument to PIN_CONFIG_LOW_POWER_MODE to get out of it instead. - Drop PIN_CONFIG_WAKEUP_ENABLE/DISABLE and just use PIN_CONFIG_WAKEUP with defined value zero to turn wakeup off. - Add PIN_CONFIG_INPUT_DEBOUNCE for configuring debounce time on input lines. - Fix a bug when we tried to configure pins for pin controllers without pinconf support. - Initialized debugfs properly so it works. - Initialize the mutex properly and lock around config tampering sections. - Check the return value from get_initial_config() properly. ChangeLog v3->v4: - Export the pin_config_get(), pin_config_set() and pin_config_group() functions. - Drop the entire concept of just getting initial config and keeping track of pin states internally, instead ask the pins what state they are in. Previous idea was plain wrong, if the device cannot keep track of its state, the driver should do it. - Drop the generic configuration layout, it seems this impose too much restriction on some pin controllers, so let them do things the way they want and split off support for generic config as an optional add-on. ChangeLog v4->v5: - Introduce two symmetric driver calls for group configuration, .pin_config_group_[get|set] and corresponding external calls. - Remove generic semantic meanings of return values from config calls, these belong in the generic config patch. Just pass the return value through instead. - Add a debugfs entry "pinconf-groups" to read status from group configuration only, also slam in a per-group debug callback in the pinconf_ops so custom drivers can display something meaningful for their pins. - Fix some dangling newline. - Drop dangling #else clause. - Update documentation to match the above. ChangeLog v5->v6: - Change to using a pin name as parameter for the [get|set]_config() functions, as suggested by Stephen Warren. This is more natural as names will be what a developer has access to in written documentation etc. ChangeLog v6->v7: - Refactor out by-pin and by-name get/set functions, only expose the by-name functions externally, expose the by-pin functions internally. - Show supported pin control functionality in the debugfs pinctrl-devices file. Acked-by: Stephen Warren <swarren@nvidia.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2011-10-19 20:14:33 +04:00
pin = pin_get_from_name(pctldev, name);
pinctrl: fix and simplify locking There are many problems with the current pinctrl locking: struct pinctrl_dev's gpio_ranges_lock isn't effective; pinctrl_match_gpio_range() only holds this lock while searching for a gpio range, but the found range is return and manipulated after releading the lock. This could allow pinctrl_remove_gpio_range() for that range while it is in use, and the caller may very well delete the range after removing it, causing pinctrl code to touch the now-free range object. Solving this requires the introduction of a higher-level lock, at least a lock per pin controller, which both gpio range registration and pinctrl_get()/put() will acquire. There is missing locking on HW programming; pin controllers may pack the configuration for different pins/groups/config options/... into one register, and hence have to read-modify-write the register. This needs to be protected, but currently isn't. Related, a future change will add a "complete" op to the pin controller drivers, the idea being that each state's programming will be programmed into the pinctrl driver followed by the "complete" call, which may e.g. flush a register cache to HW. For this to work, it must not be possible to interleave the pinctrl driver calls for different devices. As above, solving this requires the introduction of a higher-level lock, at least a lock per pin controller, which will be held for the duration of any pinctrl_enable()/disable() call. However, each pinctrl mapping table entry may affect a different pin controller if necessary. Hence, with a per-pin-controller lock, almost any pinctrl API may need to acquire multiple locks, one per controller. To avoid deadlock, these would need to be acquired in the same order in all cases. This is extremely difficult to implement in the case of pinctrl_get(), which doesn't know which pin controllers to lock until it has parsed the entire mapping table, since it contains somewhat arbitrary data. The simplest solution here is to introduce a single lock that covers all pin controllers at once. This will be acquired by all pinctrl APIs. This then makes struct pinctrl's mutex irrelevant, since that single lock will always be held whenever this mutex is currently held. Signed-off-by: Stephen Warren <swarren@nvidia.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2012-03-03 00:05:44 +04:00
if (pin < 0) {
ret = pin;
goto unlock;
}
ret = pin_config_set_for_pin(pctldev, pin, config);
pinctrl: add a pin config interface This add per-pin and per-group pin config interfaces for biasing, driving and other such electronic properties. The details of passed configurations are passed in an opaque unsigned long which may be dereferences to integer types, structs or lists on either side of the configuration interface. ChangeLog v1->v2: - Clear split of terminology: we now have pin controllers, and those may support two interfaces using vtables: pin multiplexing and pin configuration. - Break out pin configuration to its own C file, controllers may implement only config without mux, and vice versa, so keep each sub-functionality of pin controllers separate. Introduce CONFIG_PINCONF in Kconfig. - Implement some core logic around pin configuration in the pinconf.c file. - Remove UNKNOWN config states, these were just surplus baggage. - Remove FLOAT config state - HIGH_IMPEDANCE should be enough for everyone. - PIN_CONFIG_POWER_SOURCE added to handle switching the power supply for the pin logic between different sources - Explicit DISABLE config enums to turn schmitt-trigger, wakeup etc OFF. - Update documentation to reflect all the recent reasoning. ChangeLog v2->v3: - Twist API around to pass around arrays of config tuples instead of (param, value) pairs everywhere. - Explicit drive strength semantics for push/pull and similar drive modes, this shall be the number of drive stages vs nominal load impedance, which should match the actual electronics used in push/pull CMOS or TTY totempoles. - Drop load capacitance configuration - I probably don't know what I'm doing here so leave it out. - Drop PIN_CONFIG_INPUT_SCHMITT_OFF, instead the argument zero to PIN_CONFIG_INPUT_SCHMITT turns schmitt trigger off. - Drop PIN_CONFIG_NORMAL_POWER_MODE and have a well defined argument to PIN_CONFIG_LOW_POWER_MODE to get out of it instead. - Drop PIN_CONFIG_WAKEUP_ENABLE/DISABLE and just use PIN_CONFIG_WAKEUP with defined value zero to turn wakeup off. - Add PIN_CONFIG_INPUT_DEBOUNCE for configuring debounce time on input lines. - Fix a bug when we tried to configure pins for pin controllers without pinconf support. - Initialized debugfs properly so it works. - Initialize the mutex properly and lock around config tampering sections. - Check the return value from get_initial_config() properly. ChangeLog v3->v4: - Export the pin_config_get(), pin_config_set() and pin_config_group() functions. - Drop the entire concept of just getting initial config and keeping track of pin states internally, instead ask the pins what state they are in. Previous idea was plain wrong, if the device cannot keep track of its state, the driver should do it. - Drop the generic configuration layout, it seems this impose too much restriction on some pin controllers, so let them do things the way they want and split off support for generic config as an optional add-on. ChangeLog v4->v5: - Introduce two symmetric driver calls for group configuration, .pin_config_group_[get|set] and corresponding external calls. - Remove generic semantic meanings of return values from config calls, these belong in the generic config patch. Just pass the return value through instead. - Add a debugfs entry "pinconf-groups" to read status from group configuration only, also slam in a per-group debug callback in the pinconf_ops so custom drivers can display something meaningful for their pins. - Fix some dangling newline. - Drop dangling #else clause. - Update documentation to match the above. ChangeLog v5->v6: - Change to using a pin name as parameter for the [get|set]_config() functions, as suggested by Stephen Warren. This is more natural as names will be what a developer has access to in written documentation etc. ChangeLog v6->v7: - Refactor out by-pin and by-name get/set functions, only expose the by-name functions externally, expose the by-pin functions internally. - Show supported pin control functionality in the debugfs pinctrl-devices file. Acked-by: Stephen Warren <swarren@nvidia.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2011-10-19 20:14:33 +04:00
pinctrl: fix and simplify locking There are many problems with the current pinctrl locking: struct pinctrl_dev's gpio_ranges_lock isn't effective; pinctrl_match_gpio_range() only holds this lock while searching for a gpio range, but the found range is return and manipulated after releading the lock. This could allow pinctrl_remove_gpio_range() for that range while it is in use, and the caller may very well delete the range after removing it, causing pinctrl code to touch the now-free range object. Solving this requires the introduction of a higher-level lock, at least a lock per pin controller, which both gpio range registration and pinctrl_get()/put() will acquire. There is missing locking on HW programming; pin controllers may pack the configuration for different pins/groups/config options/... into one register, and hence have to read-modify-write the register. This needs to be protected, but currently isn't. Related, a future change will add a "complete" op to the pin controller drivers, the idea being that each state's programming will be programmed into the pinctrl driver followed by the "complete" call, which may e.g. flush a register cache to HW. For this to work, it must not be possible to interleave the pinctrl driver calls for different devices. As above, solving this requires the introduction of a higher-level lock, at least a lock per pin controller, which will be held for the duration of any pinctrl_enable()/disable() call. However, each pinctrl mapping table entry may affect a different pin controller if necessary. Hence, with a per-pin-controller lock, almost any pinctrl API may need to acquire multiple locks, one per controller. To avoid deadlock, these would need to be acquired in the same order in all cases. This is extremely difficult to implement in the case of pinctrl_get(), which doesn't know which pin controllers to lock until it has parsed the entire mapping table, since it contains somewhat arbitrary data. The simplest solution here is to introduce a single lock that covers all pin controllers at once. This will be acquired by all pinctrl APIs. This then makes struct pinctrl's mutex irrelevant, since that single lock will always be held whenever this mutex is currently held. Signed-off-by: Stephen Warren <swarren@nvidia.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2012-03-03 00:05:44 +04:00
unlock:
pinctrl: move subsystem mutex to pinctrl_dev struct This mutex avoids deadlock in case of use of multiple pin controllers. Before this modification, by using a global mutex, deadlock appeared when, for example, a call to pinctrl_pins_show() locked the pinctrl_mutex, called the ops->pin_dbg_show of a particular pin controller. If this pin controller needs I2C access to retrieve configuration information and I2C driver is using pinctrl to drive its pins, a call to pinctrl_select_state() try to lock again pinctrl_mutex which leads to a deadlock. Notice that the mutex grab from the two direction functions was moved into pinctrl_gpio_direction(). For several cases, we can't replace pinctrl_mutex by pctldev->mutex, because at this stage, pctldev is not accessible : - pinctrl_get()/pinctrl_put() - pinctrl_register_maps() So add respectively pinctrl_list_mutex and pinctrl_maps_mutex in order to protect pinctrl_list and pinctrl_maps list instead. Reintroduce pinctrldev_list_mutex in find_pinctrl_by_of_node(), pinctrl_find_and_add_gpio_range() pinctrl_request_gpio(), pinctrl_free_gpio(), pinctrl_gpio_direction(), pinctrl_devices_show(), pinctrl_register() and pinctrl_unregister() to protect pinctrldev_list. Changes v2->v3: - Fix a missing EXPORT_SYMBOL_GPL() for pinctrl_select_state(). Changes v1->v2: - pinctrl_select_state_locked() is removed, all lock mechanism is located inside pinctrl_select_state(). When parsing the state->setting list, take the per-pin-controller driver lock. (Patrice). - Introduce pinctrldev_list_mutex to protect pinctrldev_list in all functions which parse or modify pictrldev_list. (Patrice). - move find_pinctrl_by_of_node() from pinctrl/devicetree.c to pinctrl/core.c in order to protect pinctrldev_list. (Patrice). - Sink mutex:es into some functions and remove some _locked variants down to where the lists are actually accessed to make things simpler. (Linus) - Drop *all* mutexes completely from pinctrl_lookup_state() and pinctrl_select_state() - no relevant mutex was taken and it was unclear what this was protecting against. (Linus) Reported by : Seraphin Bonnaffe <seraphin.bonnaffe@stericsson.com> Signed-off-by: Patrice Chotard <patrice.chotard@st.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2013-04-11 13:01:27 +04:00
mutex_unlock(&pctldev->mutex);
pinctrl: fix and simplify locking There are many problems with the current pinctrl locking: struct pinctrl_dev's gpio_ranges_lock isn't effective; pinctrl_match_gpio_range() only holds this lock while searching for a gpio range, but the found range is return and manipulated after releading the lock. This could allow pinctrl_remove_gpio_range() for that range while it is in use, and the caller may very well delete the range after removing it, causing pinctrl code to touch the now-free range object. Solving this requires the introduction of a higher-level lock, at least a lock per pin controller, which both gpio range registration and pinctrl_get()/put() will acquire. There is missing locking on HW programming; pin controllers may pack the configuration for different pins/groups/config options/... into one register, and hence have to read-modify-write the register. This needs to be protected, but currently isn't. Related, a future change will add a "complete" op to the pin controller drivers, the idea being that each state's programming will be programmed into the pinctrl driver followed by the "complete" call, which may e.g. flush a register cache to HW. For this to work, it must not be possible to interleave the pinctrl driver calls for different devices. As above, solving this requires the introduction of a higher-level lock, at least a lock per pin controller, which will be held for the duration of any pinctrl_enable()/disable() call. However, each pinctrl mapping table entry may affect a different pin controller if necessary. Hence, with a per-pin-controller lock, almost any pinctrl API may need to acquire multiple locks, one per controller. To avoid deadlock, these would need to be acquired in the same order in all cases. This is extremely difficult to implement in the case of pinctrl_get(), which doesn't know which pin controllers to lock until it has parsed the entire mapping table, since it contains somewhat arbitrary data. The simplest solution here is to introduce a single lock that covers all pin controllers at once. This will be acquired by all pinctrl APIs. This then makes struct pinctrl's mutex irrelevant, since that single lock will always be held whenever this mutex is currently held. Signed-off-by: Stephen Warren <swarren@nvidia.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2012-03-03 00:05:44 +04:00
return ret;
pinctrl: add a pin config interface This add per-pin and per-group pin config interfaces for biasing, driving and other such electronic properties. The details of passed configurations are passed in an opaque unsigned long which may be dereferences to integer types, structs or lists on either side of the configuration interface. ChangeLog v1->v2: - Clear split of terminology: we now have pin controllers, and those may support two interfaces using vtables: pin multiplexing and pin configuration. - Break out pin configuration to its own C file, controllers may implement only config without mux, and vice versa, so keep each sub-functionality of pin controllers separate. Introduce CONFIG_PINCONF in Kconfig. - Implement some core logic around pin configuration in the pinconf.c file. - Remove UNKNOWN config states, these were just surplus baggage. - Remove FLOAT config state - HIGH_IMPEDANCE should be enough for everyone. - PIN_CONFIG_POWER_SOURCE added to handle switching the power supply for the pin logic between different sources - Explicit DISABLE config enums to turn schmitt-trigger, wakeup etc OFF. - Update documentation to reflect all the recent reasoning. ChangeLog v2->v3: - Twist API around to pass around arrays of config tuples instead of (param, value) pairs everywhere. - Explicit drive strength semantics for push/pull and similar drive modes, this shall be the number of drive stages vs nominal load impedance, which should match the actual electronics used in push/pull CMOS or TTY totempoles. - Drop load capacitance configuration - I probably don't know what I'm doing here so leave it out. - Drop PIN_CONFIG_INPUT_SCHMITT_OFF, instead the argument zero to PIN_CONFIG_INPUT_SCHMITT turns schmitt trigger off. - Drop PIN_CONFIG_NORMAL_POWER_MODE and have a well defined argument to PIN_CONFIG_LOW_POWER_MODE to get out of it instead. - Drop PIN_CONFIG_WAKEUP_ENABLE/DISABLE and just use PIN_CONFIG_WAKEUP with defined value zero to turn wakeup off. - Add PIN_CONFIG_INPUT_DEBOUNCE for configuring debounce time on input lines. - Fix a bug when we tried to configure pins for pin controllers without pinconf support. - Initialized debugfs properly so it works. - Initialize the mutex properly and lock around config tampering sections. - Check the return value from get_initial_config() properly. ChangeLog v3->v4: - Export the pin_config_get(), pin_config_set() and pin_config_group() functions. - Drop the entire concept of just getting initial config and keeping track of pin states internally, instead ask the pins what state they are in. Previous idea was plain wrong, if the device cannot keep track of its state, the driver should do it. - Drop the generic configuration layout, it seems this impose too much restriction on some pin controllers, so let them do things the way they want and split off support for generic config as an optional add-on. ChangeLog v4->v5: - Introduce two symmetric driver calls for group configuration, .pin_config_group_[get|set] and corresponding external calls. - Remove generic semantic meanings of return values from config calls, these belong in the generic config patch. Just pass the return value through instead. - Add a debugfs entry "pinconf-groups" to read status from group configuration only, also slam in a per-group debug callback in the pinconf_ops so custom drivers can display something meaningful for their pins. - Fix some dangling newline. - Drop dangling #else clause. - Update documentation to match the above. ChangeLog v5->v6: - Change to using a pin name as parameter for the [get|set]_config() functions, as suggested by Stephen Warren. This is more natural as names will be what a developer has access to in written documentation etc. ChangeLog v6->v7: - Refactor out by-pin and by-name get/set functions, only expose the by-name functions externally, expose the by-pin functions internally. - Show supported pin control functionality in the debugfs pinctrl-devices file. Acked-by: Stephen Warren <swarren@nvidia.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2011-10-19 20:14:33 +04:00
}
EXPORT_SYMBOL(pin_config_set);
int pin_config_group_get(const char *dev_name, const char *pin_group,
pinctrl: add a pin config interface This add per-pin and per-group pin config interfaces for biasing, driving and other such electronic properties. The details of passed configurations are passed in an opaque unsigned long which may be dereferences to integer types, structs or lists on either side of the configuration interface. ChangeLog v1->v2: - Clear split of terminology: we now have pin controllers, and those may support two interfaces using vtables: pin multiplexing and pin configuration. - Break out pin configuration to its own C file, controllers may implement only config without mux, and vice versa, so keep each sub-functionality of pin controllers separate. Introduce CONFIG_PINCONF in Kconfig. - Implement some core logic around pin configuration in the pinconf.c file. - Remove UNKNOWN config states, these were just surplus baggage. - Remove FLOAT config state - HIGH_IMPEDANCE should be enough for everyone. - PIN_CONFIG_POWER_SOURCE added to handle switching the power supply for the pin logic between different sources - Explicit DISABLE config enums to turn schmitt-trigger, wakeup etc OFF. - Update documentation to reflect all the recent reasoning. ChangeLog v2->v3: - Twist API around to pass around arrays of config tuples instead of (param, value) pairs everywhere. - Explicit drive strength semantics for push/pull and similar drive modes, this shall be the number of drive stages vs nominal load impedance, which should match the actual electronics used in push/pull CMOS or TTY totempoles. - Drop load capacitance configuration - I probably don't know what I'm doing here so leave it out. - Drop PIN_CONFIG_INPUT_SCHMITT_OFF, instead the argument zero to PIN_CONFIG_INPUT_SCHMITT turns schmitt trigger off. - Drop PIN_CONFIG_NORMAL_POWER_MODE and have a well defined argument to PIN_CONFIG_LOW_POWER_MODE to get out of it instead. - Drop PIN_CONFIG_WAKEUP_ENABLE/DISABLE and just use PIN_CONFIG_WAKEUP with defined value zero to turn wakeup off. - Add PIN_CONFIG_INPUT_DEBOUNCE for configuring debounce time on input lines. - Fix a bug when we tried to configure pins for pin controllers without pinconf support. - Initialized debugfs properly so it works. - Initialize the mutex properly and lock around config tampering sections. - Check the return value from get_initial_config() properly. ChangeLog v3->v4: - Export the pin_config_get(), pin_config_set() and pin_config_group() functions. - Drop the entire concept of just getting initial config and keeping track of pin states internally, instead ask the pins what state they are in. Previous idea was plain wrong, if the device cannot keep track of its state, the driver should do it. - Drop the generic configuration layout, it seems this impose too much restriction on some pin controllers, so let them do things the way they want and split off support for generic config as an optional add-on. ChangeLog v4->v5: - Introduce two symmetric driver calls for group configuration, .pin_config_group_[get|set] and corresponding external calls. - Remove generic semantic meanings of return values from config calls, these belong in the generic config patch. Just pass the return value through instead. - Add a debugfs entry "pinconf-groups" to read status from group configuration only, also slam in a per-group debug callback in the pinconf_ops so custom drivers can display something meaningful for their pins. - Fix some dangling newline. - Drop dangling #else clause. - Update documentation to match the above. ChangeLog v5->v6: - Change to using a pin name as parameter for the [get|set]_config() functions, as suggested by Stephen Warren. This is more natural as names will be what a developer has access to in written documentation etc. ChangeLog v6->v7: - Refactor out by-pin and by-name get/set functions, only expose the by-name functions externally, expose the by-pin functions internally. - Show supported pin control functionality in the debugfs pinctrl-devices file. Acked-by: Stephen Warren <swarren@nvidia.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2011-10-19 20:14:33 +04:00
unsigned long *config)
{
struct pinctrl_dev *pctldev;
const struct pinconf_ops *ops;
pinctrl: fix and simplify locking There are many problems with the current pinctrl locking: struct pinctrl_dev's gpio_ranges_lock isn't effective; pinctrl_match_gpio_range() only holds this lock while searching for a gpio range, but the found range is return and manipulated after releading the lock. This could allow pinctrl_remove_gpio_range() for that range while it is in use, and the caller may very well delete the range after removing it, causing pinctrl code to touch the now-free range object. Solving this requires the introduction of a higher-level lock, at least a lock per pin controller, which both gpio range registration and pinctrl_get()/put() will acquire. There is missing locking on HW programming; pin controllers may pack the configuration for different pins/groups/config options/... into one register, and hence have to read-modify-write the register. This needs to be protected, but currently isn't. Related, a future change will add a "complete" op to the pin controller drivers, the idea being that each state's programming will be programmed into the pinctrl driver followed by the "complete" call, which may e.g. flush a register cache to HW. For this to work, it must not be possible to interleave the pinctrl driver calls for different devices. As above, solving this requires the introduction of a higher-level lock, at least a lock per pin controller, which will be held for the duration of any pinctrl_enable()/disable() call. However, each pinctrl mapping table entry may affect a different pin controller if necessary. Hence, with a per-pin-controller lock, almost any pinctrl API may need to acquire multiple locks, one per controller. To avoid deadlock, these would need to be acquired in the same order in all cases. This is extremely difficult to implement in the case of pinctrl_get(), which doesn't know which pin controllers to lock until it has parsed the entire mapping table, since it contains somewhat arbitrary data. The simplest solution here is to introduce a single lock that covers all pin controllers at once. This will be acquired by all pinctrl APIs. This then makes struct pinctrl's mutex irrelevant, since that single lock will always be held whenever this mutex is currently held. Signed-off-by: Stephen Warren <swarren@nvidia.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2012-03-03 00:05:44 +04:00
int selector, ret;
pctldev = get_pinctrl_dev_from_devname(dev_name);
pinctrl: fix and simplify locking There are many problems with the current pinctrl locking: struct pinctrl_dev's gpio_ranges_lock isn't effective; pinctrl_match_gpio_range() only holds this lock while searching for a gpio range, but the found range is return and manipulated after releading the lock. This could allow pinctrl_remove_gpio_range() for that range while it is in use, and the caller may very well delete the range after removing it, causing pinctrl code to touch the now-free range object. Solving this requires the introduction of a higher-level lock, at least a lock per pin controller, which both gpio range registration and pinctrl_get()/put() will acquire. There is missing locking on HW programming; pin controllers may pack the configuration for different pins/groups/config options/... into one register, and hence have to read-modify-write the register. This needs to be protected, but currently isn't. Related, a future change will add a "complete" op to the pin controller drivers, the idea being that each state's programming will be programmed into the pinctrl driver followed by the "complete" call, which may e.g. flush a register cache to HW. For this to work, it must not be possible to interleave the pinctrl driver calls for different devices. As above, solving this requires the introduction of a higher-level lock, at least a lock per pin controller, which will be held for the duration of any pinctrl_enable()/disable() call. However, each pinctrl mapping table entry may affect a different pin controller if necessary. Hence, with a per-pin-controller lock, almost any pinctrl API may need to acquire multiple locks, one per controller. To avoid deadlock, these would need to be acquired in the same order in all cases. This is extremely difficult to implement in the case of pinctrl_get(), which doesn't know which pin controllers to lock until it has parsed the entire mapping table, since it contains somewhat arbitrary data. The simplest solution here is to introduce a single lock that covers all pin controllers at once. This will be acquired by all pinctrl APIs. This then makes struct pinctrl's mutex irrelevant, since that single lock will always be held whenever this mutex is currently held. Signed-off-by: Stephen Warren <swarren@nvidia.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2012-03-03 00:05:44 +04:00
if (!pctldev) {
ret = -EINVAL;
pinctrl: move subsystem mutex to pinctrl_dev struct This mutex avoids deadlock in case of use of multiple pin controllers. Before this modification, by using a global mutex, deadlock appeared when, for example, a call to pinctrl_pins_show() locked the pinctrl_mutex, called the ops->pin_dbg_show of a particular pin controller. If this pin controller needs I2C access to retrieve configuration information and I2C driver is using pinctrl to drive its pins, a call to pinctrl_select_state() try to lock again pinctrl_mutex which leads to a deadlock. Notice that the mutex grab from the two direction functions was moved into pinctrl_gpio_direction(). For several cases, we can't replace pinctrl_mutex by pctldev->mutex, because at this stage, pctldev is not accessible : - pinctrl_get()/pinctrl_put() - pinctrl_register_maps() So add respectively pinctrl_list_mutex and pinctrl_maps_mutex in order to protect pinctrl_list and pinctrl_maps list instead. Reintroduce pinctrldev_list_mutex in find_pinctrl_by_of_node(), pinctrl_find_and_add_gpio_range() pinctrl_request_gpio(), pinctrl_free_gpio(), pinctrl_gpio_direction(), pinctrl_devices_show(), pinctrl_register() and pinctrl_unregister() to protect pinctrldev_list. Changes v2->v3: - Fix a missing EXPORT_SYMBOL_GPL() for pinctrl_select_state(). Changes v1->v2: - pinctrl_select_state_locked() is removed, all lock mechanism is located inside pinctrl_select_state(). When parsing the state->setting list, take the per-pin-controller driver lock. (Patrice). - Introduce pinctrldev_list_mutex to protect pinctrldev_list in all functions which parse or modify pictrldev_list. (Patrice). - move find_pinctrl_by_of_node() from pinctrl/devicetree.c to pinctrl/core.c in order to protect pinctrldev_list. (Patrice). - Sink mutex:es into some functions and remove some _locked variants down to where the lists are actually accessed to make things simpler. (Linus) - Drop *all* mutexes completely from pinctrl_lookup_state() and pinctrl_select_state() - no relevant mutex was taken and it was unclear what this was protecting against. (Linus) Reported by : Seraphin Bonnaffe <seraphin.bonnaffe@stericsson.com> Signed-off-by: Patrice Chotard <patrice.chotard@st.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2013-04-11 13:01:27 +04:00
return ret;
pinctrl: fix and simplify locking There are many problems with the current pinctrl locking: struct pinctrl_dev's gpio_ranges_lock isn't effective; pinctrl_match_gpio_range() only holds this lock while searching for a gpio range, but the found range is return and manipulated after releading the lock. This could allow pinctrl_remove_gpio_range() for that range while it is in use, and the caller may very well delete the range after removing it, causing pinctrl code to touch the now-free range object. Solving this requires the introduction of a higher-level lock, at least a lock per pin controller, which both gpio range registration and pinctrl_get()/put() will acquire. There is missing locking on HW programming; pin controllers may pack the configuration for different pins/groups/config options/... into one register, and hence have to read-modify-write the register. This needs to be protected, but currently isn't. Related, a future change will add a "complete" op to the pin controller drivers, the idea being that each state's programming will be programmed into the pinctrl driver followed by the "complete" call, which may e.g. flush a register cache to HW. For this to work, it must not be possible to interleave the pinctrl driver calls for different devices. As above, solving this requires the introduction of a higher-level lock, at least a lock per pin controller, which will be held for the duration of any pinctrl_enable()/disable() call. However, each pinctrl mapping table entry may affect a different pin controller if necessary. Hence, with a per-pin-controller lock, almost any pinctrl API may need to acquire multiple locks, one per controller. To avoid deadlock, these would need to be acquired in the same order in all cases. This is extremely difficult to implement in the case of pinctrl_get(), which doesn't know which pin controllers to lock until it has parsed the entire mapping table, since it contains somewhat arbitrary data. The simplest solution here is to introduce a single lock that covers all pin controllers at once. This will be acquired by all pinctrl APIs. This then makes struct pinctrl's mutex irrelevant, since that single lock will always be held whenever this mutex is currently held. Signed-off-by: Stephen Warren <swarren@nvidia.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2012-03-03 00:05:44 +04:00
}
pinctrl: move subsystem mutex to pinctrl_dev struct This mutex avoids deadlock in case of use of multiple pin controllers. Before this modification, by using a global mutex, deadlock appeared when, for example, a call to pinctrl_pins_show() locked the pinctrl_mutex, called the ops->pin_dbg_show of a particular pin controller. If this pin controller needs I2C access to retrieve configuration information and I2C driver is using pinctrl to drive its pins, a call to pinctrl_select_state() try to lock again pinctrl_mutex which leads to a deadlock. Notice that the mutex grab from the two direction functions was moved into pinctrl_gpio_direction(). For several cases, we can't replace pinctrl_mutex by pctldev->mutex, because at this stage, pctldev is not accessible : - pinctrl_get()/pinctrl_put() - pinctrl_register_maps() So add respectively pinctrl_list_mutex and pinctrl_maps_mutex in order to protect pinctrl_list and pinctrl_maps list instead. Reintroduce pinctrldev_list_mutex in find_pinctrl_by_of_node(), pinctrl_find_and_add_gpio_range() pinctrl_request_gpio(), pinctrl_free_gpio(), pinctrl_gpio_direction(), pinctrl_devices_show(), pinctrl_register() and pinctrl_unregister() to protect pinctrldev_list. Changes v2->v3: - Fix a missing EXPORT_SYMBOL_GPL() for pinctrl_select_state(). Changes v1->v2: - pinctrl_select_state_locked() is removed, all lock mechanism is located inside pinctrl_select_state(). When parsing the state->setting list, take the per-pin-controller driver lock. (Patrice). - Introduce pinctrldev_list_mutex to protect pinctrldev_list in all functions which parse or modify pictrldev_list. (Patrice). - move find_pinctrl_by_of_node() from pinctrl/devicetree.c to pinctrl/core.c in order to protect pinctrldev_list. (Patrice). - Sink mutex:es into some functions and remove some _locked variants down to where the lists are actually accessed to make things simpler. (Linus) - Drop *all* mutexes completely from pinctrl_lookup_state() and pinctrl_select_state() - no relevant mutex was taken and it was unclear what this was protecting against. (Linus) Reported by : Seraphin Bonnaffe <seraphin.bonnaffe@stericsson.com> Signed-off-by: Patrice Chotard <patrice.chotard@st.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2013-04-11 13:01:27 +04:00
mutex_lock(&pctldev->mutex);
ops = pctldev->desc->confops;
pinctrl: add a pin config interface This add per-pin and per-group pin config interfaces for biasing, driving and other such electronic properties. The details of passed configurations are passed in an opaque unsigned long which may be dereferences to integer types, structs or lists on either side of the configuration interface. ChangeLog v1->v2: - Clear split of terminology: we now have pin controllers, and those may support two interfaces using vtables: pin multiplexing and pin configuration. - Break out pin configuration to its own C file, controllers may implement only config without mux, and vice versa, so keep each sub-functionality of pin controllers separate. Introduce CONFIG_PINCONF in Kconfig. - Implement some core logic around pin configuration in the pinconf.c file. - Remove UNKNOWN config states, these were just surplus baggage. - Remove FLOAT config state - HIGH_IMPEDANCE should be enough for everyone. - PIN_CONFIG_POWER_SOURCE added to handle switching the power supply for the pin logic between different sources - Explicit DISABLE config enums to turn schmitt-trigger, wakeup etc OFF. - Update documentation to reflect all the recent reasoning. ChangeLog v2->v3: - Twist API around to pass around arrays of config tuples instead of (param, value) pairs everywhere. - Explicit drive strength semantics for push/pull and similar drive modes, this shall be the number of drive stages vs nominal load impedance, which should match the actual electronics used in push/pull CMOS or TTY totempoles. - Drop load capacitance configuration - I probably don't know what I'm doing here so leave it out. - Drop PIN_CONFIG_INPUT_SCHMITT_OFF, instead the argument zero to PIN_CONFIG_INPUT_SCHMITT turns schmitt trigger off. - Drop PIN_CONFIG_NORMAL_POWER_MODE and have a well defined argument to PIN_CONFIG_LOW_POWER_MODE to get out of it instead. - Drop PIN_CONFIG_WAKEUP_ENABLE/DISABLE and just use PIN_CONFIG_WAKEUP with defined value zero to turn wakeup off. - Add PIN_CONFIG_INPUT_DEBOUNCE for configuring debounce time on input lines. - Fix a bug when we tried to configure pins for pin controllers without pinconf support. - Initialized debugfs properly so it works. - Initialize the mutex properly and lock around config tampering sections. - Check the return value from get_initial_config() properly. ChangeLog v3->v4: - Export the pin_config_get(), pin_config_set() and pin_config_group() functions. - Drop the entire concept of just getting initial config and keeping track of pin states internally, instead ask the pins what state they are in. Previous idea was plain wrong, if the device cannot keep track of its state, the driver should do it. - Drop the generic configuration layout, it seems this impose too much restriction on some pin controllers, so let them do things the way they want and split off support for generic config as an optional add-on. ChangeLog v4->v5: - Introduce two symmetric driver calls for group configuration, .pin_config_group_[get|set] and corresponding external calls. - Remove generic semantic meanings of return values from config calls, these belong in the generic config patch. Just pass the return value through instead. - Add a debugfs entry "pinconf-groups" to read status from group configuration only, also slam in a per-group debug callback in the pinconf_ops so custom drivers can display something meaningful for their pins. - Fix some dangling newline. - Drop dangling #else clause. - Update documentation to match the above. ChangeLog v5->v6: - Change to using a pin name as parameter for the [get|set]_config() functions, as suggested by Stephen Warren. This is more natural as names will be what a developer has access to in written documentation etc. ChangeLog v6->v7: - Refactor out by-pin and by-name get/set functions, only expose the by-name functions externally, expose the by-pin functions internally. - Show supported pin control functionality in the debugfs pinctrl-devices file. Acked-by: Stephen Warren <swarren@nvidia.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2011-10-19 20:14:33 +04:00
if (!ops || !ops->pin_config_group_get) {
dev_err(pctldev->dev, "cannot get configuration for pin "
pinctrl: add a pin config interface This add per-pin and per-group pin config interfaces for biasing, driving and other such electronic properties. The details of passed configurations are passed in an opaque unsigned long which may be dereferences to integer types, structs or lists on either side of the configuration interface. ChangeLog v1->v2: - Clear split of terminology: we now have pin controllers, and those may support two interfaces using vtables: pin multiplexing and pin configuration. - Break out pin configuration to its own C file, controllers may implement only config without mux, and vice versa, so keep each sub-functionality of pin controllers separate. Introduce CONFIG_PINCONF in Kconfig. - Implement some core logic around pin configuration in the pinconf.c file. - Remove UNKNOWN config states, these were just surplus baggage. - Remove FLOAT config state - HIGH_IMPEDANCE should be enough for everyone. - PIN_CONFIG_POWER_SOURCE added to handle switching the power supply for the pin logic between different sources - Explicit DISABLE config enums to turn schmitt-trigger, wakeup etc OFF. - Update documentation to reflect all the recent reasoning. ChangeLog v2->v3: - Twist API around to pass around arrays of config tuples instead of (param, value) pairs everywhere. - Explicit drive strength semantics for push/pull and similar drive modes, this shall be the number of drive stages vs nominal load impedance, which should match the actual electronics used in push/pull CMOS or TTY totempoles. - Drop load capacitance configuration - I probably don't know what I'm doing here so leave it out. - Drop PIN_CONFIG_INPUT_SCHMITT_OFF, instead the argument zero to PIN_CONFIG_INPUT_SCHMITT turns schmitt trigger off. - Drop PIN_CONFIG_NORMAL_POWER_MODE and have a well defined argument to PIN_CONFIG_LOW_POWER_MODE to get out of it instead. - Drop PIN_CONFIG_WAKEUP_ENABLE/DISABLE and just use PIN_CONFIG_WAKEUP with defined value zero to turn wakeup off. - Add PIN_CONFIG_INPUT_DEBOUNCE for configuring debounce time on input lines. - Fix a bug when we tried to configure pins for pin controllers without pinconf support. - Initialized debugfs properly so it works. - Initialize the mutex properly and lock around config tampering sections. - Check the return value from get_initial_config() properly. ChangeLog v3->v4: - Export the pin_config_get(), pin_config_set() and pin_config_group() functions. - Drop the entire concept of just getting initial config and keeping track of pin states internally, instead ask the pins what state they are in. Previous idea was plain wrong, if the device cannot keep track of its state, the driver should do it. - Drop the generic configuration layout, it seems this impose too much restriction on some pin controllers, so let them do things the way they want and split off support for generic config as an optional add-on. ChangeLog v4->v5: - Introduce two symmetric driver calls for group configuration, .pin_config_group_[get|set] and corresponding external calls. - Remove generic semantic meanings of return values from config calls, these belong in the generic config patch. Just pass the return value through instead. - Add a debugfs entry "pinconf-groups" to read status from group configuration only, also slam in a per-group debug callback in the pinconf_ops so custom drivers can display something meaningful for their pins. - Fix some dangling newline. - Drop dangling #else clause. - Update documentation to match the above. ChangeLog v5->v6: - Change to using a pin name as parameter for the [get|set]_config() functions, as suggested by Stephen Warren. This is more natural as names will be what a developer has access to in written documentation etc. ChangeLog v6->v7: - Refactor out by-pin and by-name get/set functions, only expose the by-name functions externally, expose the by-pin functions internally. - Show supported pin control functionality in the debugfs pinctrl-devices file. Acked-by: Stephen Warren <swarren@nvidia.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2011-10-19 20:14:33 +04:00
"group, missing group config get function in "
"driver\n");
pinctrl: fix and simplify locking There are many problems with the current pinctrl locking: struct pinctrl_dev's gpio_ranges_lock isn't effective; pinctrl_match_gpio_range() only holds this lock while searching for a gpio range, but the found range is return and manipulated after releading the lock. This could allow pinctrl_remove_gpio_range() for that range while it is in use, and the caller may very well delete the range after removing it, causing pinctrl code to touch the now-free range object. Solving this requires the introduction of a higher-level lock, at least a lock per pin controller, which both gpio range registration and pinctrl_get()/put() will acquire. There is missing locking on HW programming; pin controllers may pack the configuration for different pins/groups/config options/... into one register, and hence have to read-modify-write the register. This needs to be protected, but currently isn't. Related, a future change will add a "complete" op to the pin controller drivers, the idea being that each state's programming will be programmed into the pinctrl driver followed by the "complete" call, which may e.g. flush a register cache to HW. For this to work, it must not be possible to interleave the pinctrl driver calls for different devices. As above, solving this requires the introduction of a higher-level lock, at least a lock per pin controller, which will be held for the duration of any pinctrl_enable()/disable() call. However, each pinctrl mapping table entry may affect a different pin controller if necessary. Hence, with a per-pin-controller lock, almost any pinctrl API may need to acquire multiple locks, one per controller. To avoid deadlock, these would need to be acquired in the same order in all cases. This is extremely difficult to implement in the case of pinctrl_get(), which doesn't know which pin controllers to lock until it has parsed the entire mapping table, since it contains somewhat arbitrary data. The simplest solution here is to introduce a single lock that covers all pin controllers at once. This will be acquired by all pinctrl APIs. This then makes struct pinctrl's mutex irrelevant, since that single lock will always be held whenever this mutex is currently held. Signed-off-by: Stephen Warren <swarren@nvidia.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2012-03-03 00:05:44 +04:00
ret = -EINVAL;
goto unlock;
pinctrl: add a pin config interface This add per-pin and per-group pin config interfaces for biasing, driving and other such electronic properties. The details of passed configurations are passed in an opaque unsigned long which may be dereferences to integer types, structs or lists on either side of the configuration interface. ChangeLog v1->v2: - Clear split of terminology: we now have pin controllers, and those may support two interfaces using vtables: pin multiplexing and pin configuration. - Break out pin configuration to its own C file, controllers may implement only config without mux, and vice versa, so keep each sub-functionality of pin controllers separate. Introduce CONFIG_PINCONF in Kconfig. - Implement some core logic around pin configuration in the pinconf.c file. - Remove UNKNOWN config states, these were just surplus baggage. - Remove FLOAT config state - HIGH_IMPEDANCE should be enough for everyone. - PIN_CONFIG_POWER_SOURCE added to handle switching the power supply for the pin logic between different sources - Explicit DISABLE config enums to turn schmitt-trigger, wakeup etc OFF. - Update documentation to reflect all the recent reasoning. ChangeLog v2->v3: - Twist API around to pass around arrays of config tuples instead of (param, value) pairs everywhere. - Explicit drive strength semantics for push/pull and similar drive modes, this shall be the number of drive stages vs nominal load impedance, which should match the actual electronics used in push/pull CMOS or TTY totempoles. - Drop load capacitance configuration - I probably don't know what I'm doing here so leave it out. - Drop PIN_CONFIG_INPUT_SCHMITT_OFF, instead the argument zero to PIN_CONFIG_INPUT_SCHMITT turns schmitt trigger off. - Drop PIN_CONFIG_NORMAL_POWER_MODE and have a well defined argument to PIN_CONFIG_LOW_POWER_MODE to get out of it instead. - Drop PIN_CONFIG_WAKEUP_ENABLE/DISABLE and just use PIN_CONFIG_WAKEUP with defined value zero to turn wakeup off. - Add PIN_CONFIG_INPUT_DEBOUNCE for configuring debounce time on input lines. - Fix a bug when we tried to configure pins for pin controllers without pinconf support. - Initialized debugfs properly so it works. - Initialize the mutex properly and lock around config tampering sections. - Check the return value from get_initial_config() properly. ChangeLog v3->v4: - Export the pin_config_get(), pin_config_set() and pin_config_group() functions. - Drop the entire concept of just getting initial config and keeping track of pin states internally, instead ask the pins what state they are in. Previous idea was plain wrong, if the device cannot keep track of its state, the driver should do it. - Drop the generic configuration layout, it seems this impose too much restriction on some pin controllers, so let them do things the way they want and split off support for generic config as an optional add-on. ChangeLog v4->v5: - Introduce two symmetric driver calls for group configuration, .pin_config_group_[get|set] and corresponding external calls. - Remove generic semantic meanings of return values from config calls, these belong in the generic config patch. Just pass the return value through instead. - Add a debugfs entry "pinconf-groups" to read status from group configuration only, also slam in a per-group debug callback in the pinconf_ops so custom drivers can display something meaningful for their pins. - Fix some dangling newline. - Drop dangling #else clause. - Update documentation to match the above. ChangeLog v5->v6: - Change to using a pin name as parameter for the [get|set]_config() functions, as suggested by Stephen Warren. This is more natural as names will be what a developer has access to in written documentation etc. ChangeLog v6->v7: - Refactor out by-pin and by-name get/set functions, only expose the by-name functions externally, expose the by-pin functions internally. - Show supported pin control functionality in the debugfs pinctrl-devices file. Acked-by: Stephen Warren <swarren@nvidia.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2011-10-19 20:14:33 +04:00
}
selector = pinctrl_get_group_selector(pctldev, pin_group);
pinctrl: fix and simplify locking There are many problems with the current pinctrl locking: struct pinctrl_dev's gpio_ranges_lock isn't effective; pinctrl_match_gpio_range() only holds this lock while searching for a gpio range, but the found range is return and manipulated after releading the lock. This could allow pinctrl_remove_gpio_range() for that range while it is in use, and the caller may very well delete the range after removing it, causing pinctrl code to touch the now-free range object. Solving this requires the introduction of a higher-level lock, at least a lock per pin controller, which both gpio range registration and pinctrl_get()/put() will acquire. There is missing locking on HW programming; pin controllers may pack the configuration for different pins/groups/config options/... into one register, and hence have to read-modify-write the register. This needs to be protected, but currently isn't. Related, a future change will add a "complete" op to the pin controller drivers, the idea being that each state's programming will be programmed into the pinctrl driver followed by the "complete" call, which may e.g. flush a register cache to HW. For this to work, it must not be possible to interleave the pinctrl driver calls for different devices. As above, solving this requires the introduction of a higher-level lock, at least a lock per pin controller, which will be held for the duration of any pinctrl_enable()/disable() call. However, each pinctrl mapping table entry may affect a different pin controller if necessary. Hence, with a per-pin-controller lock, almost any pinctrl API may need to acquire multiple locks, one per controller. To avoid deadlock, these would need to be acquired in the same order in all cases. This is extremely difficult to implement in the case of pinctrl_get(), which doesn't know which pin controllers to lock until it has parsed the entire mapping table, since it contains somewhat arbitrary data. The simplest solution here is to introduce a single lock that covers all pin controllers at once. This will be acquired by all pinctrl APIs. This then makes struct pinctrl's mutex irrelevant, since that single lock will always be held whenever this mutex is currently held. Signed-off-by: Stephen Warren <swarren@nvidia.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2012-03-03 00:05:44 +04:00
if (selector < 0) {
ret = selector;
goto unlock;
}
pinctrl: add a pin config interface This add per-pin and per-group pin config interfaces for biasing, driving and other such electronic properties. The details of passed configurations are passed in an opaque unsigned long which may be dereferences to integer types, structs or lists on either side of the configuration interface. ChangeLog v1->v2: - Clear split of terminology: we now have pin controllers, and those may support two interfaces using vtables: pin multiplexing and pin configuration. - Break out pin configuration to its own C file, controllers may implement only config without mux, and vice versa, so keep each sub-functionality of pin controllers separate. Introduce CONFIG_PINCONF in Kconfig. - Implement some core logic around pin configuration in the pinconf.c file. - Remove UNKNOWN config states, these were just surplus baggage. - Remove FLOAT config state - HIGH_IMPEDANCE should be enough for everyone. - PIN_CONFIG_POWER_SOURCE added to handle switching the power supply for the pin logic between different sources - Explicit DISABLE config enums to turn schmitt-trigger, wakeup etc OFF. - Update documentation to reflect all the recent reasoning. ChangeLog v2->v3: - Twist API around to pass around arrays of config tuples instead of (param, value) pairs everywhere. - Explicit drive strength semantics for push/pull and similar drive modes, this shall be the number of drive stages vs nominal load impedance, which should match the actual electronics used in push/pull CMOS or TTY totempoles. - Drop load capacitance configuration - I probably don't know what I'm doing here so leave it out. - Drop PIN_CONFIG_INPUT_SCHMITT_OFF, instead the argument zero to PIN_CONFIG_INPUT_SCHMITT turns schmitt trigger off. - Drop PIN_CONFIG_NORMAL_POWER_MODE and have a well defined argument to PIN_CONFIG_LOW_POWER_MODE to get out of it instead. - Drop PIN_CONFIG_WAKEUP_ENABLE/DISABLE and just use PIN_CONFIG_WAKEUP with defined value zero to turn wakeup off. - Add PIN_CONFIG_INPUT_DEBOUNCE for configuring debounce time on input lines. - Fix a bug when we tried to configure pins for pin controllers without pinconf support. - Initialized debugfs properly so it works. - Initialize the mutex properly and lock around config tampering sections. - Check the return value from get_initial_config() properly. ChangeLog v3->v4: - Export the pin_config_get(), pin_config_set() and pin_config_group() functions. - Drop the entire concept of just getting initial config and keeping track of pin states internally, instead ask the pins what state they are in. Previous idea was plain wrong, if the device cannot keep track of its state, the driver should do it. - Drop the generic configuration layout, it seems this impose too much restriction on some pin controllers, so let them do things the way they want and split off support for generic config as an optional add-on. ChangeLog v4->v5: - Introduce two symmetric driver calls for group configuration, .pin_config_group_[get|set] and corresponding external calls. - Remove generic semantic meanings of return values from config calls, these belong in the generic config patch. Just pass the return value through instead. - Add a debugfs entry "pinconf-groups" to read status from group configuration only, also slam in a per-group debug callback in the pinconf_ops so custom drivers can display something meaningful for their pins. - Fix some dangling newline. - Drop dangling #else clause. - Update documentation to match the above. ChangeLog v5->v6: - Change to using a pin name as parameter for the [get|set]_config() functions, as suggested by Stephen Warren. This is more natural as names will be what a developer has access to in written documentation etc. ChangeLog v6->v7: - Refactor out by-pin and by-name get/set functions, only expose the by-name functions externally, expose the by-pin functions internally. - Show supported pin control functionality in the debugfs pinctrl-devices file. Acked-by: Stephen Warren <swarren@nvidia.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2011-10-19 20:14:33 +04:00
pinctrl: fix and simplify locking There are many problems with the current pinctrl locking: struct pinctrl_dev's gpio_ranges_lock isn't effective; pinctrl_match_gpio_range() only holds this lock while searching for a gpio range, but the found range is return and manipulated after releading the lock. This could allow pinctrl_remove_gpio_range() for that range while it is in use, and the caller may very well delete the range after removing it, causing pinctrl code to touch the now-free range object. Solving this requires the introduction of a higher-level lock, at least a lock per pin controller, which both gpio range registration and pinctrl_get()/put() will acquire. There is missing locking on HW programming; pin controllers may pack the configuration for different pins/groups/config options/... into one register, and hence have to read-modify-write the register. This needs to be protected, but currently isn't. Related, a future change will add a "complete" op to the pin controller drivers, the idea being that each state's programming will be programmed into the pinctrl driver followed by the "complete" call, which may e.g. flush a register cache to HW. For this to work, it must not be possible to interleave the pinctrl driver calls for different devices. As above, solving this requires the introduction of a higher-level lock, at least a lock per pin controller, which will be held for the duration of any pinctrl_enable()/disable() call. However, each pinctrl mapping table entry may affect a different pin controller if necessary. Hence, with a per-pin-controller lock, almost any pinctrl API may need to acquire multiple locks, one per controller. To avoid deadlock, these would need to be acquired in the same order in all cases. This is extremely difficult to implement in the case of pinctrl_get(), which doesn't know which pin controllers to lock until it has parsed the entire mapping table, since it contains somewhat arbitrary data. The simplest solution here is to introduce a single lock that covers all pin controllers at once. This will be acquired by all pinctrl APIs. This then makes struct pinctrl's mutex irrelevant, since that single lock will always be held whenever this mutex is currently held. Signed-off-by: Stephen Warren <swarren@nvidia.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2012-03-03 00:05:44 +04:00
ret = ops->pin_config_group_get(pctldev, selector, config);
unlock:
pinctrl: move subsystem mutex to pinctrl_dev struct This mutex avoids deadlock in case of use of multiple pin controllers. Before this modification, by using a global mutex, deadlock appeared when, for example, a call to pinctrl_pins_show() locked the pinctrl_mutex, called the ops->pin_dbg_show of a particular pin controller. If this pin controller needs I2C access to retrieve configuration information and I2C driver is using pinctrl to drive its pins, a call to pinctrl_select_state() try to lock again pinctrl_mutex which leads to a deadlock. Notice that the mutex grab from the two direction functions was moved into pinctrl_gpio_direction(). For several cases, we can't replace pinctrl_mutex by pctldev->mutex, because at this stage, pctldev is not accessible : - pinctrl_get()/pinctrl_put() - pinctrl_register_maps() So add respectively pinctrl_list_mutex and pinctrl_maps_mutex in order to protect pinctrl_list and pinctrl_maps list instead. Reintroduce pinctrldev_list_mutex in find_pinctrl_by_of_node(), pinctrl_find_and_add_gpio_range() pinctrl_request_gpio(), pinctrl_free_gpio(), pinctrl_gpio_direction(), pinctrl_devices_show(), pinctrl_register() and pinctrl_unregister() to protect pinctrldev_list. Changes v2->v3: - Fix a missing EXPORT_SYMBOL_GPL() for pinctrl_select_state(). Changes v1->v2: - pinctrl_select_state_locked() is removed, all lock mechanism is located inside pinctrl_select_state(). When parsing the state->setting list, take the per-pin-controller driver lock. (Patrice). - Introduce pinctrldev_list_mutex to protect pinctrldev_list in all functions which parse or modify pictrldev_list. (Patrice). - move find_pinctrl_by_of_node() from pinctrl/devicetree.c to pinctrl/core.c in order to protect pinctrldev_list. (Patrice). - Sink mutex:es into some functions and remove some _locked variants down to where the lists are actually accessed to make things simpler. (Linus) - Drop *all* mutexes completely from pinctrl_lookup_state() and pinctrl_select_state() - no relevant mutex was taken and it was unclear what this was protecting against. (Linus) Reported by : Seraphin Bonnaffe <seraphin.bonnaffe@stericsson.com> Signed-off-by: Patrice Chotard <patrice.chotard@st.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2013-04-11 13:01:27 +04:00
mutex_unlock(&pctldev->mutex);
pinctrl: fix and simplify locking There are many problems with the current pinctrl locking: struct pinctrl_dev's gpio_ranges_lock isn't effective; pinctrl_match_gpio_range() only holds this lock while searching for a gpio range, but the found range is return and manipulated after releading the lock. This could allow pinctrl_remove_gpio_range() for that range while it is in use, and the caller may very well delete the range after removing it, causing pinctrl code to touch the now-free range object. Solving this requires the introduction of a higher-level lock, at least a lock per pin controller, which both gpio range registration and pinctrl_get()/put() will acquire. There is missing locking on HW programming; pin controllers may pack the configuration for different pins/groups/config options/... into one register, and hence have to read-modify-write the register. This needs to be protected, but currently isn't. Related, a future change will add a "complete" op to the pin controller drivers, the idea being that each state's programming will be programmed into the pinctrl driver followed by the "complete" call, which may e.g. flush a register cache to HW. For this to work, it must not be possible to interleave the pinctrl driver calls for different devices. As above, solving this requires the introduction of a higher-level lock, at least a lock per pin controller, which will be held for the duration of any pinctrl_enable()/disable() call. However, each pinctrl mapping table entry may affect a different pin controller if necessary. Hence, with a per-pin-controller lock, almost any pinctrl API may need to acquire multiple locks, one per controller. To avoid deadlock, these would need to be acquired in the same order in all cases. This is extremely difficult to implement in the case of pinctrl_get(), which doesn't know which pin controllers to lock until it has parsed the entire mapping table, since it contains somewhat arbitrary data. The simplest solution here is to introduce a single lock that covers all pin controllers at once. This will be acquired by all pinctrl APIs. This then makes struct pinctrl's mutex irrelevant, since that single lock will always be held whenever this mutex is currently held. Signed-off-by: Stephen Warren <swarren@nvidia.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2012-03-03 00:05:44 +04:00
return ret;
pinctrl: add a pin config interface This add per-pin and per-group pin config interfaces for biasing, driving and other such electronic properties. The details of passed configurations are passed in an opaque unsigned long which may be dereferences to integer types, structs or lists on either side of the configuration interface. ChangeLog v1->v2: - Clear split of terminology: we now have pin controllers, and those may support two interfaces using vtables: pin multiplexing and pin configuration. - Break out pin configuration to its own C file, controllers may implement only config without mux, and vice versa, so keep each sub-functionality of pin controllers separate. Introduce CONFIG_PINCONF in Kconfig. - Implement some core logic around pin configuration in the pinconf.c file. - Remove UNKNOWN config states, these were just surplus baggage. - Remove FLOAT config state - HIGH_IMPEDANCE should be enough for everyone. - PIN_CONFIG_POWER_SOURCE added to handle switching the power supply for the pin logic between different sources - Explicit DISABLE config enums to turn schmitt-trigger, wakeup etc OFF. - Update documentation to reflect all the recent reasoning. ChangeLog v2->v3: - Twist API around to pass around arrays of config tuples instead of (param, value) pairs everywhere. - Explicit drive strength semantics for push/pull and similar drive modes, this shall be the number of drive stages vs nominal load impedance, which should match the actual electronics used in push/pull CMOS or TTY totempoles. - Drop load capacitance configuration - I probably don't know what I'm doing here so leave it out. - Drop PIN_CONFIG_INPUT_SCHMITT_OFF, instead the argument zero to PIN_CONFIG_INPUT_SCHMITT turns schmitt trigger off. - Drop PIN_CONFIG_NORMAL_POWER_MODE and have a well defined argument to PIN_CONFIG_LOW_POWER_MODE to get out of it instead. - Drop PIN_CONFIG_WAKEUP_ENABLE/DISABLE and just use PIN_CONFIG_WAKEUP with defined value zero to turn wakeup off. - Add PIN_CONFIG_INPUT_DEBOUNCE for configuring debounce time on input lines. - Fix a bug when we tried to configure pins for pin controllers without pinconf support. - Initialized debugfs properly so it works. - Initialize the mutex properly and lock around config tampering sections. - Check the return value from get_initial_config() properly. ChangeLog v3->v4: - Export the pin_config_get(), pin_config_set() and pin_config_group() functions. - Drop the entire concept of just getting initial config and keeping track of pin states internally, instead ask the pins what state they are in. Previous idea was plain wrong, if the device cannot keep track of its state, the driver should do it. - Drop the generic configuration layout, it seems this impose too much restriction on some pin controllers, so let them do things the way they want and split off support for generic config as an optional add-on. ChangeLog v4->v5: - Introduce two symmetric driver calls for group configuration, .pin_config_group_[get|set] and corresponding external calls. - Remove generic semantic meanings of return values from config calls, these belong in the generic config patch. Just pass the return value through instead. - Add a debugfs entry "pinconf-groups" to read status from group configuration only, also slam in a per-group debug callback in the pinconf_ops so custom drivers can display something meaningful for their pins. - Fix some dangling newline. - Drop dangling #else clause. - Update documentation to match the above. ChangeLog v5->v6: - Change to using a pin name as parameter for the [get|set]_config() functions, as suggested by Stephen Warren. This is more natural as names will be what a developer has access to in written documentation etc. ChangeLog v6->v7: - Refactor out by-pin and by-name get/set functions, only expose the by-name functions externally, expose the by-pin functions internally. - Show supported pin control functionality in the debugfs pinctrl-devices file. Acked-by: Stephen Warren <swarren@nvidia.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2011-10-19 20:14:33 +04:00
}
EXPORT_SYMBOL(pin_config_group_get);
int pin_config_group_set(const char *dev_name, const char *pin_group,
pinctrl: add a pin config interface This add per-pin and per-group pin config interfaces for biasing, driving and other such electronic properties. The details of passed configurations are passed in an opaque unsigned long which may be dereferences to integer types, structs or lists on either side of the configuration interface. ChangeLog v1->v2: - Clear split of terminology: we now have pin controllers, and those may support two interfaces using vtables: pin multiplexing and pin configuration. - Break out pin configuration to its own C file, controllers may implement only config without mux, and vice versa, so keep each sub-functionality of pin controllers separate. Introduce CONFIG_PINCONF in Kconfig. - Implement some core logic around pin configuration in the pinconf.c file. - Remove UNKNOWN config states, these were just surplus baggage. - Remove FLOAT config state - HIGH_IMPEDANCE should be enough for everyone. - PIN_CONFIG_POWER_SOURCE added to handle switching the power supply for the pin logic between different sources - Explicit DISABLE config enums to turn schmitt-trigger, wakeup etc OFF. - Update documentation to reflect all the recent reasoning. ChangeLog v2->v3: - Twist API around to pass around arrays of config tuples instead of (param, value) pairs everywhere. - Explicit drive strength semantics for push/pull and similar drive modes, this shall be the number of drive stages vs nominal load impedance, which should match the actual electronics used in push/pull CMOS or TTY totempoles. - Drop load capacitance configuration - I probably don't know what I'm doing here so leave it out. - Drop PIN_CONFIG_INPUT_SCHMITT_OFF, instead the argument zero to PIN_CONFIG_INPUT_SCHMITT turns schmitt trigger off. - Drop PIN_CONFIG_NORMAL_POWER_MODE and have a well defined argument to PIN_CONFIG_LOW_POWER_MODE to get out of it instead. - Drop PIN_CONFIG_WAKEUP_ENABLE/DISABLE and just use PIN_CONFIG_WAKEUP with defined value zero to turn wakeup off. - Add PIN_CONFIG_INPUT_DEBOUNCE for configuring debounce time on input lines. - Fix a bug when we tried to configure pins for pin controllers without pinconf support. - Initialized debugfs properly so it works. - Initialize the mutex properly and lock around config tampering sections. - Check the return value from get_initial_config() properly. ChangeLog v3->v4: - Export the pin_config_get(), pin_config_set() and pin_config_group() functions. - Drop the entire concept of just getting initial config and keeping track of pin states internally, instead ask the pins what state they are in. Previous idea was plain wrong, if the device cannot keep track of its state, the driver should do it. - Drop the generic configuration layout, it seems this impose too much restriction on some pin controllers, so let them do things the way they want and split off support for generic config as an optional add-on. ChangeLog v4->v5: - Introduce two symmetric driver calls for group configuration, .pin_config_group_[get|set] and corresponding external calls. - Remove generic semantic meanings of return values from config calls, these belong in the generic config patch. Just pass the return value through instead. - Add a debugfs entry "pinconf-groups" to read status from group configuration only, also slam in a per-group debug callback in the pinconf_ops so custom drivers can display something meaningful for their pins. - Fix some dangling newline. - Drop dangling #else clause. - Update documentation to match the above. ChangeLog v5->v6: - Change to using a pin name as parameter for the [get|set]_config() functions, as suggested by Stephen Warren. This is more natural as names will be what a developer has access to in written documentation etc. ChangeLog v6->v7: - Refactor out by-pin and by-name get/set functions, only expose the by-name functions externally, expose the by-pin functions internally. - Show supported pin control functionality in the debugfs pinctrl-devices file. Acked-by: Stephen Warren <swarren@nvidia.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2011-10-19 20:14:33 +04:00
unsigned long config)
{
struct pinctrl_dev *pctldev;
const struct pinconf_ops *ops;
const struct pinctrl_ops *pctlops;
pinctrl: add a pin config interface This add per-pin and per-group pin config interfaces for biasing, driving and other such electronic properties. The details of passed configurations are passed in an opaque unsigned long which may be dereferences to integer types, structs or lists on either side of the configuration interface. ChangeLog v1->v2: - Clear split of terminology: we now have pin controllers, and those may support two interfaces using vtables: pin multiplexing and pin configuration. - Break out pin configuration to its own C file, controllers may implement only config without mux, and vice versa, so keep each sub-functionality of pin controllers separate. Introduce CONFIG_PINCONF in Kconfig. - Implement some core logic around pin configuration in the pinconf.c file. - Remove UNKNOWN config states, these were just surplus baggage. - Remove FLOAT config state - HIGH_IMPEDANCE should be enough for everyone. - PIN_CONFIG_POWER_SOURCE added to handle switching the power supply for the pin logic between different sources - Explicit DISABLE config enums to turn schmitt-trigger, wakeup etc OFF. - Update documentation to reflect all the recent reasoning. ChangeLog v2->v3: - Twist API around to pass around arrays of config tuples instead of (param, value) pairs everywhere. - Explicit drive strength semantics for push/pull and similar drive modes, this shall be the number of drive stages vs nominal load impedance, which should match the actual electronics used in push/pull CMOS or TTY totempoles. - Drop load capacitance configuration - I probably don't know what I'm doing here so leave it out. - Drop PIN_CONFIG_INPUT_SCHMITT_OFF, instead the argument zero to PIN_CONFIG_INPUT_SCHMITT turns schmitt trigger off. - Drop PIN_CONFIG_NORMAL_POWER_MODE and have a well defined argument to PIN_CONFIG_LOW_POWER_MODE to get out of it instead. - Drop PIN_CONFIG_WAKEUP_ENABLE/DISABLE and just use PIN_CONFIG_WAKEUP with defined value zero to turn wakeup off. - Add PIN_CONFIG_INPUT_DEBOUNCE for configuring debounce time on input lines. - Fix a bug when we tried to configure pins for pin controllers without pinconf support. - Initialized debugfs properly so it works. - Initialize the mutex properly and lock around config tampering sections. - Check the return value from get_initial_config() properly. ChangeLog v3->v4: - Export the pin_config_get(), pin_config_set() and pin_config_group() functions. - Drop the entire concept of just getting initial config and keeping track of pin states internally, instead ask the pins what state they are in. Previous idea was plain wrong, if the device cannot keep track of its state, the driver should do it. - Drop the generic configuration layout, it seems this impose too much restriction on some pin controllers, so let them do things the way they want and split off support for generic config as an optional add-on. ChangeLog v4->v5: - Introduce two symmetric driver calls for group configuration, .pin_config_group_[get|set] and corresponding external calls. - Remove generic semantic meanings of return values from config calls, these belong in the generic config patch. Just pass the return value through instead. - Add a debugfs entry "pinconf-groups" to read status from group configuration only, also slam in a per-group debug callback in the pinconf_ops so custom drivers can display something meaningful for their pins. - Fix some dangling newline. - Drop dangling #else clause. - Update documentation to match the above. ChangeLog v5->v6: - Change to using a pin name as parameter for the [get|set]_config() functions, as suggested by Stephen Warren. This is more natural as names will be what a developer has access to in written documentation etc. ChangeLog v6->v7: - Refactor out by-pin and by-name get/set functions, only expose the by-name functions externally, expose the by-pin functions internally. - Show supported pin control functionality in the debugfs pinctrl-devices file. Acked-by: Stephen Warren <swarren@nvidia.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2011-10-19 20:14:33 +04:00
int selector;
const unsigned *pins;
unsigned num_pins;
int ret;
int i;
pctldev = get_pinctrl_dev_from_devname(dev_name);
pinctrl: fix and simplify locking There are many problems with the current pinctrl locking: struct pinctrl_dev's gpio_ranges_lock isn't effective; pinctrl_match_gpio_range() only holds this lock while searching for a gpio range, but the found range is return and manipulated after releading the lock. This could allow pinctrl_remove_gpio_range() for that range while it is in use, and the caller may very well delete the range after removing it, causing pinctrl code to touch the now-free range object. Solving this requires the introduction of a higher-level lock, at least a lock per pin controller, which both gpio range registration and pinctrl_get()/put() will acquire. There is missing locking on HW programming; pin controllers may pack the configuration for different pins/groups/config options/... into one register, and hence have to read-modify-write the register. This needs to be protected, but currently isn't. Related, a future change will add a "complete" op to the pin controller drivers, the idea being that each state's programming will be programmed into the pinctrl driver followed by the "complete" call, which may e.g. flush a register cache to HW. For this to work, it must not be possible to interleave the pinctrl driver calls for different devices. As above, solving this requires the introduction of a higher-level lock, at least a lock per pin controller, which will be held for the duration of any pinctrl_enable()/disable() call. However, each pinctrl mapping table entry may affect a different pin controller if necessary. Hence, with a per-pin-controller lock, almost any pinctrl API may need to acquire multiple locks, one per controller. To avoid deadlock, these would need to be acquired in the same order in all cases. This is extremely difficult to implement in the case of pinctrl_get(), which doesn't know which pin controllers to lock until it has parsed the entire mapping table, since it contains somewhat arbitrary data. The simplest solution here is to introduce a single lock that covers all pin controllers at once. This will be acquired by all pinctrl APIs. This then makes struct pinctrl's mutex irrelevant, since that single lock will always be held whenever this mutex is currently held. Signed-off-by: Stephen Warren <swarren@nvidia.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2012-03-03 00:05:44 +04:00
if (!pctldev) {
ret = -EINVAL;
pinctrl: move subsystem mutex to pinctrl_dev struct This mutex avoids deadlock in case of use of multiple pin controllers. Before this modification, by using a global mutex, deadlock appeared when, for example, a call to pinctrl_pins_show() locked the pinctrl_mutex, called the ops->pin_dbg_show of a particular pin controller. If this pin controller needs I2C access to retrieve configuration information and I2C driver is using pinctrl to drive its pins, a call to pinctrl_select_state() try to lock again pinctrl_mutex which leads to a deadlock. Notice that the mutex grab from the two direction functions was moved into pinctrl_gpio_direction(). For several cases, we can't replace pinctrl_mutex by pctldev->mutex, because at this stage, pctldev is not accessible : - pinctrl_get()/pinctrl_put() - pinctrl_register_maps() So add respectively pinctrl_list_mutex and pinctrl_maps_mutex in order to protect pinctrl_list and pinctrl_maps list instead. Reintroduce pinctrldev_list_mutex in find_pinctrl_by_of_node(), pinctrl_find_and_add_gpio_range() pinctrl_request_gpio(), pinctrl_free_gpio(), pinctrl_gpio_direction(), pinctrl_devices_show(), pinctrl_register() and pinctrl_unregister() to protect pinctrldev_list. Changes v2->v3: - Fix a missing EXPORT_SYMBOL_GPL() for pinctrl_select_state(). Changes v1->v2: - pinctrl_select_state_locked() is removed, all lock mechanism is located inside pinctrl_select_state(). When parsing the state->setting list, take the per-pin-controller driver lock. (Patrice). - Introduce pinctrldev_list_mutex to protect pinctrldev_list in all functions which parse or modify pictrldev_list. (Patrice). - move find_pinctrl_by_of_node() from pinctrl/devicetree.c to pinctrl/core.c in order to protect pinctrldev_list. (Patrice). - Sink mutex:es into some functions and remove some _locked variants down to where the lists are actually accessed to make things simpler. (Linus) - Drop *all* mutexes completely from pinctrl_lookup_state() and pinctrl_select_state() - no relevant mutex was taken and it was unclear what this was protecting against. (Linus) Reported by : Seraphin Bonnaffe <seraphin.bonnaffe@stericsson.com> Signed-off-by: Patrice Chotard <patrice.chotard@st.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2013-04-11 13:01:27 +04:00
return ret;
pinctrl: fix and simplify locking There are many problems with the current pinctrl locking: struct pinctrl_dev's gpio_ranges_lock isn't effective; pinctrl_match_gpio_range() only holds this lock while searching for a gpio range, but the found range is return and manipulated after releading the lock. This could allow pinctrl_remove_gpio_range() for that range while it is in use, and the caller may very well delete the range after removing it, causing pinctrl code to touch the now-free range object. Solving this requires the introduction of a higher-level lock, at least a lock per pin controller, which both gpio range registration and pinctrl_get()/put() will acquire. There is missing locking on HW programming; pin controllers may pack the configuration for different pins/groups/config options/... into one register, and hence have to read-modify-write the register. This needs to be protected, but currently isn't. Related, a future change will add a "complete" op to the pin controller drivers, the idea being that each state's programming will be programmed into the pinctrl driver followed by the "complete" call, which may e.g. flush a register cache to HW. For this to work, it must not be possible to interleave the pinctrl driver calls for different devices. As above, solving this requires the introduction of a higher-level lock, at least a lock per pin controller, which will be held for the duration of any pinctrl_enable()/disable() call. However, each pinctrl mapping table entry may affect a different pin controller if necessary. Hence, with a per-pin-controller lock, almost any pinctrl API may need to acquire multiple locks, one per controller. To avoid deadlock, these would need to be acquired in the same order in all cases. This is extremely difficult to implement in the case of pinctrl_get(), which doesn't know which pin controllers to lock until it has parsed the entire mapping table, since it contains somewhat arbitrary data. The simplest solution here is to introduce a single lock that covers all pin controllers at once. This will be acquired by all pinctrl APIs. This then makes struct pinctrl's mutex irrelevant, since that single lock will always be held whenever this mutex is currently held. Signed-off-by: Stephen Warren <swarren@nvidia.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2012-03-03 00:05:44 +04:00
}
pinctrl: move subsystem mutex to pinctrl_dev struct This mutex avoids deadlock in case of use of multiple pin controllers. Before this modification, by using a global mutex, deadlock appeared when, for example, a call to pinctrl_pins_show() locked the pinctrl_mutex, called the ops->pin_dbg_show of a particular pin controller. If this pin controller needs I2C access to retrieve configuration information and I2C driver is using pinctrl to drive its pins, a call to pinctrl_select_state() try to lock again pinctrl_mutex which leads to a deadlock. Notice that the mutex grab from the two direction functions was moved into pinctrl_gpio_direction(). For several cases, we can't replace pinctrl_mutex by pctldev->mutex, because at this stage, pctldev is not accessible : - pinctrl_get()/pinctrl_put() - pinctrl_register_maps() So add respectively pinctrl_list_mutex and pinctrl_maps_mutex in order to protect pinctrl_list and pinctrl_maps list instead. Reintroduce pinctrldev_list_mutex in find_pinctrl_by_of_node(), pinctrl_find_and_add_gpio_range() pinctrl_request_gpio(), pinctrl_free_gpio(), pinctrl_gpio_direction(), pinctrl_devices_show(), pinctrl_register() and pinctrl_unregister() to protect pinctrldev_list. Changes v2->v3: - Fix a missing EXPORT_SYMBOL_GPL() for pinctrl_select_state(). Changes v1->v2: - pinctrl_select_state_locked() is removed, all lock mechanism is located inside pinctrl_select_state(). When parsing the state->setting list, take the per-pin-controller driver lock. (Patrice). - Introduce pinctrldev_list_mutex to protect pinctrldev_list in all functions which parse or modify pictrldev_list. (Patrice). - move find_pinctrl_by_of_node() from pinctrl/devicetree.c to pinctrl/core.c in order to protect pinctrldev_list. (Patrice). - Sink mutex:es into some functions and remove some _locked variants down to where the lists are actually accessed to make things simpler. (Linus) - Drop *all* mutexes completely from pinctrl_lookup_state() and pinctrl_select_state() - no relevant mutex was taken and it was unclear what this was protecting against. (Linus) Reported by : Seraphin Bonnaffe <seraphin.bonnaffe@stericsson.com> Signed-off-by: Patrice Chotard <patrice.chotard@st.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2013-04-11 13:01:27 +04:00
mutex_lock(&pctldev->mutex);
ops = pctldev->desc->confops;
pctlops = pctldev->desc->pctlops;
pinctrl: add a pin config interface This add per-pin and per-group pin config interfaces for biasing, driving and other such electronic properties. The details of passed configurations are passed in an opaque unsigned long which may be dereferences to integer types, structs or lists on either side of the configuration interface. ChangeLog v1->v2: - Clear split of terminology: we now have pin controllers, and those may support two interfaces using vtables: pin multiplexing and pin configuration. - Break out pin configuration to its own C file, controllers may implement only config without mux, and vice versa, so keep each sub-functionality of pin controllers separate. Introduce CONFIG_PINCONF in Kconfig. - Implement some core logic around pin configuration in the pinconf.c file. - Remove UNKNOWN config states, these were just surplus baggage. - Remove FLOAT config state - HIGH_IMPEDANCE should be enough for everyone. - PIN_CONFIG_POWER_SOURCE added to handle switching the power supply for the pin logic between different sources - Explicit DISABLE config enums to turn schmitt-trigger, wakeup etc OFF. - Update documentation to reflect all the recent reasoning. ChangeLog v2->v3: - Twist API around to pass around arrays of config tuples instead of (param, value) pairs everywhere. - Explicit drive strength semantics for push/pull and similar drive modes, this shall be the number of drive stages vs nominal load impedance, which should match the actual electronics used in push/pull CMOS or TTY totempoles. - Drop load capacitance configuration - I probably don't know what I'm doing here so leave it out. - Drop PIN_CONFIG_INPUT_SCHMITT_OFF, instead the argument zero to PIN_CONFIG_INPUT_SCHMITT turns schmitt trigger off. - Drop PIN_CONFIG_NORMAL_POWER_MODE and have a well defined argument to PIN_CONFIG_LOW_POWER_MODE to get out of it instead. - Drop PIN_CONFIG_WAKEUP_ENABLE/DISABLE and just use PIN_CONFIG_WAKEUP with defined value zero to turn wakeup off. - Add PIN_CONFIG_INPUT_DEBOUNCE for configuring debounce time on input lines. - Fix a bug when we tried to configure pins for pin controllers without pinconf support. - Initialized debugfs properly so it works. - Initialize the mutex properly and lock around config tampering sections. - Check the return value from get_initial_config() properly. ChangeLog v3->v4: - Export the pin_config_get(), pin_config_set() and pin_config_group() functions. - Drop the entire concept of just getting initial config and keeping track of pin states internally, instead ask the pins what state they are in. Previous idea was plain wrong, if the device cannot keep track of its state, the driver should do it. - Drop the generic configuration layout, it seems this impose too much restriction on some pin controllers, so let them do things the way they want and split off support for generic config as an optional add-on. ChangeLog v4->v5: - Introduce two symmetric driver calls for group configuration, .pin_config_group_[get|set] and corresponding external calls. - Remove generic semantic meanings of return values from config calls, these belong in the generic config patch. Just pass the return value through instead. - Add a debugfs entry "pinconf-groups" to read status from group configuration only, also slam in a per-group debug callback in the pinconf_ops so custom drivers can display something meaningful for their pins. - Fix some dangling newline. - Drop dangling #else clause. - Update documentation to match the above. ChangeLog v5->v6: - Change to using a pin name as parameter for the [get|set]_config() functions, as suggested by Stephen Warren. This is more natural as names will be what a developer has access to in written documentation etc. ChangeLog v6->v7: - Refactor out by-pin and by-name get/set functions, only expose the by-name functions externally, expose the by-pin functions internally. - Show supported pin control functionality in the debugfs pinctrl-devices file. Acked-by: Stephen Warren <swarren@nvidia.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2011-10-19 20:14:33 +04:00
if (!ops || (!ops->pin_config_group_set && !ops->pin_config_set)) {
dev_err(pctldev->dev, "cannot configure pin group, missing "
pinctrl: add a pin config interface This add per-pin and per-group pin config interfaces for biasing, driving and other such electronic properties. The details of passed configurations are passed in an opaque unsigned long which may be dereferences to integer types, structs or lists on either side of the configuration interface. ChangeLog v1->v2: - Clear split of terminology: we now have pin controllers, and those may support two interfaces using vtables: pin multiplexing and pin configuration. - Break out pin configuration to its own C file, controllers may implement only config without mux, and vice versa, so keep each sub-functionality of pin controllers separate. Introduce CONFIG_PINCONF in Kconfig. - Implement some core logic around pin configuration in the pinconf.c file. - Remove UNKNOWN config states, these were just surplus baggage. - Remove FLOAT config state - HIGH_IMPEDANCE should be enough for everyone. - PIN_CONFIG_POWER_SOURCE added to handle switching the power supply for the pin logic between different sources - Explicit DISABLE config enums to turn schmitt-trigger, wakeup etc OFF. - Update documentation to reflect all the recent reasoning. ChangeLog v2->v3: - Twist API around to pass around arrays of config tuples instead of (param, value) pairs everywhere. - Explicit drive strength semantics for push/pull and similar drive modes, this shall be the number of drive stages vs nominal load impedance, which should match the actual electronics used in push/pull CMOS or TTY totempoles. - Drop load capacitance configuration - I probably don't know what I'm doing here so leave it out. - Drop PIN_CONFIG_INPUT_SCHMITT_OFF, instead the argument zero to PIN_CONFIG_INPUT_SCHMITT turns schmitt trigger off. - Drop PIN_CONFIG_NORMAL_POWER_MODE and have a well defined argument to PIN_CONFIG_LOW_POWER_MODE to get out of it instead. - Drop PIN_CONFIG_WAKEUP_ENABLE/DISABLE and just use PIN_CONFIG_WAKEUP with defined value zero to turn wakeup off. - Add PIN_CONFIG_INPUT_DEBOUNCE for configuring debounce time on input lines. - Fix a bug when we tried to configure pins for pin controllers without pinconf support. - Initialized debugfs properly so it works. - Initialize the mutex properly and lock around config tampering sections. - Check the return value from get_initial_config() properly. ChangeLog v3->v4: - Export the pin_config_get(), pin_config_set() and pin_config_group() functions. - Drop the entire concept of just getting initial config and keeping track of pin states internally, instead ask the pins what state they are in. Previous idea was plain wrong, if the device cannot keep track of its state, the driver should do it. - Drop the generic configuration layout, it seems this impose too much restriction on some pin controllers, so let them do things the way they want and split off support for generic config as an optional add-on. ChangeLog v4->v5: - Introduce two symmetric driver calls for group configuration, .pin_config_group_[get|set] and corresponding external calls. - Remove generic semantic meanings of return values from config calls, these belong in the generic config patch. Just pass the return value through instead. - Add a debugfs entry "pinconf-groups" to read status from group configuration only, also slam in a per-group debug callback in the pinconf_ops so custom drivers can display something meaningful for their pins. - Fix some dangling newline. - Drop dangling #else clause. - Update documentation to match the above. ChangeLog v5->v6: - Change to using a pin name as parameter for the [get|set]_config() functions, as suggested by Stephen Warren. This is more natural as names will be what a developer has access to in written documentation etc. ChangeLog v6->v7: - Refactor out by-pin and by-name get/set functions, only expose the by-name functions externally, expose the by-pin functions internally. - Show supported pin control functionality in the debugfs pinctrl-devices file. Acked-by: Stephen Warren <swarren@nvidia.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2011-10-19 20:14:33 +04:00
"config function in driver\n");
pinctrl: fix and simplify locking There are many problems with the current pinctrl locking: struct pinctrl_dev's gpio_ranges_lock isn't effective; pinctrl_match_gpio_range() only holds this lock while searching for a gpio range, but the found range is return and manipulated after releading the lock. This could allow pinctrl_remove_gpio_range() for that range while it is in use, and the caller may very well delete the range after removing it, causing pinctrl code to touch the now-free range object. Solving this requires the introduction of a higher-level lock, at least a lock per pin controller, which both gpio range registration and pinctrl_get()/put() will acquire. There is missing locking on HW programming; pin controllers may pack the configuration for different pins/groups/config options/... into one register, and hence have to read-modify-write the register. This needs to be protected, but currently isn't. Related, a future change will add a "complete" op to the pin controller drivers, the idea being that each state's programming will be programmed into the pinctrl driver followed by the "complete" call, which may e.g. flush a register cache to HW. For this to work, it must not be possible to interleave the pinctrl driver calls for different devices. As above, solving this requires the introduction of a higher-level lock, at least a lock per pin controller, which will be held for the duration of any pinctrl_enable()/disable() call. However, each pinctrl mapping table entry may affect a different pin controller if necessary. Hence, with a per-pin-controller lock, almost any pinctrl API may need to acquire multiple locks, one per controller. To avoid deadlock, these would need to be acquired in the same order in all cases. This is extremely difficult to implement in the case of pinctrl_get(), which doesn't know which pin controllers to lock until it has parsed the entire mapping table, since it contains somewhat arbitrary data. The simplest solution here is to introduce a single lock that covers all pin controllers at once. This will be acquired by all pinctrl APIs. This then makes struct pinctrl's mutex irrelevant, since that single lock will always be held whenever this mutex is currently held. Signed-off-by: Stephen Warren <swarren@nvidia.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2012-03-03 00:05:44 +04:00
ret = -EINVAL;
goto unlock;
pinctrl: add a pin config interface This add per-pin and per-group pin config interfaces for biasing, driving and other such electronic properties. The details of passed configurations are passed in an opaque unsigned long which may be dereferences to integer types, structs or lists on either side of the configuration interface. ChangeLog v1->v2: - Clear split of terminology: we now have pin controllers, and those may support two interfaces using vtables: pin multiplexing and pin configuration. - Break out pin configuration to its own C file, controllers may implement only config without mux, and vice versa, so keep each sub-functionality of pin controllers separate. Introduce CONFIG_PINCONF in Kconfig. - Implement some core logic around pin configuration in the pinconf.c file. - Remove UNKNOWN config states, these were just surplus baggage. - Remove FLOAT config state - HIGH_IMPEDANCE should be enough for everyone. - PIN_CONFIG_POWER_SOURCE added to handle switching the power supply for the pin logic between different sources - Explicit DISABLE config enums to turn schmitt-trigger, wakeup etc OFF. - Update documentation to reflect all the recent reasoning. ChangeLog v2->v3: - Twist API around to pass around arrays of config tuples instead of (param, value) pairs everywhere. - Explicit drive strength semantics for push/pull and similar drive modes, this shall be the number of drive stages vs nominal load impedance, which should match the actual electronics used in push/pull CMOS or TTY totempoles. - Drop load capacitance configuration - I probably don't know what I'm doing here so leave it out. - Drop PIN_CONFIG_INPUT_SCHMITT_OFF, instead the argument zero to PIN_CONFIG_INPUT_SCHMITT turns schmitt trigger off. - Drop PIN_CONFIG_NORMAL_POWER_MODE and have a well defined argument to PIN_CONFIG_LOW_POWER_MODE to get out of it instead. - Drop PIN_CONFIG_WAKEUP_ENABLE/DISABLE and just use PIN_CONFIG_WAKEUP with defined value zero to turn wakeup off. - Add PIN_CONFIG_INPUT_DEBOUNCE for configuring debounce time on input lines. - Fix a bug when we tried to configure pins for pin controllers without pinconf support. - Initialized debugfs properly so it works. - Initialize the mutex properly and lock around config tampering sections. - Check the return value from get_initial_config() properly. ChangeLog v3->v4: - Export the pin_config_get(), pin_config_set() and pin_config_group() functions. - Drop the entire concept of just getting initial config and keeping track of pin states internally, instead ask the pins what state they are in. Previous idea was plain wrong, if the device cannot keep track of its state, the driver should do it. - Drop the generic configuration layout, it seems this impose too much restriction on some pin controllers, so let them do things the way they want and split off support for generic config as an optional add-on. ChangeLog v4->v5: - Introduce two symmetric driver calls for group configuration, .pin_config_group_[get|set] and corresponding external calls. - Remove generic semantic meanings of return values from config calls, these belong in the generic config patch. Just pass the return value through instead. - Add a debugfs entry "pinconf-groups" to read status from group configuration only, also slam in a per-group debug callback in the pinconf_ops so custom drivers can display something meaningful for their pins. - Fix some dangling newline. - Drop dangling #else clause. - Update documentation to match the above. ChangeLog v5->v6: - Change to using a pin name as parameter for the [get|set]_config() functions, as suggested by Stephen Warren. This is more natural as names will be what a developer has access to in written documentation etc. ChangeLog v6->v7: - Refactor out by-pin and by-name get/set functions, only expose the by-name functions externally, expose the by-pin functions internally. - Show supported pin control functionality in the debugfs pinctrl-devices file. Acked-by: Stephen Warren <swarren@nvidia.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2011-10-19 20:14:33 +04:00
}
selector = pinctrl_get_group_selector(pctldev, pin_group);
pinctrl: fix and simplify locking There are many problems with the current pinctrl locking: struct pinctrl_dev's gpio_ranges_lock isn't effective; pinctrl_match_gpio_range() only holds this lock while searching for a gpio range, but the found range is return and manipulated after releading the lock. This could allow pinctrl_remove_gpio_range() for that range while it is in use, and the caller may very well delete the range after removing it, causing pinctrl code to touch the now-free range object. Solving this requires the introduction of a higher-level lock, at least a lock per pin controller, which both gpio range registration and pinctrl_get()/put() will acquire. There is missing locking on HW programming; pin controllers may pack the configuration for different pins/groups/config options/... into one register, and hence have to read-modify-write the register. This needs to be protected, but currently isn't. Related, a future change will add a "complete" op to the pin controller drivers, the idea being that each state's programming will be programmed into the pinctrl driver followed by the "complete" call, which may e.g. flush a register cache to HW. For this to work, it must not be possible to interleave the pinctrl driver calls for different devices. As above, solving this requires the introduction of a higher-level lock, at least a lock per pin controller, which will be held for the duration of any pinctrl_enable()/disable() call. However, each pinctrl mapping table entry may affect a different pin controller if necessary. Hence, with a per-pin-controller lock, almost any pinctrl API may need to acquire multiple locks, one per controller. To avoid deadlock, these would need to be acquired in the same order in all cases. This is extremely difficult to implement in the case of pinctrl_get(), which doesn't know which pin controllers to lock until it has parsed the entire mapping table, since it contains somewhat arbitrary data. The simplest solution here is to introduce a single lock that covers all pin controllers at once. This will be acquired by all pinctrl APIs. This then makes struct pinctrl's mutex irrelevant, since that single lock will always be held whenever this mutex is currently held. Signed-off-by: Stephen Warren <swarren@nvidia.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2012-03-03 00:05:44 +04:00
if (selector < 0) {
ret = selector;
goto unlock;
}
pinctrl: add a pin config interface This add per-pin and per-group pin config interfaces for biasing, driving and other such electronic properties. The details of passed configurations are passed in an opaque unsigned long which may be dereferences to integer types, structs or lists on either side of the configuration interface. ChangeLog v1->v2: - Clear split of terminology: we now have pin controllers, and those may support two interfaces using vtables: pin multiplexing and pin configuration. - Break out pin configuration to its own C file, controllers may implement only config without mux, and vice versa, so keep each sub-functionality of pin controllers separate. Introduce CONFIG_PINCONF in Kconfig. - Implement some core logic around pin configuration in the pinconf.c file. - Remove UNKNOWN config states, these were just surplus baggage. - Remove FLOAT config state - HIGH_IMPEDANCE should be enough for everyone. - PIN_CONFIG_POWER_SOURCE added to handle switching the power supply for the pin logic between different sources - Explicit DISABLE config enums to turn schmitt-trigger, wakeup etc OFF. - Update documentation to reflect all the recent reasoning. ChangeLog v2->v3: - Twist API around to pass around arrays of config tuples instead of (param, value) pairs everywhere. - Explicit drive strength semantics for push/pull and similar drive modes, this shall be the number of drive stages vs nominal load impedance, which should match the actual electronics used in push/pull CMOS or TTY totempoles. - Drop load capacitance configuration - I probably don't know what I'm doing here so leave it out. - Drop PIN_CONFIG_INPUT_SCHMITT_OFF, instead the argument zero to PIN_CONFIG_INPUT_SCHMITT turns schmitt trigger off. - Drop PIN_CONFIG_NORMAL_POWER_MODE and have a well defined argument to PIN_CONFIG_LOW_POWER_MODE to get out of it instead. - Drop PIN_CONFIG_WAKEUP_ENABLE/DISABLE and just use PIN_CONFIG_WAKEUP with defined value zero to turn wakeup off. - Add PIN_CONFIG_INPUT_DEBOUNCE for configuring debounce time on input lines. - Fix a bug when we tried to configure pins for pin controllers without pinconf support. - Initialized debugfs properly so it works. - Initialize the mutex properly and lock around config tampering sections. - Check the return value from get_initial_config() properly. ChangeLog v3->v4: - Export the pin_config_get(), pin_config_set() and pin_config_group() functions. - Drop the entire concept of just getting initial config and keeping track of pin states internally, instead ask the pins what state they are in. Previous idea was plain wrong, if the device cannot keep track of its state, the driver should do it. - Drop the generic configuration layout, it seems this impose too much restriction on some pin controllers, so let them do things the way they want and split off support for generic config as an optional add-on. ChangeLog v4->v5: - Introduce two symmetric driver calls for group configuration, .pin_config_group_[get|set] and corresponding external calls. - Remove generic semantic meanings of return values from config calls, these belong in the generic config patch. Just pass the return value through instead. - Add a debugfs entry "pinconf-groups" to read status from group configuration only, also slam in a per-group debug callback in the pinconf_ops so custom drivers can display something meaningful for their pins. - Fix some dangling newline. - Drop dangling #else clause. - Update documentation to match the above. ChangeLog v5->v6: - Change to using a pin name as parameter for the [get|set]_config() functions, as suggested by Stephen Warren. This is more natural as names will be what a developer has access to in written documentation etc. ChangeLog v6->v7: - Refactor out by-pin and by-name get/set functions, only expose the by-name functions externally, expose the by-pin functions internally. - Show supported pin control functionality in the debugfs pinctrl-devices file. Acked-by: Stephen Warren <swarren@nvidia.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2011-10-19 20:14:33 +04:00
ret = pctlops->get_group_pins(pctldev, selector, &pins, &num_pins);
if (ret) {
dev_err(pctldev->dev, "cannot configure pin group, error "
pinctrl: add a pin config interface This add per-pin and per-group pin config interfaces for biasing, driving and other such electronic properties. The details of passed configurations are passed in an opaque unsigned long which may be dereferences to integer types, structs or lists on either side of the configuration interface. ChangeLog v1->v2: - Clear split of terminology: we now have pin controllers, and those may support two interfaces using vtables: pin multiplexing and pin configuration. - Break out pin configuration to its own C file, controllers may implement only config without mux, and vice versa, so keep each sub-functionality of pin controllers separate. Introduce CONFIG_PINCONF in Kconfig. - Implement some core logic around pin configuration in the pinconf.c file. - Remove UNKNOWN config states, these were just surplus baggage. - Remove FLOAT config state - HIGH_IMPEDANCE should be enough for everyone. - PIN_CONFIG_POWER_SOURCE added to handle switching the power supply for the pin logic between different sources - Explicit DISABLE config enums to turn schmitt-trigger, wakeup etc OFF. - Update documentation to reflect all the recent reasoning. ChangeLog v2->v3: - Twist API around to pass around arrays of config tuples instead of (param, value) pairs everywhere. - Explicit drive strength semantics for push/pull and similar drive modes, this shall be the number of drive stages vs nominal load impedance, which should match the actual electronics used in push/pull CMOS or TTY totempoles. - Drop load capacitance configuration - I probably don't know what I'm doing here so leave it out. - Drop PIN_CONFIG_INPUT_SCHMITT_OFF, instead the argument zero to PIN_CONFIG_INPUT_SCHMITT turns schmitt trigger off. - Drop PIN_CONFIG_NORMAL_POWER_MODE and have a well defined argument to PIN_CONFIG_LOW_POWER_MODE to get out of it instead. - Drop PIN_CONFIG_WAKEUP_ENABLE/DISABLE and just use PIN_CONFIG_WAKEUP with defined value zero to turn wakeup off. - Add PIN_CONFIG_INPUT_DEBOUNCE for configuring debounce time on input lines. - Fix a bug when we tried to configure pins for pin controllers without pinconf support. - Initialized debugfs properly so it works. - Initialize the mutex properly and lock around config tampering sections. - Check the return value from get_initial_config() properly. ChangeLog v3->v4: - Export the pin_config_get(), pin_config_set() and pin_config_group() functions. - Drop the entire concept of just getting initial config and keeping track of pin states internally, instead ask the pins what state they are in. Previous idea was plain wrong, if the device cannot keep track of its state, the driver should do it. - Drop the generic configuration layout, it seems this impose too much restriction on some pin controllers, so let them do things the way they want and split off support for generic config as an optional add-on. ChangeLog v4->v5: - Introduce two symmetric driver calls for group configuration, .pin_config_group_[get|set] and corresponding external calls. - Remove generic semantic meanings of return values from config calls, these belong in the generic config patch. Just pass the return value through instead. - Add a debugfs entry "pinconf-groups" to read status from group configuration only, also slam in a per-group debug callback in the pinconf_ops so custom drivers can display something meaningful for their pins. - Fix some dangling newline. - Drop dangling #else clause. - Update documentation to match the above. ChangeLog v5->v6: - Change to using a pin name as parameter for the [get|set]_config() functions, as suggested by Stephen Warren. This is more natural as names will be what a developer has access to in written documentation etc. ChangeLog v6->v7: - Refactor out by-pin and by-name get/set functions, only expose the by-name functions externally, expose the by-pin functions internally. - Show supported pin control functionality in the debugfs pinctrl-devices file. Acked-by: Stephen Warren <swarren@nvidia.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2011-10-19 20:14:33 +04:00
"getting pins\n");
pinctrl: fix and simplify locking There are many problems with the current pinctrl locking: struct pinctrl_dev's gpio_ranges_lock isn't effective; pinctrl_match_gpio_range() only holds this lock while searching for a gpio range, but the found range is return and manipulated after releading the lock. This could allow pinctrl_remove_gpio_range() for that range while it is in use, and the caller may very well delete the range after removing it, causing pinctrl code to touch the now-free range object. Solving this requires the introduction of a higher-level lock, at least a lock per pin controller, which both gpio range registration and pinctrl_get()/put() will acquire. There is missing locking on HW programming; pin controllers may pack the configuration for different pins/groups/config options/... into one register, and hence have to read-modify-write the register. This needs to be protected, but currently isn't. Related, a future change will add a "complete" op to the pin controller drivers, the idea being that each state's programming will be programmed into the pinctrl driver followed by the "complete" call, which may e.g. flush a register cache to HW. For this to work, it must not be possible to interleave the pinctrl driver calls for different devices. As above, solving this requires the introduction of a higher-level lock, at least a lock per pin controller, which will be held for the duration of any pinctrl_enable()/disable() call. However, each pinctrl mapping table entry may affect a different pin controller if necessary. Hence, with a per-pin-controller lock, almost any pinctrl API may need to acquire multiple locks, one per controller. To avoid deadlock, these would need to be acquired in the same order in all cases. This is extremely difficult to implement in the case of pinctrl_get(), which doesn't know which pin controllers to lock until it has parsed the entire mapping table, since it contains somewhat arbitrary data. The simplest solution here is to introduce a single lock that covers all pin controllers at once. This will be acquired by all pinctrl APIs. This then makes struct pinctrl's mutex irrelevant, since that single lock will always be held whenever this mutex is currently held. Signed-off-by: Stephen Warren <swarren@nvidia.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2012-03-03 00:05:44 +04:00
goto unlock;
pinctrl: add a pin config interface This add per-pin and per-group pin config interfaces for biasing, driving and other such electronic properties. The details of passed configurations are passed in an opaque unsigned long which may be dereferences to integer types, structs or lists on either side of the configuration interface. ChangeLog v1->v2: - Clear split of terminology: we now have pin controllers, and those may support two interfaces using vtables: pin multiplexing and pin configuration. - Break out pin configuration to its own C file, controllers may implement only config without mux, and vice versa, so keep each sub-functionality of pin controllers separate. Introduce CONFIG_PINCONF in Kconfig. - Implement some core logic around pin configuration in the pinconf.c file. - Remove UNKNOWN config states, these were just surplus baggage. - Remove FLOAT config state - HIGH_IMPEDANCE should be enough for everyone. - PIN_CONFIG_POWER_SOURCE added to handle switching the power supply for the pin logic between different sources - Explicit DISABLE config enums to turn schmitt-trigger, wakeup etc OFF. - Update documentation to reflect all the recent reasoning. ChangeLog v2->v3: - Twist API around to pass around arrays of config tuples instead of (param, value) pairs everywhere. - Explicit drive strength semantics for push/pull and similar drive modes, this shall be the number of drive stages vs nominal load impedance, which should match the actual electronics used in push/pull CMOS or TTY totempoles. - Drop load capacitance configuration - I probably don't know what I'm doing here so leave it out. - Drop PIN_CONFIG_INPUT_SCHMITT_OFF, instead the argument zero to PIN_CONFIG_INPUT_SCHMITT turns schmitt trigger off. - Drop PIN_CONFIG_NORMAL_POWER_MODE and have a well defined argument to PIN_CONFIG_LOW_POWER_MODE to get out of it instead. - Drop PIN_CONFIG_WAKEUP_ENABLE/DISABLE and just use PIN_CONFIG_WAKEUP with defined value zero to turn wakeup off. - Add PIN_CONFIG_INPUT_DEBOUNCE for configuring debounce time on input lines. - Fix a bug when we tried to configure pins for pin controllers without pinconf support. - Initialized debugfs properly so it works. - Initialize the mutex properly and lock around config tampering sections. - Check the return value from get_initial_config() properly. ChangeLog v3->v4: - Export the pin_config_get(), pin_config_set() and pin_config_group() functions. - Drop the entire concept of just getting initial config and keeping track of pin states internally, instead ask the pins what state they are in. Previous idea was plain wrong, if the device cannot keep track of its state, the driver should do it. - Drop the generic configuration layout, it seems this impose too much restriction on some pin controllers, so let them do things the way they want and split off support for generic config as an optional add-on. ChangeLog v4->v5: - Introduce two symmetric driver calls for group configuration, .pin_config_group_[get|set] and corresponding external calls. - Remove generic semantic meanings of return values from config calls, these belong in the generic config patch. Just pass the return value through instead. - Add a debugfs entry "pinconf-groups" to read status from group configuration only, also slam in a per-group debug callback in the pinconf_ops so custom drivers can display something meaningful for their pins. - Fix some dangling newline. - Drop dangling #else clause. - Update documentation to match the above. ChangeLog v5->v6: - Change to using a pin name as parameter for the [get|set]_config() functions, as suggested by Stephen Warren. This is more natural as names will be what a developer has access to in written documentation etc. ChangeLog v6->v7: - Refactor out by-pin and by-name get/set functions, only expose the by-name functions externally, expose the by-pin functions internally. - Show supported pin control functionality in the debugfs pinctrl-devices file. Acked-by: Stephen Warren <swarren@nvidia.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2011-10-19 20:14:33 +04:00
}
/*
* If the pin controller supports handling entire groups we use that
* capability.
*/
if (ops->pin_config_group_set) {
ret = ops->pin_config_group_set(pctldev, selector, config);
/*
* If the pin controller prefer that a certain group be handled
* pin-by-pin as well, it returns -EAGAIN.
*/
if (ret != -EAGAIN)
pinctrl: fix and simplify locking There are many problems with the current pinctrl locking: struct pinctrl_dev's gpio_ranges_lock isn't effective; pinctrl_match_gpio_range() only holds this lock while searching for a gpio range, but the found range is return and manipulated after releading the lock. This could allow pinctrl_remove_gpio_range() for that range while it is in use, and the caller may very well delete the range after removing it, causing pinctrl code to touch the now-free range object. Solving this requires the introduction of a higher-level lock, at least a lock per pin controller, which both gpio range registration and pinctrl_get()/put() will acquire. There is missing locking on HW programming; pin controllers may pack the configuration for different pins/groups/config options/... into one register, and hence have to read-modify-write the register. This needs to be protected, but currently isn't. Related, a future change will add a "complete" op to the pin controller drivers, the idea being that each state's programming will be programmed into the pinctrl driver followed by the "complete" call, which may e.g. flush a register cache to HW. For this to work, it must not be possible to interleave the pinctrl driver calls for different devices. As above, solving this requires the introduction of a higher-level lock, at least a lock per pin controller, which will be held for the duration of any pinctrl_enable()/disable() call. However, each pinctrl mapping table entry may affect a different pin controller if necessary. Hence, with a per-pin-controller lock, almost any pinctrl API may need to acquire multiple locks, one per controller. To avoid deadlock, these would need to be acquired in the same order in all cases. This is extremely difficult to implement in the case of pinctrl_get(), which doesn't know which pin controllers to lock until it has parsed the entire mapping table, since it contains somewhat arbitrary data. The simplest solution here is to introduce a single lock that covers all pin controllers at once. This will be acquired by all pinctrl APIs. This then makes struct pinctrl's mutex irrelevant, since that single lock will always be held whenever this mutex is currently held. Signed-off-by: Stephen Warren <swarren@nvidia.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2012-03-03 00:05:44 +04:00
goto unlock;
pinctrl: add a pin config interface This add per-pin and per-group pin config interfaces for biasing, driving and other such electronic properties. The details of passed configurations are passed in an opaque unsigned long which may be dereferences to integer types, structs or lists on either side of the configuration interface. ChangeLog v1->v2: - Clear split of terminology: we now have pin controllers, and those may support two interfaces using vtables: pin multiplexing and pin configuration. - Break out pin configuration to its own C file, controllers may implement only config without mux, and vice versa, so keep each sub-functionality of pin controllers separate. Introduce CONFIG_PINCONF in Kconfig. - Implement some core logic around pin configuration in the pinconf.c file. - Remove UNKNOWN config states, these were just surplus baggage. - Remove FLOAT config state - HIGH_IMPEDANCE should be enough for everyone. - PIN_CONFIG_POWER_SOURCE added to handle switching the power supply for the pin logic between different sources - Explicit DISABLE config enums to turn schmitt-trigger, wakeup etc OFF. - Update documentation to reflect all the recent reasoning. ChangeLog v2->v3: - Twist API around to pass around arrays of config tuples instead of (param, value) pairs everywhere. - Explicit drive strength semantics for push/pull and similar drive modes, this shall be the number of drive stages vs nominal load impedance, which should match the actual electronics used in push/pull CMOS or TTY totempoles. - Drop load capacitance configuration - I probably don't know what I'm doing here so leave it out. - Drop PIN_CONFIG_INPUT_SCHMITT_OFF, instead the argument zero to PIN_CONFIG_INPUT_SCHMITT turns schmitt trigger off. - Drop PIN_CONFIG_NORMAL_POWER_MODE and have a well defined argument to PIN_CONFIG_LOW_POWER_MODE to get out of it instead. - Drop PIN_CONFIG_WAKEUP_ENABLE/DISABLE and just use PIN_CONFIG_WAKEUP with defined value zero to turn wakeup off. - Add PIN_CONFIG_INPUT_DEBOUNCE for configuring debounce time on input lines. - Fix a bug when we tried to configure pins for pin controllers without pinconf support. - Initialized debugfs properly so it works. - Initialize the mutex properly and lock around config tampering sections. - Check the return value from get_initial_config() properly. ChangeLog v3->v4: - Export the pin_config_get(), pin_config_set() and pin_config_group() functions. - Drop the entire concept of just getting initial config and keeping track of pin states internally, instead ask the pins what state they are in. Previous idea was plain wrong, if the device cannot keep track of its state, the driver should do it. - Drop the generic configuration layout, it seems this impose too much restriction on some pin controllers, so let them do things the way they want and split off support for generic config as an optional add-on. ChangeLog v4->v5: - Introduce two symmetric driver calls for group configuration, .pin_config_group_[get|set] and corresponding external calls. - Remove generic semantic meanings of return values from config calls, these belong in the generic config patch. Just pass the return value through instead. - Add a debugfs entry "pinconf-groups" to read status from group configuration only, also slam in a per-group debug callback in the pinconf_ops so custom drivers can display something meaningful for their pins. - Fix some dangling newline. - Drop dangling #else clause. - Update documentation to match the above. ChangeLog v5->v6: - Change to using a pin name as parameter for the [get|set]_config() functions, as suggested by Stephen Warren. This is more natural as names will be what a developer has access to in written documentation etc. ChangeLog v6->v7: - Refactor out by-pin and by-name get/set functions, only expose the by-name functions externally, expose the by-pin functions internally. - Show supported pin control functionality in the debugfs pinctrl-devices file. Acked-by: Stephen Warren <swarren@nvidia.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2011-10-19 20:14:33 +04:00
}
/*
* If the controller cannot handle entire groups, we configure each pin
* individually.
*/
pinctrl: fix and simplify locking There are many problems with the current pinctrl locking: struct pinctrl_dev's gpio_ranges_lock isn't effective; pinctrl_match_gpio_range() only holds this lock while searching for a gpio range, but the found range is return and manipulated after releading the lock. This could allow pinctrl_remove_gpio_range() for that range while it is in use, and the caller may very well delete the range after removing it, causing pinctrl code to touch the now-free range object. Solving this requires the introduction of a higher-level lock, at least a lock per pin controller, which both gpio range registration and pinctrl_get()/put() will acquire. There is missing locking on HW programming; pin controllers may pack the configuration for different pins/groups/config options/... into one register, and hence have to read-modify-write the register. This needs to be protected, but currently isn't. Related, a future change will add a "complete" op to the pin controller drivers, the idea being that each state's programming will be programmed into the pinctrl driver followed by the "complete" call, which may e.g. flush a register cache to HW. For this to work, it must not be possible to interleave the pinctrl driver calls for different devices. As above, solving this requires the introduction of a higher-level lock, at least a lock per pin controller, which will be held for the duration of any pinctrl_enable()/disable() call. However, each pinctrl mapping table entry may affect a different pin controller if necessary. Hence, with a per-pin-controller lock, almost any pinctrl API may need to acquire multiple locks, one per controller. To avoid deadlock, these would need to be acquired in the same order in all cases. This is extremely difficult to implement in the case of pinctrl_get(), which doesn't know which pin controllers to lock until it has parsed the entire mapping table, since it contains somewhat arbitrary data. The simplest solution here is to introduce a single lock that covers all pin controllers at once. This will be acquired by all pinctrl APIs. This then makes struct pinctrl's mutex irrelevant, since that single lock will always be held whenever this mutex is currently held. Signed-off-by: Stephen Warren <swarren@nvidia.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2012-03-03 00:05:44 +04:00
if (!ops->pin_config_set) {
ret = 0;
goto unlock;
}
pinctrl: add a pin config interface This add per-pin and per-group pin config interfaces for biasing, driving and other such electronic properties. The details of passed configurations are passed in an opaque unsigned long which may be dereferences to integer types, structs or lists on either side of the configuration interface. ChangeLog v1->v2: - Clear split of terminology: we now have pin controllers, and those may support two interfaces using vtables: pin multiplexing and pin configuration. - Break out pin configuration to its own C file, controllers may implement only config without mux, and vice versa, so keep each sub-functionality of pin controllers separate. Introduce CONFIG_PINCONF in Kconfig. - Implement some core logic around pin configuration in the pinconf.c file. - Remove UNKNOWN config states, these were just surplus baggage. - Remove FLOAT config state - HIGH_IMPEDANCE should be enough for everyone. - PIN_CONFIG_POWER_SOURCE added to handle switching the power supply for the pin logic between different sources - Explicit DISABLE config enums to turn schmitt-trigger, wakeup etc OFF. - Update documentation to reflect all the recent reasoning. ChangeLog v2->v3: - Twist API around to pass around arrays of config tuples instead of (param, value) pairs everywhere. - Explicit drive strength semantics for push/pull and similar drive modes, this shall be the number of drive stages vs nominal load impedance, which should match the actual electronics used in push/pull CMOS or TTY totempoles. - Drop load capacitance configuration - I probably don't know what I'm doing here so leave it out. - Drop PIN_CONFIG_INPUT_SCHMITT_OFF, instead the argument zero to PIN_CONFIG_INPUT_SCHMITT turns schmitt trigger off. - Drop PIN_CONFIG_NORMAL_POWER_MODE and have a well defined argument to PIN_CONFIG_LOW_POWER_MODE to get out of it instead. - Drop PIN_CONFIG_WAKEUP_ENABLE/DISABLE and just use PIN_CONFIG_WAKEUP with defined value zero to turn wakeup off. - Add PIN_CONFIG_INPUT_DEBOUNCE for configuring debounce time on input lines. - Fix a bug when we tried to configure pins for pin controllers without pinconf support. - Initialized debugfs properly so it works. - Initialize the mutex properly and lock around config tampering sections. - Check the return value from get_initial_config() properly. ChangeLog v3->v4: - Export the pin_config_get(), pin_config_set() and pin_config_group() functions. - Drop the entire concept of just getting initial config and keeping track of pin states internally, instead ask the pins what state they are in. Previous idea was plain wrong, if the device cannot keep track of its state, the driver should do it. - Drop the generic configuration layout, it seems this impose too much restriction on some pin controllers, so let them do things the way they want and split off support for generic config as an optional add-on. ChangeLog v4->v5: - Introduce two symmetric driver calls for group configuration, .pin_config_group_[get|set] and corresponding external calls. - Remove generic semantic meanings of return values from config calls, these belong in the generic config patch. Just pass the return value through instead. - Add a debugfs entry "pinconf-groups" to read status from group configuration only, also slam in a per-group debug callback in the pinconf_ops so custom drivers can display something meaningful for their pins. - Fix some dangling newline. - Drop dangling #else clause. - Update documentation to match the above. ChangeLog v5->v6: - Change to using a pin name as parameter for the [get|set]_config() functions, as suggested by Stephen Warren. This is more natural as names will be what a developer has access to in written documentation etc. ChangeLog v6->v7: - Refactor out by-pin and by-name get/set functions, only expose the by-name functions externally, expose the by-pin functions internally. - Show supported pin control functionality in the debugfs pinctrl-devices file. Acked-by: Stephen Warren <swarren@nvidia.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2011-10-19 20:14:33 +04:00
for (i = 0; i < num_pins; i++) {
ret = ops->pin_config_set(pctldev, pins[i], config);
if (ret < 0)
pinctrl: fix and simplify locking There are many problems with the current pinctrl locking: struct pinctrl_dev's gpio_ranges_lock isn't effective; pinctrl_match_gpio_range() only holds this lock while searching for a gpio range, but the found range is return and manipulated after releading the lock. This could allow pinctrl_remove_gpio_range() for that range while it is in use, and the caller may very well delete the range after removing it, causing pinctrl code to touch the now-free range object. Solving this requires the introduction of a higher-level lock, at least a lock per pin controller, which both gpio range registration and pinctrl_get()/put() will acquire. There is missing locking on HW programming; pin controllers may pack the configuration for different pins/groups/config options/... into one register, and hence have to read-modify-write the register. This needs to be protected, but currently isn't. Related, a future change will add a "complete" op to the pin controller drivers, the idea being that each state's programming will be programmed into the pinctrl driver followed by the "complete" call, which may e.g. flush a register cache to HW. For this to work, it must not be possible to interleave the pinctrl driver calls for different devices. As above, solving this requires the introduction of a higher-level lock, at least a lock per pin controller, which will be held for the duration of any pinctrl_enable()/disable() call. However, each pinctrl mapping table entry may affect a different pin controller if necessary. Hence, with a per-pin-controller lock, almost any pinctrl API may need to acquire multiple locks, one per controller. To avoid deadlock, these would need to be acquired in the same order in all cases. This is extremely difficult to implement in the case of pinctrl_get(), which doesn't know which pin controllers to lock until it has parsed the entire mapping table, since it contains somewhat arbitrary data. The simplest solution here is to introduce a single lock that covers all pin controllers at once. This will be acquired by all pinctrl APIs. This then makes struct pinctrl's mutex irrelevant, since that single lock will always be held whenever this mutex is currently held. Signed-off-by: Stephen Warren <swarren@nvidia.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2012-03-03 00:05:44 +04:00
goto unlock;
pinctrl: add a pin config interface This add per-pin and per-group pin config interfaces for biasing, driving and other such electronic properties. The details of passed configurations are passed in an opaque unsigned long which may be dereferences to integer types, structs or lists on either side of the configuration interface. ChangeLog v1->v2: - Clear split of terminology: we now have pin controllers, and those may support two interfaces using vtables: pin multiplexing and pin configuration. - Break out pin configuration to its own C file, controllers may implement only config without mux, and vice versa, so keep each sub-functionality of pin controllers separate. Introduce CONFIG_PINCONF in Kconfig. - Implement some core logic around pin configuration in the pinconf.c file. - Remove UNKNOWN config states, these were just surplus baggage. - Remove FLOAT config state - HIGH_IMPEDANCE should be enough for everyone. - PIN_CONFIG_POWER_SOURCE added to handle switching the power supply for the pin logic between different sources - Explicit DISABLE config enums to turn schmitt-trigger, wakeup etc OFF. - Update documentation to reflect all the recent reasoning. ChangeLog v2->v3: - Twist API around to pass around arrays of config tuples instead of (param, value) pairs everywhere. - Explicit drive strength semantics for push/pull and similar drive modes, this shall be the number of drive stages vs nominal load impedance, which should match the actual electronics used in push/pull CMOS or TTY totempoles. - Drop load capacitance configuration - I probably don't know what I'm doing here so leave it out. - Drop PIN_CONFIG_INPUT_SCHMITT_OFF, instead the argument zero to PIN_CONFIG_INPUT_SCHMITT turns schmitt trigger off. - Drop PIN_CONFIG_NORMAL_POWER_MODE and have a well defined argument to PIN_CONFIG_LOW_POWER_MODE to get out of it instead. - Drop PIN_CONFIG_WAKEUP_ENABLE/DISABLE and just use PIN_CONFIG_WAKEUP with defined value zero to turn wakeup off. - Add PIN_CONFIG_INPUT_DEBOUNCE for configuring debounce time on input lines. - Fix a bug when we tried to configure pins for pin controllers without pinconf support. - Initialized debugfs properly so it works. - Initialize the mutex properly and lock around config tampering sections. - Check the return value from get_initial_config() properly. ChangeLog v3->v4: - Export the pin_config_get(), pin_config_set() and pin_config_group() functions. - Drop the entire concept of just getting initial config and keeping track of pin states internally, instead ask the pins what state they are in. Previous idea was plain wrong, if the device cannot keep track of its state, the driver should do it. - Drop the generic configuration layout, it seems this impose too much restriction on some pin controllers, so let them do things the way they want and split off support for generic config as an optional add-on. ChangeLog v4->v5: - Introduce two symmetric driver calls for group configuration, .pin_config_group_[get|set] and corresponding external calls. - Remove generic semantic meanings of return values from config calls, these belong in the generic config patch. Just pass the return value through instead. - Add a debugfs entry "pinconf-groups" to read status from group configuration only, also slam in a per-group debug callback in the pinconf_ops so custom drivers can display something meaningful for their pins. - Fix some dangling newline. - Drop dangling #else clause. - Update documentation to match the above. ChangeLog v5->v6: - Change to using a pin name as parameter for the [get|set]_config() functions, as suggested by Stephen Warren. This is more natural as names will be what a developer has access to in written documentation etc. ChangeLog v6->v7: - Refactor out by-pin and by-name get/set functions, only expose the by-name functions externally, expose the by-pin functions internally. - Show supported pin control functionality in the debugfs pinctrl-devices file. Acked-by: Stephen Warren <swarren@nvidia.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2011-10-19 20:14:33 +04:00
}
pinctrl: fix and simplify locking There are many problems with the current pinctrl locking: struct pinctrl_dev's gpio_ranges_lock isn't effective; pinctrl_match_gpio_range() only holds this lock while searching for a gpio range, but the found range is return and manipulated after releading the lock. This could allow pinctrl_remove_gpio_range() for that range while it is in use, and the caller may very well delete the range after removing it, causing pinctrl code to touch the now-free range object. Solving this requires the introduction of a higher-level lock, at least a lock per pin controller, which both gpio range registration and pinctrl_get()/put() will acquire. There is missing locking on HW programming; pin controllers may pack the configuration for different pins/groups/config options/... into one register, and hence have to read-modify-write the register. This needs to be protected, but currently isn't. Related, a future change will add a "complete" op to the pin controller drivers, the idea being that each state's programming will be programmed into the pinctrl driver followed by the "complete" call, which may e.g. flush a register cache to HW. For this to work, it must not be possible to interleave the pinctrl driver calls for different devices. As above, solving this requires the introduction of a higher-level lock, at least a lock per pin controller, which will be held for the duration of any pinctrl_enable()/disable() call. However, each pinctrl mapping table entry may affect a different pin controller if necessary. Hence, with a per-pin-controller lock, almost any pinctrl API may need to acquire multiple locks, one per controller. To avoid deadlock, these would need to be acquired in the same order in all cases. This is extremely difficult to implement in the case of pinctrl_get(), which doesn't know which pin controllers to lock until it has parsed the entire mapping table, since it contains somewhat arbitrary data. The simplest solution here is to introduce a single lock that covers all pin controllers at once. This will be acquired by all pinctrl APIs. This then makes struct pinctrl's mutex irrelevant, since that single lock will always be held whenever this mutex is currently held. Signed-off-by: Stephen Warren <swarren@nvidia.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2012-03-03 00:05:44 +04:00
ret = 0;
unlock:
pinctrl: move subsystem mutex to pinctrl_dev struct This mutex avoids deadlock in case of use of multiple pin controllers. Before this modification, by using a global mutex, deadlock appeared when, for example, a call to pinctrl_pins_show() locked the pinctrl_mutex, called the ops->pin_dbg_show of a particular pin controller. If this pin controller needs I2C access to retrieve configuration information and I2C driver is using pinctrl to drive its pins, a call to pinctrl_select_state() try to lock again pinctrl_mutex which leads to a deadlock. Notice that the mutex grab from the two direction functions was moved into pinctrl_gpio_direction(). For several cases, we can't replace pinctrl_mutex by pctldev->mutex, because at this stage, pctldev is not accessible : - pinctrl_get()/pinctrl_put() - pinctrl_register_maps() So add respectively pinctrl_list_mutex and pinctrl_maps_mutex in order to protect pinctrl_list and pinctrl_maps list instead. Reintroduce pinctrldev_list_mutex in find_pinctrl_by_of_node(), pinctrl_find_and_add_gpio_range() pinctrl_request_gpio(), pinctrl_free_gpio(), pinctrl_gpio_direction(), pinctrl_devices_show(), pinctrl_register() and pinctrl_unregister() to protect pinctrldev_list. Changes v2->v3: - Fix a missing EXPORT_SYMBOL_GPL() for pinctrl_select_state(). Changes v1->v2: - pinctrl_select_state_locked() is removed, all lock mechanism is located inside pinctrl_select_state(). When parsing the state->setting list, take the per-pin-controller driver lock. (Patrice). - Introduce pinctrldev_list_mutex to protect pinctrldev_list in all functions which parse or modify pictrldev_list. (Patrice). - move find_pinctrl_by_of_node() from pinctrl/devicetree.c to pinctrl/core.c in order to protect pinctrldev_list. (Patrice). - Sink mutex:es into some functions and remove some _locked variants down to where the lists are actually accessed to make things simpler. (Linus) - Drop *all* mutexes completely from pinctrl_lookup_state() and pinctrl_select_state() - no relevant mutex was taken and it was unclear what this was protecting against. (Linus) Reported by : Seraphin Bonnaffe <seraphin.bonnaffe@stericsson.com> Signed-off-by: Patrice Chotard <patrice.chotard@st.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2013-04-11 13:01:27 +04:00
mutex_unlock(&pctldev->mutex);
pinctrl: fix and simplify locking There are many problems with the current pinctrl locking: struct pinctrl_dev's gpio_ranges_lock isn't effective; pinctrl_match_gpio_range() only holds this lock while searching for a gpio range, but the found range is return and manipulated after releading the lock. This could allow pinctrl_remove_gpio_range() for that range while it is in use, and the caller may very well delete the range after removing it, causing pinctrl code to touch the now-free range object. Solving this requires the introduction of a higher-level lock, at least a lock per pin controller, which both gpio range registration and pinctrl_get()/put() will acquire. There is missing locking on HW programming; pin controllers may pack the configuration for different pins/groups/config options/... into one register, and hence have to read-modify-write the register. This needs to be protected, but currently isn't. Related, a future change will add a "complete" op to the pin controller drivers, the idea being that each state's programming will be programmed into the pinctrl driver followed by the "complete" call, which may e.g. flush a register cache to HW. For this to work, it must not be possible to interleave the pinctrl driver calls for different devices. As above, solving this requires the introduction of a higher-level lock, at least a lock per pin controller, which will be held for the duration of any pinctrl_enable()/disable() call. However, each pinctrl mapping table entry may affect a different pin controller if necessary. Hence, with a per-pin-controller lock, almost any pinctrl API may need to acquire multiple locks, one per controller. To avoid deadlock, these would need to be acquired in the same order in all cases. This is extremely difficult to implement in the case of pinctrl_get(), which doesn't know which pin controllers to lock until it has parsed the entire mapping table, since it contains somewhat arbitrary data. The simplest solution here is to introduce a single lock that covers all pin controllers at once. This will be acquired by all pinctrl APIs. This then makes struct pinctrl's mutex irrelevant, since that single lock will always be held whenever this mutex is currently held. Signed-off-by: Stephen Warren <swarren@nvidia.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2012-03-03 00:05:44 +04:00
return ret;
pinctrl: add a pin config interface This add per-pin and per-group pin config interfaces for biasing, driving and other such electronic properties. The details of passed configurations are passed in an opaque unsigned long which may be dereferences to integer types, structs or lists on either side of the configuration interface. ChangeLog v1->v2: - Clear split of terminology: we now have pin controllers, and those may support two interfaces using vtables: pin multiplexing and pin configuration. - Break out pin configuration to its own C file, controllers may implement only config without mux, and vice versa, so keep each sub-functionality of pin controllers separate. Introduce CONFIG_PINCONF in Kconfig. - Implement some core logic around pin configuration in the pinconf.c file. - Remove UNKNOWN config states, these were just surplus baggage. - Remove FLOAT config state - HIGH_IMPEDANCE should be enough for everyone. - PIN_CONFIG_POWER_SOURCE added to handle switching the power supply for the pin logic between different sources - Explicit DISABLE config enums to turn schmitt-trigger, wakeup etc OFF. - Update documentation to reflect all the recent reasoning. ChangeLog v2->v3: - Twist API around to pass around arrays of config tuples instead of (param, value) pairs everywhere. - Explicit drive strength semantics for push/pull and similar drive modes, this shall be the number of drive stages vs nominal load impedance, which should match the actual electronics used in push/pull CMOS or TTY totempoles. - Drop load capacitance configuration - I probably don't know what I'm doing here so leave it out. - Drop PIN_CONFIG_INPUT_SCHMITT_OFF, instead the argument zero to PIN_CONFIG_INPUT_SCHMITT turns schmitt trigger off. - Drop PIN_CONFIG_NORMAL_POWER_MODE and have a well defined argument to PIN_CONFIG_LOW_POWER_MODE to get out of it instead. - Drop PIN_CONFIG_WAKEUP_ENABLE/DISABLE and just use PIN_CONFIG_WAKEUP with defined value zero to turn wakeup off. - Add PIN_CONFIG_INPUT_DEBOUNCE for configuring debounce time on input lines. - Fix a bug when we tried to configure pins for pin controllers without pinconf support. - Initialized debugfs properly so it works. - Initialize the mutex properly and lock around config tampering sections. - Check the return value from get_initial_config() properly. ChangeLog v3->v4: - Export the pin_config_get(), pin_config_set() and pin_config_group() functions. - Drop the entire concept of just getting initial config and keeping track of pin states internally, instead ask the pins what state they are in. Previous idea was plain wrong, if the device cannot keep track of its state, the driver should do it. - Drop the generic configuration layout, it seems this impose too much restriction on some pin controllers, so let them do things the way they want and split off support for generic config as an optional add-on. ChangeLog v4->v5: - Introduce two symmetric driver calls for group configuration, .pin_config_group_[get|set] and corresponding external calls. - Remove generic semantic meanings of return values from config calls, these belong in the generic config patch. Just pass the return value through instead. - Add a debugfs entry "pinconf-groups" to read status from group configuration only, also slam in a per-group debug callback in the pinconf_ops so custom drivers can display something meaningful for their pins. - Fix some dangling newline. - Drop dangling #else clause. - Update documentation to match the above. ChangeLog v5->v6: - Change to using a pin name as parameter for the [get|set]_config() functions, as suggested by Stephen Warren. This is more natural as names will be what a developer has access to in written documentation etc. ChangeLog v6->v7: - Refactor out by-pin and by-name get/set functions, only expose the by-name functions externally, expose the by-pin functions internally. - Show supported pin control functionality in the debugfs pinctrl-devices file. Acked-by: Stephen Warren <swarren@nvidia.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2011-10-19 20:14:33 +04:00
}
EXPORT_SYMBOL(pin_config_group_set);
int pinconf_map_to_setting(struct pinctrl_map const *map,
struct pinctrl_setting *setting)
{
struct pinctrl_dev *pctldev = setting->pctldev;
int pin;
switch (setting->type) {
case PIN_MAP_TYPE_CONFIGS_PIN:
pin = pin_get_from_name(pctldev,
map->data.configs.group_or_pin);
if (pin < 0) {
dev_err(pctldev->dev, "could not map pin config for \"%s\"",
map->data.configs.group_or_pin);
return pin;
}
setting->data.configs.group_or_pin = pin;
break;
case PIN_MAP_TYPE_CONFIGS_GROUP:
pin = pinctrl_get_group_selector(pctldev,
map->data.configs.group_or_pin);
if (pin < 0) {
dev_err(pctldev->dev, "could not map group config for \"%s\"",
map->data.configs.group_or_pin);
return pin;
}
setting->data.configs.group_or_pin = pin;
break;
default:
return -EINVAL;
}
setting->data.configs.num_configs = map->data.configs.num_configs;
setting->data.configs.configs = map->data.configs.configs;
return 0;
}
void pinconf_free_setting(struct pinctrl_setting const *setting)
{
}
int pinconf_apply_setting(struct pinctrl_setting const *setting)
{
struct pinctrl_dev *pctldev = setting->pctldev;
const struct pinconf_ops *ops = pctldev->desc->confops;
int i, ret;
if (!ops) {
dev_err(pctldev->dev, "missing confops\n");
return -EINVAL;
}
switch (setting->type) {
case PIN_MAP_TYPE_CONFIGS_PIN:
if (!ops->pin_config_set) {
dev_err(pctldev->dev, "missing pin_config_set op\n");
return -EINVAL;
}
for (i = 0; i < setting->data.configs.num_configs; i++) {
ret = ops->pin_config_set(pctldev,
setting->data.configs.group_or_pin,
setting->data.configs.configs[i]);
if (ret < 0) {
dev_err(pctldev->dev,
"pin_config_set op failed for pin %d config %08lx\n",
setting->data.configs.group_or_pin,
setting->data.configs.configs[i]);
return ret;
}
}
break;
case PIN_MAP_TYPE_CONFIGS_GROUP:
if (!ops->pin_config_group_set) {
dev_err(pctldev->dev,
"missing pin_config_group_set op\n");
return -EINVAL;
}
for (i = 0; i < setting->data.configs.num_configs; i++) {
ret = ops->pin_config_group_set(pctldev,
setting->data.configs.group_or_pin,
setting->data.configs.configs[i]);
if (ret < 0) {
dev_err(pctldev->dev,
"pin_config_group_set op failed for group %d config %08lx\n",
setting->data.configs.group_or_pin,
setting->data.configs.configs[i]);
return ret;
}
}
break;
default:
return -EINVAL;
}
return 0;
}
pinctrl: add a pin config interface This add per-pin and per-group pin config interfaces for biasing, driving and other such electronic properties. The details of passed configurations are passed in an opaque unsigned long which may be dereferences to integer types, structs or lists on either side of the configuration interface. ChangeLog v1->v2: - Clear split of terminology: we now have pin controllers, and those may support two interfaces using vtables: pin multiplexing and pin configuration. - Break out pin configuration to its own C file, controllers may implement only config without mux, and vice versa, so keep each sub-functionality of pin controllers separate. Introduce CONFIG_PINCONF in Kconfig. - Implement some core logic around pin configuration in the pinconf.c file. - Remove UNKNOWN config states, these were just surplus baggage. - Remove FLOAT config state - HIGH_IMPEDANCE should be enough for everyone. - PIN_CONFIG_POWER_SOURCE added to handle switching the power supply for the pin logic between different sources - Explicit DISABLE config enums to turn schmitt-trigger, wakeup etc OFF. - Update documentation to reflect all the recent reasoning. ChangeLog v2->v3: - Twist API around to pass around arrays of config tuples instead of (param, value) pairs everywhere. - Explicit drive strength semantics for push/pull and similar drive modes, this shall be the number of drive stages vs nominal load impedance, which should match the actual electronics used in push/pull CMOS or TTY totempoles. - Drop load capacitance configuration - I probably don't know what I'm doing here so leave it out. - Drop PIN_CONFIG_INPUT_SCHMITT_OFF, instead the argument zero to PIN_CONFIG_INPUT_SCHMITT turns schmitt trigger off. - Drop PIN_CONFIG_NORMAL_POWER_MODE and have a well defined argument to PIN_CONFIG_LOW_POWER_MODE to get out of it instead. - Drop PIN_CONFIG_WAKEUP_ENABLE/DISABLE and just use PIN_CONFIG_WAKEUP with defined value zero to turn wakeup off. - Add PIN_CONFIG_INPUT_DEBOUNCE for configuring debounce time on input lines. - Fix a bug when we tried to configure pins for pin controllers without pinconf support. - Initialized debugfs properly so it works. - Initialize the mutex properly and lock around config tampering sections. - Check the return value from get_initial_config() properly. ChangeLog v3->v4: - Export the pin_config_get(), pin_config_set() and pin_config_group() functions. - Drop the entire concept of just getting initial config and keeping track of pin states internally, instead ask the pins what state they are in. Previous idea was plain wrong, if the device cannot keep track of its state, the driver should do it. - Drop the generic configuration layout, it seems this impose too much restriction on some pin controllers, so let them do things the way they want and split off support for generic config as an optional add-on. ChangeLog v4->v5: - Introduce two symmetric driver calls for group configuration, .pin_config_group_[get|set] and corresponding external calls. - Remove generic semantic meanings of return values from config calls, these belong in the generic config patch. Just pass the return value through instead. - Add a debugfs entry "pinconf-groups" to read status from group configuration only, also slam in a per-group debug callback in the pinconf_ops so custom drivers can display something meaningful for their pins. - Fix some dangling newline. - Drop dangling #else clause. - Update documentation to match the above. ChangeLog v5->v6: - Change to using a pin name as parameter for the [get|set]_config() functions, as suggested by Stephen Warren. This is more natural as names will be what a developer has access to in written documentation etc. ChangeLog v6->v7: - Refactor out by-pin and by-name get/set functions, only expose the by-name functions externally, expose the by-pin functions internally. - Show supported pin control functionality in the debugfs pinctrl-devices file. Acked-by: Stephen Warren <swarren@nvidia.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2011-10-19 20:14:33 +04:00
#ifdef CONFIG_DEBUG_FS
void pinconf_show_map(struct seq_file *s, struct pinctrl_map const *map)
{
struct pinctrl_dev *pctldev;
const struct pinconf_ops *confops;
int i;
pctldev = get_pinctrl_dev_from_devname(map->ctrl_dev_name);
if (pctldev)
confops = pctldev->desc->confops;
else
confops = NULL;
switch (map->type) {
case PIN_MAP_TYPE_CONFIGS_PIN:
seq_printf(s, "pin ");
break;
case PIN_MAP_TYPE_CONFIGS_GROUP:
seq_printf(s, "group ");
break;
default:
break;
}
seq_printf(s, "%s\n", map->data.configs.group_or_pin);
for (i = 0; i < map->data.configs.num_configs; i++) {
seq_printf(s, "config ");
if (confops && confops->pin_config_config_dbg_show)
confops->pin_config_config_dbg_show(pctldev, s,
map->data.configs.configs[i]);
else
seq_printf(s, "%08lx", map->data.configs.configs[i]);
seq_printf(s, "\n");
}
}
void pinconf_show_setting(struct seq_file *s,
struct pinctrl_setting const *setting)
{
struct pinctrl_dev *pctldev = setting->pctldev;
const struct pinctrl_ops *pctlops = pctldev->desc->pctlops;
const struct pinconf_ops *confops = pctldev->desc->confops;
struct pin_desc *desc;
int i;
switch (setting->type) {
case PIN_MAP_TYPE_CONFIGS_PIN:
desc = pin_desc_get(setting->pctldev,
setting->data.configs.group_or_pin);
seq_printf(s, "pin %s (%d)",
desc->name ? desc->name : "unnamed",
setting->data.configs.group_or_pin);
break;
case PIN_MAP_TYPE_CONFIGS_GROUP:
seq_printf(s, "group %s (%d)",
pctlops->get_group_name(pctldev,
setting->data.configs.group_or_pin),
setting->data.configs.group_or_pin);
break;
default:
break;
}
/*
* FIXME: We should really get the pin controler to dump the config
* values, so they can be decoded to something meaningful.
*/
for (i = 0; i < setting->data.configs.num_configs; i++) {
seq_printf(s, " ");
if (confops && confops->pin_config_config_dbg_show)
confops->pin_config_config_dbg_show(pctldev, s,
setting->data.configs.configs[i]);
else
seq_printf(s, "%08lx",
setting->data.configs.configs[i]);
}
seq_printf(s, "\n");
}
pinctrl: add a pin config interface This add per-pin and per-group pin config interfaces for biasing, driving and other such electronic properties. The details of passed configurations are passed in an opaque unsigned long which may be dereferences to integer types, structs or lists on either side of the configuration interface. ChangeLog v1->v2: - Clear split of terminology: we now have pin controllers, and those may support two interfaces using vtables: pin multiplexing and pin configuration. - Break out pin configuration to its own C file, controllers may implement only config without mux, and vice versa, so keep each sub-functionality of pin controllers separate. Introduce CONFIG_PINCONF in Kconfig. - Implement some core logic around pin configuration in the pinconf.c file. - Remove UNKNOWN config states, these were just surplus baggage. - Remove FLOAT config state - HIGH_IMPEDANCE should be enough for everyone. - PIN_CONFIG_POWER_SOURCE added to handle switching the power supply for the pin logic between different sources - Explicit DISABLE config enums to turn schmitt-trigger, wakeup etc OFF. - Update documentation to reflect all the recent reasoning. ChangeLog v2->v3: - Twist API around to pass around arrays of config tuples instead of (param, value) pairs everywhere. - Explicit drive strength semantics for push/pull and similar drive modes, this shall be the number of drive stages vs nominal load impedance, which should match the actual electronics used in push/pull CMOS or TTY totempoles. - Drop load capacitance configuration - I probably don't know what I'm doing here so leave it out. - Drop PIN_CONFIG_INPUT_SCHMITT_OFF, instead the argument zero to PIN_CONFIG_INPUT_SCHMITT turns schmitt trigger off. - Drop PIN_CONFIG_NORMAL_POWER_MODE and have a well defined argument to PIN_CONFIG_LOW_POWER_MODE to get out of it instead. - Drop PIN_CONFIG_WAKEUP_ENABLE/DISABLE and just use PIN_CONFIG_WAKEUP with defined value zero to turn wakeup off. - Add PIN_CONFIG_INPUT_DEBOUNCE for configuring debounce time on input lines. - Fix a bug when we tried to configure pins for pin controllers without pinconf support. - Initialized debugfs properly so it works. - Initialize the mutex properly and lock around config tampering sections. - Check the return value from get_initial_config() properly. ChangeLog v3->v4: - Export the pin_config_get(), pin_config_set() and pin_config_group() functions. - Drop the entire concept of just getting initial config and keeping track of pin states internally, instead ask the pins what state they are in. Previous idea was plain wrong, if the device cannot keep track of its state, the driver should do it. - Drop the generic configuration layout, it seems this impose too much restriction on some pin controllers, so let them do things the way they want and split off support for generic config as an optional add-on. ChangeLog v4->v5: - Introduce two symmetric driver calls for group configuration, .pin_config_group_[get|set] and corresponding external calls. - Remove generic semantic meanings of return values from config calls, these belong in the generic config patch. Just pass the return value through instead. - Add a debugfs entry "pinconf-groups" to read status from group configuration only, also slam in a per-group debug callback in the pinconf_ops so custom drivers can display something meaningful for their pins. - Fix some dangling newline. - Drop dangling #else clause. - Update documentation to match the above. ChangeLog v5->v6: - Change to using a pin name as parameter for the [get|set]_config() functions, as suggested by Stephen Warren. This is more natural as names will be what a developer has access to in written documentation etc. ChangeLog v6->v7: - Refactor out by-pin and by-name get/set functions, only expose the by-name functions externally, expose the by-pin functions internally. - Show supported pin control functionality in the debugfs pinctrl-devices file. Acked-by: Stephen Warren <swarren@nvidia.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2011-10-19 20:14:33 +04:00
static void pinconf_dump_pin(struct pinctrl_dev *pctldev,
struct seq_file *s, int pin)
{
const struct pinconf_ops *ops = pctldev->desc->confops;
/* no-op when not using generic pin config */
pinconf_generic_dump_pin(pctldev, s, pin);
pinctrl: add a pin config interface This add per-pin and per-group pin config interfaces for biasing, driving and other such electronic properties. The details of passed configurations are passed in an opaque unsigned long which may be dereferences to integer types, structs or lists on either side of the configuration interface. ChangeLog v1->v2: - Clear split of terminology: we now have pin controllers, and those may support two interfaces using vtables: pin multiplexing and pin configuration. - Break out pin configuration to its own C file, controllers may implement only config without mux, and vice versa, so keep each sub-functionality of pin controllers separate. Introduce CONFIG_PINCONF in Kconfig. - Implement some core logic around pin configuration in the pinconf.c file. - Remove UNKNOWN config states, these were just surplus baggage. - Remove FLOAT config state - HIGH_IMPEDANCE should be enough for everyone. - PIN_CONFIG_POWER_SOURCE added to handle switching the power supply for the pin logic between different sources - Explicit DISABLE config enums to turn schmitt-trigger, wakeup etc OFF. - Update documentation to reflect all the recent reasoning. ChangeLog v2->v3: - Twist API around to pass around arrays of config tuples instead of (param, value) pairs everywhere. - Explicit drive strength semantics for push/pull and similar drive modes, this shall be the number of drive stages vs nominal load impedance, which should match the actual electronics used in push/pull CMOS or TTY totempoles. - Drop load capacitance configuration - I probably don't know what I'm doing here so leave it out. - Drop PIN_CONFIG_INPUT_SCHMITT_OFF, instead the argument zero to PIN_CONFIG_INPUT_SCHMITT turns schmitt trigger off. - Drop PIN_CONFIG_NORMAL_POWER_MODE and have a well defined argument to PIN_CONFIG_LOW_POWER_MODE to get out of it instead. - Drop PIN_CONFIG_WAKEUP_ENABLE/DISABLE and just use PIN_CONFIG_WAKEUP with defined value zero to turn wakeup off. - Add PIN_CONFIG_INPUT_DEBOUNCE for configuring debounce time on input lines. - Fix a bug when we tried to configure pins for pin controllers without pinconf support. - Initialized debugfs properly so it works. - Initialize the mutex properly and lock around config tampering sections. - Check the return value from get_initial_config() properly. ChangeLog v3->v4: - Export the pin_config_get(), pin_config_set() and pin_config_group() functions. - Drop the entire concept of just getting initial config and keeping track of pin states internally, instead ask the pins what state they are in. Previous idea was plain wrong, if the device cannot keep track of its state, the driver should do it. - Drop the generic configuration layout, it seems this impose too much restriction on some pin controllers, so let them do things the way they want and split off support for generic config as an optional add-on. ChangeLog v4->v5: - Introduce two symmetric driver calls for group configuration, .pin_config_group_[get|set] and corresponding external calls. - Remove generic semantic meanings of return values from config calls, these belong in the generic config patch. Just pass the return value through instead. - Add a debugfs entry "pinconf-groups" to read status from group configuration only, also slam in a per-group debug callback in the pinconf_ops so custom drivers can display something meaningful for their pins. - Fix some dangling newline. - Drop dangling #else clause. - Update documentation to match the above. ChangeLog v5->v6: - Change to using a pin name as parameter for the [get|set]_config() functions, as suggested by Stephen Warren. This is more natural as names will be what a developer has access to in written documentation etc. ChangeLog v6->v7: - Refactor out by-pin and by-name get/set functions, only expose the by-name functions externally, expose the by-pin functions internally. - Show supported pin control functionality in the debugfs pinctrl-devices file. Acked-by: Stephen Warren <swarren@nvidia.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2011-10-19 20:14:33 +04:00
if (ops && ops->pin_config_dbg_show)
ops->pin_config_dbg_show(pctldev, s, pin);
}
static int pinconf_pins_show(struct seq_file *s, void *what)
{
struct pinctrl_dev *pctldev = s->private;
const struct pinconf_ops *ops = pctldev->desc->confops;
unsigned i, pin;
pinctrl: add a pin config interface This add per-pin and per-group pin config interfaces for biasing, driving and other such electronic properties. The details of passed configurations are passed in an opaque unsigned long which may be dereferences to integer types, structs or lists on either side of the configuration interface. ChangeLog v1->v2: - Clear split of terminology: we now have pin controllers, and those may support two interfaces using vtables: pin multiplexing and pin configuration. - Break out pin configuration to its own C file, controllers may implement only config without mux, and vice versa, so keep each sub-functionality of pin controllers separate. Introduce CONFIG_PINCONF in Kconfig. - Implement some core logic around pin configuration in the pinconf.c file. - Remove UNKNOWN config states, these were just surplus baggage. - Remove FLOAT config state - HIGH_IMPEDANCE should be enough for everyone. - PIN_CONFIG_POWER_SOURCE added to handle switching the power supply for the pin logic between different sources - Explicit DISABLE config enums to turn schmitt-trigger, wakeup etc OFF. - Update documentation to reflect all the recent reasoning. ChangeLog v2->v3: - Twist API around to pass around arrays of config tuples instead of (param, value) pairs everywhere. - Explicit drive strength semantics for push/pull and similar drive modes, this shall be the number of drive stages vs nominal load impedance, which should match the actual electronics used in push/pull CMOS or TTY totempoles. - Drop load capacitance configuration - I probably don't know what I'm doing here so leave it out. - Drop PIN_CONFIG_INPUT_SCHMITT_OFF, instead the argument zero to PIN_CONFIG_INPUT_SCHMITT turns schmitt trigger off. - Drop PIN_CONFIG_NORMAL_POWER_MODE and have a well defined argument to PIN_CONFIG_LOW_POWER_MODE to get out of it instead. - Drop PIN_CONFIG_WAKEUP_ENABLE/DISABLE and just use PIN_CONFIG_WAKEUP with defined value zero to turn wakeup off. - Add PIN_CONFIG_INPUT_DEBOUNCE for configuring debounce time on input lines. - Fix a bug when we tried to configure pins for pin controllers without pinconf support. - Initialized debugfs properly so it works. - Initialize the mutex properly and lock around config tampering sections. - Check the return value from get_initial_config() properly. ChangeLog v3->v4: - Export the pin_config_get(), pin_config_set() and pin_config_group() functions. - Drop the entire concept of just getting initial config and keeping track of pin states internally, instead ask the pins what state they are in. Previous idea was plain wrong, if the device cannot keep track of its state, the driver should do it. - Drop the generic configuration layout, it seems this impose too much restriction on some pin controllers, so let them do things the way they want and split off support for generic config as an optional add-on. ChangeLog v4->v5: - Introduce two symmetric driver calls for group configuration, .pin_config_group_[get|set] and corresponding external calls. - Remove generic semantic meanings of return values from config calls, these belong in the generic config patch. Just pass the return value through instead. - Add a debugfs entry "pinconf-groups" to read status from group configuration only, also slam in a per-group debug callback in the pinconf_ops so custom drivers can display something meaningful for their pins. - Fix some dangling newline. - Drop dangling #else clause. - Update documentation to match the above. ChangeLog v5->v6: - Change to using a pin name as parameter for the [get|set]_config() functions, as suggested by Stephen Warren. This is more natural as names will be what a developer has access to in written documentation etc. ChangeLog v6->v7: - Refactor out by-pin and by-name get/set functions, only expose the by-name functions externally, expose the by-pin functions internally. - Show supported pin control functionality in the debugfs pinctrl-devices file. Acked-by: Stephen Warren <swarren@nvidia.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2011-10-19 20:14:33 +04:00
if (!ops || !ops->pin_config_get)
return 0;
pinctrl: add a pin config interface This add per-pin and per-group pin config interfaces for biasing, driving and other such electronic properties. The details of passed configurations are passed in an opaque unsigned long which may be dereferences to integer types, structs or lists on either side of the configuration interface. ChangeLog v1->v2: - Clear split of terminology: we now have pin controllers, and those may support two interfaces using vtables: pin multiplexing and pin configuration. - Break out pin configuration to its own C file, controllers may implement only config without mux, and vice versa, so keep each sub-functionality of pin controllers separate. Introduce CONFIG_PINCONF in Kconfig. - Implement some core logic around pin configuration in the pinconf.c file. - Remove UNKNOWN config states, these were just surplus baggage. - Remove FLOAT config state - HIGH_IMPEDANCE should be enough for everyone. - PIN_CONFIG_POWER_SOURCE added to handle switching the power supply for the pin logic between different sources - Explicit DISABLE config enums to turn schmitt-trigger, wakeup etc OFF. - Update documentation to reflect all the recent reasoning. ChangeLog v2->v3: - Twist API around to pass around arrays of config tuples instead of (param, value) pairs everywhere. - Explicit drive strength semantics for push/pull and similar drive modes, this shall be the number of drive stages vs nominal load impedance, which should match the actual electronics used in push/pull CMOS or TTY totempoles. - Drop load capacitance configuration - I probably don't know what I'm doing here so leave it out. - Drop PIN_CONFIG_INPUT_SCHMITT_OFF, instead the argument zero to PIN_CONFIG_INPUT_SCHMITT turns schmitt trigger off. - Drop PIN_CONFIG_NORMAL_POWER_MODE and have a well defined argument to PIN_CONFIG_LOW_POWER_MODE to get out of it instead. - Drop PIN_CONFIG_WAKEUP_ENABLE/DISABLE and just use PIN_CONFIG_WAKEUP with defined value zero to turn wakeup off. - Add PIN_CONFIG_INPUT_DEBOUNCE for configuring debounce time on input lines. - Fix a bug when we tried to configure pins for pin controllers without pinconf support. - Initialized debugfs properly so it works. - Initialize the mutex properly and lock around config tampering sections. - Check the return value from get_initial_config() properly. ChangeLog v3->v4: - Export the pin_config_get(), pin_config_set() and pin_config_group() functions. - Drop the entire concept of just getting initial config and keeping track of pin states internally, instead ask the pins what state they are in. Previous idea was plain wrong, if the device cannot keep track of its state, the driver should do it. - Drop the generic configuration layout, it seems this impose too much restriction on some pin controllers, so let them do things the way they want and split off support for generic config as an optional add-on. ChangeLog v4->v5: - Introduce two symmetric driver calls for group configuration, .pin_config_group_[get|set] and corresponding external calls. - Remove generic semantic meanings of return values from config calls, these belong in the generic config patch. Just pass the return value through instead. - Add a debugfs entry "pinconf-groups" to read status from group configuration only, also slam in a per-group debug callback in the pinconf_ops so custom drivers can display something meaningful for their pins. - Fix some dangling newline. - Drop dangling #else clause. - Update documentation to match the above. ChangeLog v5->v6: - Change to using a pin name as parameter for the [get|set]_config() functions, as suggested by Stephen Warren. This is more natural as names will be what a developer has access to in written documentation etc. ChangeLog v6->v7: - Refactor out by-pin and by-name get/set functions, only expose the by-name functions externally, expose the by-pin functions internally. - Show supported pin control functionality in the debugfs pinctrl-devices file. Acked-by: Stephen Warren <swarren@nvidia.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2011-10-19 20:14:33 +04:00
seq_puts(s, "Pin config settings per pin\n");
seq_puts(s, "Format: pin (name): configs\n");
pinctrl: add a pin config interface This add per-pin and per-group pin config interfaces for biasing, driving and other such electronic properties. The details of passed configurations are passed in an opaque unsigned long which may be dereferences to integer types, structs or lists on either side of the configuration interface. ChangeLog v1->v2: - Clear split of terminology: we now have pin controllers, and those may support two interfaces using vtables: pin multiplexing and pin configuration. - Break out pin configuration to its own C file, controllers may implement only config without mux, and vice versa, so keep each sub-functionality of pin controllers separate. Introduce CONFIG_PINCONF in Kconfig. - Implement some core logic around pin configuration in the pinconf.c file. - Remove UNKNOWN config states, these were just surplus baggage. - Remove FLOAT config state - HIGH_IMPEDANCE should be enough for everyone. - PIN_CONFIG_POWER_SOURCE added to handle switching the power supply for the pin logic between different sources - Explicit DISABLE config enums to turn schmitt-trigger, wakeup etc OFF. - Update documentation to reflect all the recent reasoning. ChangeLog v2->v3: - Twist API around to pass around arrays of config tuples instead of (param, value) pairs everywhere. - Explicit drive strength semantics for push/pull and similar drive modes, this shall be the number of drive stages vs nominal load impedance, which should match the actual electronics used in push/pull CMOS or TTY totempoles. - Drop load capacitance configuration - I probably don't know what I'm doing here so leave it out. - Drop PIN_CONFIG_INPUT_SCHMITT_OFF, instead the argument zero to PIN_CONFIG_INPUT_SCHMITT turns schmitt trigger off. - Drop PIN_CONFIG_NORMAL_POWER_MODE and have a well defined argument to PIN_CONFIG_LOW_POWER_MODE to get out of it instead. - Drop PIN_CONFIG_WAKEUP_ENABLE/DISABLE and just use PIN_CONFIG_WAKEUP with defined value zero to turn wakeup off. - Add PIN_CONFIG_INPUT_DEBOUNCE for configuring debounce time on input lines. - Fix a bug when we tried to configure pins for pin controllers without pinconf support. - Initialized debugfs properly so it works. - Initialize the mutex properly and lock around config tampering sections. - Check the return value from get_initial_config() properly. ChangeLog v3->v4: - Export the pin_config_get(), pin_config_set() and pin_config_group() functions. - Drop the entire concept of just getting initial config and keeping track of pin states internally, instead ask the pins what state they are in. Previous idea was plain wrong, if the device cannot keep track of its state, the driver should do it. - Drop the generic configuration layout, it seems this impose too much restriction on some pin controllers, so let them do things the way they want and split off support for generic config as an optional add-on. ChangeLog v4->v5: - Introduce two symmetric driver calls for group configuration, .pin_config_group_[get|set] and corresponding external calls. - Remove generic semantic meanings of return values from config calls, these belong in the generic config patch. Just pass the return value through instead. - Add a debugfs entry "pinconf-groups" to read status from group configuration only, also slam in a per-group debug callback in the pinconf_ops so custom drivers can display something meaningful for their pins. - Fix some dangling newline. - Drop dangling #else clause. - Update documentation to match the above. ChangeLog v5->v6: - Change to using a pin name as parameter for the [get|set]_config() functions, as suggested by Stephen Warren. This is more natural as names will be what a developer has access to in written documentation etc. ChangeLog v6->v7: - Refactor out by-pin and by-name get/set functions, only expose the by-name functions externally, expose the by-pin functions internally. - Show supported pin control functionality in the debugfs pinctrl-devices file. Acked-by: Stephen Warren <swarren@nvidia.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2011-10-19 20:14:33 +04:00
pinctrl: move subsystem mutex to pinctrl_dev struct This mutex avoids deadlock in case of use of multiple pin controllers. Before this modification, by using a global mutex, deadlock appeared when, for example, a call to pinctrl_pins_show() locked the pinctrl_mutex, called the ops->pin_dbg_show of a particular pin controller. If this pin controller needs I2C access to retrieve configuration information and I2C driver is using pinctrl to drive its pins, a call to pinctrl_select_state() try to lock again pinctrl_mutex which leads to a deadlock. Notice that the mutex grab from the two direction functions was moved into pinctrl_gpio_direction(). For several cases, we can't replace pinctrl_mutex by pctldev->mutex, because at this stage, pctldev is not accessible : - pinctrl_get()/pinctrl_put() - pinctrl_register_maps() So add respectively pinctrl_list_mutex and pinctrl_maps_mutex in order to protect pinctrl_list and pinctrl_maps list instead. Reintroduce pinctrldev_list_mutex in find_pinctrl_by_of_node(), pinctrl_find_and_add_gpio_range() pinctrl_request_gpio(), pinctrl_free_gpio(), pinctrl_gpio_direction(), pinctrl_devices_show(), pinctrl_register() and pinctrl_unregister() to protect pinctrldev_list. Changes v2->v3: - Fix a missing EXPORT_SYMBOL_GPL() for pinctrl_select_state(). Changes v1->v2: - pinctrl_select_state_locked() is removed, all lock mechanism is located inside pinctrl_select_state(). When parsing the state->setting list, take the per-pin-controller driver lock. (Patrice). - Introduce pinctrldev_list_mutex to protect pinctrldev_list in all functions which parse or modify pictrldev_list. (Patrice). - move find_pinctrl_by_of_node() from pinctrl/devicetree.c to pinctrl/core.c in order to protect pinctrldev_list. (Patrice). - Sink mutex:es into some functions and remove some _locked variants down to where the lists are actually accessed to make things simpler. (Linus) - Drop *all* mutexes completely from pinctrl_lookup_state() and pinctrl_select_state() - no relevant mutex was taken and it was unclear what this was protecting against. (Linus) Reported by : Seraphin Bonnaffe <seraphin.bonnaffe@stericsson.com> Signed-off-by: Patrice Chotard <patrice.chotard@st.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2013-04-11 13:01:27 +04:00
mutex_lock(&pctldev->mutex);
pinctrl: fix and simplify locking There are many problems with the current pinctrl locking: struct pinctrl_dev's gpio_ranges_lock isn't effective; pinctrl_match_gpio_range() only holds this lock while searching for a gpio range, but the found range is return and manipulated after releading the lock. This could allow pinctrl_remove_gpio_range() for that range while it is in use, and the caller may very well delete the range after removing it, causing pinctrl code to touch the now-free range object. Solving this requires the introduction of a higher-level lock, at least a lock per pin controller, which both gpio range registration and pinctrl_get()/put() will acquire. There is missing locking on HW programming; pin controllers may pack the configuration for different pins/groups/config options/... into one register, and hence have to read-modify-write the register. This needs to be protected, but currently isn't. Related, a future change will add a "complete" op to the pin controller drivers, the idea being that each state's programming will be programmed into the pinctrl driver followed by the "complete" call, which may e.g. flush a register cache to HW. For this to work, it must not be possible to interleave the pinctrl driver calls for different devices. As above, solving this requires the introduction of a higher-level lock, at least a lock per pin controller, which will be held for the duration of any pinctrl_enable()/disable() call. However, each pinctrl mapping table entry may affect a different pin controller if necessary. Hence, with a per-pin-controller lock, almost any pinctrl API may need to acquire multiple locks, one per controller. To avoid deadlock, these would need to be acquired in the same order in all cases. This is extremely difficult to implement in the case of pinctrl_get(), which doesn't know which pin controllers to lock until it has parsed the entire mapping table, since it contains somewhat arbitrary data. The simplest solution here is to introduce a single lock that covers all pin controllers at once. This will be acquired by all pinctrl APIs. This then makes struct pinctrl's mutex irrelevant, since that single lock will always be held whenever this mutex is currently held. Signed-off-by: Stephen Warren <swarren@nvidia.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2012-03-03 00:05:44 +04:00
/* The pin number can be retrived from the pin controller descriptor */
for (i = 0; i < pctldev->desc->npins; i++) {
pinctrl: add a pin config interface This add per-pin and per-group pin config interfaces for biasing, driving and other such electronic properties. The details of passed configurations are passed in an opaque unsigned long which may be dereferences to integer types, structs or lists on either side of the configuration interface. ChangeLog v1->v2: - Clear split of terminology: we now have pin controllers, and those may support two interfaces using vtables: pin multiplexing and pin configuration. - Break out pin configuration to its own C file, controllers may implement only config without mux, and vice versa, so keep each sub-functionality of pin controllers separate. Introduce CONFIG_PINCONF in Kconfig. - Implement some core logic around pin configuration in the pinconf.c file. - Remove UNKNOWN config states, these were just surplus baggage. - Remove FLOAT config state - HIGH_IMPEDANCE should be enough for everyone. - PIN_CONFIG_POWER_SOURCE added to handle switching the power supply for the pin logic between different sources - Explicit DISABLE config enums to turn schmitt-trigger, wakeup etc OFF. - Update documentation to reflect all the recent reasoning. ChangeLog v2->v3: - Twist API around to pass around arrays of config tuples instead of (param, value) pairs everywhere. - Explicit drive strength semantics for push/pull and similar drive modes, this shall be the number of drive stages vs nominal load impedance, which should match the actual electronics used in push/pull CMOS or TTY totempoles. - Drop load capacitance configuration - I probably don't know what I'm doing here so leave it out. - Drop PIN_CONFIG_INPUT_SCHMITT_OFF, instead the argument zero to PIN_CONFIG_INPUT_SCHMITT turns schmitt trigger off. - Drop PIN_CONFIG_NORMAL_POWER_MODE and have a well defined argument to PIN_CONFIG_LOW_POWER_MODE to get out of it instead. - Drop PIN_CONFIG_WAKEUP_ENABLE/DISABLE and just use PIN_CONFIG_WAKEUP with defined value zero to turn wakeup off. - Add PIN_CONFIG_INPUT_DEBOUNCE for configuring debounce time on input lines. - Fix a bug when we tried to configure pins for pin controllers without pinconf support. - Initialized debugfs properly so it works. - Initialize the mutex properly and lock around config tampering sections. - Check the return value from get_initial_config() properly. ChangeLog v3->v4: - Export the pin_config_get(), pin_config_set() and pin_config_group() functions. - Drop the entire concept of just getting initial config and keeping track of pin states internally, instead ask the pins what state they are in. Previous idea was plain wrong, if the device cannot keep track of its state, the driver should do it. - Drop the generic configuration layout, it seems this impose too much restriction on some pin controllers, so let them do things the way they want and split off support for generic config as an optional add-on. ChangeLog v4->v5: - Introduce two symmetric driver calls for group configuration, .pin_config_group_[get|set] and corresponding external calls. - Remove generic semantic meanings of return values from config calls, these belong in the generic config patch. Just pass the return value through instead. - Add a debugfs entry "pinconf-groups" to read status from group configuration only, also slam in a per-group debug callback in the pinconf_ops so custom drivers can display something meaningful for their pins. - Fix some dangling newline. - Drop dangling #else clause. - Update documentation to match the above. ChangeLog v5->v6: - Change to using a pin name as parameter for the [get|set]_config() functions, as suggested by Stephen Warren. This is more natural as names will be what a developer has access to in written documentation etc. ChangeLog v6->v7: - Refactor out by-pin and by-name get/set functions, only expose the by-name functions externally, expose the by-pin functions internally. - Show supported pin control functionality in the debugfs pinctrl-devices file. Acked-by: Stephen Warren <swarren@nvidia.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2011-10-19 20:14:33 +04:00
struct pin_desc *desc;
pin = pctldev->desc->pins[i].number;
pinctrl: add a pin config interface This add per-pin and per-group pin config interfaces for biasing, driving and other such electronic properties. The details of passed configurations are passed in an opaque unsigned long which may be dereferences to integer types, structs or lists on either side of the configuration interface. ChangeLog v1->v2: - Clear split of terminology: we now have pin controllers, and those may support two interfaces using vtables: pin multiplexing and pin configuration. - Break out pin configuration to its own C file, controllers may implement only config without mux, and vice versa, so keep each sub-functionality of pin controllers separate. Introduce CONFIG_PINCONF in Kconfig. - Implement some core logic around pin configuration in the pinconf.c file. - Remove UNKNOWN config states, these were just surplus baggage. - Remove FLOAT config state - HIGH_IMPEDANCE should be enough for everyone. - PIN_CONFIG_POWER_SOURCE added to handle switching the power supply for the pin logic between different sources - Explicit DISABLE config enums to turn schmitt-trigger, wakeup etc OFF. - Update documentation to reflect all the recent reasoning. ChangeLog v2->v3: - Twist API around to pass around arrays of config tuples instead of (param, value) pairs everywhere. - Explicit drive strength semantics for push/pull and similar drive modes, this shall be the number of drive stages vs nominal load impedance, which should match the actual electronics used in push/pull CMOS or TTY totempoles. - Drop load capacitance configuration - I probably don't know what I'm doing here so leave it out. - Drop PIN_CONFIG_INPUT_SCHMITT_OFF, instead the argument zero to PIN_CONFIG_INPUT_SCHMITT turns schmitt trigger off. - Drop PIN_CONFIG_NORMAL_POWER_MODE and have a well defined argument to PIN_CONFIG_LOW_POWER_MODE to get out of it instead. - Drop PIN_CONFIG_WAKEUP_ENABLE/DISABLE and just use PIN_CONFIG_WAKEUP with defined value zero to turn wakeup off. - Add PIN_CONFIG_INPUT_DEBOUNCE for configuring debounce time on input lines. - Fix a bug when we tried to configure pins for pin controllers without pinconf support. - Initialized debugfs properly so it works. - Initialize the mutex properly and lock around config tampering sections. - Check the return value from get_initial_config() properly. ChangeLog v3->v4: - Export the pin_config_get(), pin_config_set() and pin_config_group() functions. - Drop the entire concept of just getting initial config and keeping track of pin states internally, instead ask the pins what state they are in. Previous idea was plain wrong, if the device cannot keep track of its state, the driver should do it. - Drop the generic configuration layout, it seems this impose too much restriction on some pin controllers, so let them do things the way they want and split off support for generic config as an optional add-on. ChangeLog v4->v5: - Introduce two symmetric driver calls for group configuration, .pin_config_group_[get|set] and corresponding external calls. - Remove generic semantic meanings of return values from config calls, these belong in the generic config patch. Just pass the return value through instead. - Add a debugfs entry "pinconf-groups" to read status from group configuration only, also slam in a per-group debug callback in the pinconf_ops so custom drivers can display something meaningful for their pins. - Fix some dangling newline. - Drop dangling #else clause. - Update documentation to match the above. ChangeLog v5->v6: - Change to using a pin name as parameter for the [get|set]_config() functions, as suggested by Stephen Warren. This is more natural as names will be what a developer has access to in written documentation etc. ChangeLog v6->v7: - Refactor out by-pin and by-name get/set functions, only expose the by-name functions externally, expose the by-pin functions internally. - Show supported pin control functionality in the debugfs pinctrl-devices file. Acked-by: Stephen Warren <swarren@nvidia.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2011-10-19 20:14:33 +04:00
desc = pin_desc_get(pctldev, pin);
/* Skip if we cannot search the pin */
pinctrl: add a pin config interface This add per-pin and per-group pin config interfaces for biasing, driving and other such electronic properties. The details of passed configurations are passed in an opaque unsigned long which may be dereferences to integer types, structs or lists on either side of the configuration interface. ChangeLog v1->v2: - Clear split of terminology: we now have pin controllers, and those may support two interfaces using vtables: pin multiplexing and pin configuration. - Break out pin configuration to its own C file, controllers may implement only config without mux, and vice versa, so keep each sub-functionality of pin controllers separate. Introduce CONFIG_PINCONF in Kconfig. - Implement some core logic around pin configuration in the pinconf.c file. - Remove UNKNOWN config states, these were just surplus baggage. - Remove FLOAT config state - HIGH_IMPEDANCE should be enough for everyone. - PIN_CONFIG_POWER_SOURCE added to handle switching the power supply for the pin logic between different sources - Explicit DISABLE config enums to turn schmitt-trigger, wakeup etc OFF. - Update documentation to reflect all the recent reasoning. ChangeLog v2->v3: - Twist API around to pass around arrays of config tuples instead of (param, value) pairs everywhere. - Explicit drive strength semantics for push/pull and similar drive modes, this shall be the number of drive stages vs nominal load impedance, which should match the actual electronics used in push/pull CMOS or TTY totempoles. - Drop load capacitance configuration - I probably don't know what I'm doing here so leave it out. - Drop PIN_CONFIG_INPUT_SCHMITT_OFF, instead the argument zero to PIN_CONFIG_INPUT_SCHMITT turns schmitt trigger off. - Drop PIN_CONFIG_NORMAL_POWER_MODE and have a well defined argument to PIN_CONFIG_LOW_POWER_MODE to get out of it instead. - Drop PIN_CONFIG_WAKEUP_ENABLE/DISABLE and just use PIN_CONFIG_WAKEUP with defined value zero to turn wakeup off. - Add PIN_CONFIG_INPUT_DEBOUNCE for configuring debounce time on input lines. - Fix a bug when we tried to configure pins for pin controllers without pinconf support. - Initialized debugfs properly so it works. - Initialize the mutex properly and lock around config tampering sections. - Check the return value from get_initial_config() properly. ChangeLog v3->v4: - Export the pin_config_get(), pin_config_set() and pin_config_group() functions. - Drop the entire concept of just getting initial config and keeping track of pin states internally, instead ask the pins what state they are in. Previous idea was plain wrong, if the device cannot keep track of its state, the driver should do it. - Drop the generic configuration layout, it seems this impose too much restriction on some pin controllers, so let them do things the way they want and split off support for generic config as an optional add-on. ChangeLog v4->v5: - Introduce two symmetric driver calls for group configuration, .pin_config_group_[get|set] and corresponding external calls. - Remove generic semantic meanings of return values from config calls, these belong in the generic config patch. Just pass the return value through instead. - Add a debugfs entry "pinconf-groups" to read status from group configuration only, also slam in a per-group debug callback in the pinconf_ops so custom drivers can display something meaningful for their pins. - Fix some dangling newline. - Drop dangling #else clause. - Update documentation to match the above. ChangeLog v5->v6: - Change to using a pin name as parameter for the [get|set]_config() functions, as suggested by Stephen Warren. This is more natural as names will be what a developer has access to in written documentation etc. ChangeLog v6->v7: - Refactor out by-pin and by-name get/set functions, only expose the by-name functions externally, expose the by-pin functions internally. - Show supported pin control functionality in the debugfs pinctrl-devices file. Acked-by: Stephen Warren <swarren@nvidia.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2011-10-19 20:14:33 +04:00
if (desc == NULL)
continue;
seq_printf(s, "pin %d (%s):", pin,
desc->name ? desc->name : "unnamed");
pinconf_dump_pin(pctldev, s, pin);
seq_printf(s, "\n");
}
pinctrl: move subsystem mutex to pinctrl_dev struct This mutex avoids deadlock in case of use of multiple pin controllers. Before this modification, by using a global mutex, deadlock appeared when, for example, a call to pinctrl_pins_show() locked the pinctrl_mutex, called the ops->pin_dbg_show of a particular pin controller. If this pin controller needs I2C access to retrieve configuration information and I2C driver is using pinctrl to drive its pins, a call to pinctrl_select_state() try to lock again pinctrl_mutex which leads to a deadlock. Notice that the mutex grab from the two direction functions was moved into pinctrl_gpio_direction(). For several cases, we can't replace pinctrl_mutex by pctldev->mutex, because at this stage, pctldev is not accessible : - pinctrl_get()/pinctrl_put() - pinctrl_register_maps() So add respectively pinctrl_list_mutex and pinctrl_maps_mutex in order to protect pinctrl_list and pinctrl_maps list instead. Reintroduce pinctrldev_list_mutex in find_pinctrl_by_of_node(), pinctrl_find_and_add_gpio_range() pinctrl_request_gpio(), pinctrl_free_gpio(), pinctrl_gpio_direction(), pinctrl_devices_show(), pinctrl_register() and pinctrl_unregister() to protect pinctrldev_list. Changes v2->v3: - Fix a missing EXPORT_SYMBOL_GPL() for pinctrl_select_state(). Changes v1->v2: - pinctrl_select_state_locked() is removed, all lock mechanism is located inside pinctrl_select_state(). When parsing the state->setting list, take the per-pin-controller driver lock. (Patrice). - Introduce pinctrldev_list_mutex to protect pinctrldev_list in all functions which parse or modify pictrldev_list. (Patrice). - move find_pinctrl_by_of_node() from pinctrl/devicetree.c to pinctrl/core.c in order to protect pinctrldev_list. (Patrice). - Sink mutex:es into some functions and remove some _locked variants down to where the lists are actually accessed to make things simpler. (Linus) - Drop *all* mutexes completely from pinctrl_lookup_state() and pinctrl_select_state() - no relevant mutex was taken and it was unclear what this was protecting against. (Linus) Reported by : Seraphin Bonnaffe <seraphin.bonnaffe@stericsson.com> Signed-off-by: Patrice Chotard <patrice.chotard@st.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2013-04-11 13:01:27 +04:00
mutex_unlock(&pctldev->mutex);
pinctrl: fix and simplify locking There are many problems with the current pinctrl locking: struct pinctrl_dev's gpio_ranges_lock isn't effective; pinctrl_match_gpio_range() only holds this lock while searching for a gpio range, but the found range is return and manipulated after releading the lock. This could allow pinctrl_remove_gpio_range() for that range while it is in use, and the caller may very well delete the range after removing it, causing pinctrl code to touch the now-free range object. Solving this requires the introduction of a higher-level lock, at least a lock per pin controller, which both gpio range registration and pinctrl_get()/put() will acquire. There is missing locking on HW programming; pin controllers may pack the configuration for different pins/groups/config options/... into one register, and hence have to read-modify-write the register. This needs to be protected, but currently isn't. Related, a future change will add a "complete" op to the pin controller drivers, the idea being that each state's programming will be programmed into the pinctrl driver followed by the "complete" call, which may e.g. flush a register cache to HW. For this to work, it must not be possible to interleave the pinctrl driver calls for different devices. As above, solving this requires the introduction of a higher-level lock, at least a lock per pin controller, which will be held for the duration of any pinctrl_enable()/disable() call. However, each pinctrl mapping table entry may affect a different pin controller if necessary. Hence, with a per-pin-controller lock, almost any pinctrl API may need to acquire multiple locks, one per controller. To avoid deadlock, these would need to be acquired in the same order in all cases. This is extremely difficult to implement in the case of pinctrl_get(), which doesn't know which pin controllers to lock until it has parsed the entire mapping table, since it contains somewhat arbitrary data. The simplest solution here is to introduce a single lock that covers all pin controllers at once. This will be acquired by all pinctrl APIs. This then makes struct pinctrl's mutex irrelevant, since that single lock will always be held whenever this mutex is currently held. Signed-off-by: Stephen Warren <swarren@nvidia.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2012-03-03 00:05:44 +04:00
pinctrl: add a pin config interface This add per-pin and per-group pin config interfaces for biasing, driving and other such electronic properties. The details of passed configurations are passed in an opaque unsigned long which may be dereferences to integer types, structs or lists on either side of the configuration interface. ChangeLog v1->v2: - Clear split of terminology: we now have pin controllers, and those may support two interfaces using vtables: pin multiplexing and pin configuration. - Break out pin configuration to its own C file, controllers may implement only config without mux, and vice versa, so keep each sub-functionality of pin controllers separate. Introduce CONFIG_PINCONF in Kconfig. - Implement some core logic around pin configuration in the pinconf.c file. - Remove UNKNOWN config states, these were just surplus baggage. - Remove FLOAT config state - HIGH_IMPEDANCE should be enough for everyone. - PIN_CONFIG_POWER_SOURCE added to handle switching the power supply for the pin logic between different sources - Explicit DISABLE config enums to turn schmitt-trigger, wakeup etc OFF. - Update documentation to reflect all the recent reasoning. ChangeLog v2->v3: - Twist API around to pass around arrays of config tuples instead of (param, value) pairs everywhere. - Explicit drive strength semantics for push/pull and similar drive modes, this shall be the number of drive stages vs nominal load impedance, which should match the actual electronics used in push/pull CMOS or TTY totempoles. - Drop load capacitance configuration - I probably don't know what I'm doing here so leave it out. - Drop PIN_CONFIG_INPUT_SCHMITT_OFF, instead the argument zero to PIN_CONFIG_INPUT_SCHMITT turns schmitt trigger off. - Drop PIN_CONFIG_NORMAL_POWER_MODE and have a well defined argument to PIN_CONFIG_LOW_POWER_MODE to get out of it instead. - Drop PIN_CONFIG_WAKEUP_ENABLE/DISABLE and just use PIN_CONFIG_WAKEUP with defined value zero to turn wakeup off. - Add PIN_CONFIG_INPUT_DEBOUNCE for configuring debounce time on input lines. - Fix a bug when we tried to configure pins for pin controllers without pinconf support. - Initialized debugfs properly so it works. - Initialize the mutex properly and lock around config tampering sections. - Check the return value from get_initial_config() properly. ChangeLog v3->v4: - Export the pin_config_get(), pin_config_set() and pin_config_group() functions. - Drop the entire concept of just getting initial config and keeping track of pin states internally, instead ask the pins what state they are in. Previous idea was plain wrong, if the device cannot keep track of its state, the driver should do it. - Drop the generic configuration layout, it seems this impose too much restriction on some pin controllers, so let them do things the way they want and split off support for generic config as an optional add-on. ChangeLog v4->v5: - Introduce two symmetric driver calls for group configuration, .pin_config_group_[get|set] and corresponding external calls. - Remove generic semantic meanings of return values from config calls, these belong in the generic config patch. Just pass the return value through instead. - Add a debugfs entry "pinconf-groups" to read status from group configuration only, also slam in a per-group debug callback in the pinconf_ops so custom drivers can display something meaningful for their pins. - Fix some dangling newline. - Drop dangling #else clause. - Update documentation to match the above. ChangeLog v5->v6: - Change to using a pin name as parameter for the [get|set]_config() functions, as suggested by Stephen Warren. This is more natural as names will be what a developer has access to in written documentation etc. ChangeLog v6->v7: - Refactor out by-pin and by-name get/set functions, only expose the by-name functions externally, expose the by-pin functions internally. - Show supported pin control functionality in the debugfs pinctrl-devices file. Acked-by: Stephen Warren <swarren@nvidia.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2011-10-19 20:14:33 +04:00
return 0;
}
static void pinconf_dump_group(struct pinctrl_dev *pctldev,
struct seq_file *s, unsigned selector,
const char *gname)
{
const struct pinconf_ops *ops = pctldev->desc->confops;
/* no-op when not using generic pin config */
pinconf_generic_dump_group(pctldev, s, gname);
pinctrl: add a pin config interface This add per-pin and per-group pin config interfaces for biasing, driving and other such electronic properties. The details of passed configurations are passed in an opaque unsigned long which may be dereferences to integer types, structs or lists on either side of the configuration interface. ChangeLog v1->v2: - Clear split of terminology: we now have pin controllers, and those may support two interfaces using vtables: pin multiplexing and pin configuration. - Break out pin configuration to its own C file, controllers may implement only config without mux, and vice versa, so keep each sub-functionality of pin controllers separate. Introduce CONFIG_PINCONF in Kconfig. - Implement some core logic around pin configuration in the pinconf.c file. - Remove UNKNOWN config states, these were just surplus baggage. - Remove FLOAT config state - HIGH_IMPEDANCE should be enough for everyone. - PIN_CONFIG_POWER_SOURCE added to handle switching the power supply for the pin logic between different sources - Explicit DISABLE config enums to turn schmitt-trigger, wakeup etc OFF. - Update documentation to reflect all the recent reasoning. ChangeLog v2->v3: - Twist API around to pass around arrays of config tuples instead of (param, value) pairs everywhere. - Explicit drive strength semantics for push/pull and similar drive modes, this shall be the number of drive stages vs nominal load impedance, which should match the actual electronics used in push/pull CMOS or TTY totempoles. - Drop load capacitance configuration - I probably don't know what I'm doing here so leave it out. - Drop PIN_CONFIG_INPUT_SCHMITT_OFF, instead the argument zero to PIN_CONFIG_INPUT_SCHMITT turns schmitt trigger off. - Drop PIN_CONFIG_NORMAL_POWER_MODE and have a well defined argument to PIN_CONFIG_LOW_POWER_MODE to get out of it instead. - Drop PIN_CONFIG_WAKEUP_ENABLE/DISABLE and just use PIN_CONFIG_WAKEUP with defined value zero to turn wakeup off. - Add PIN_CONFIG_INPUT_DEBOUNCE for configuring debounce time on input lines. - Fix a bug when we tried to configure pins for pin controllers without pinconf support. - Initialized debugfs properly so it works. - Initialize the mutex properly and lock around config tampering sections. - Check the return value from get_initial_config() properly. ChangeLog v3->v4: - Export the pin_config_get(), pin_config_set() and pin_config_group() functions. - Drop the entire concept of just getting initial config and keeping track of pin states internally, instead ask the pins what state they are in. Previous idea was plain wrong, if the device cannot keep track of its state, the driver should do it. - Drop the generic configuration layout, it seems this impose too much restriction on some pin controllers, so let them do things the way they want and split off support for generic config as an optional add-on. ChangeLog v4->v5: - Introduce two symmetric driver calls for group configuration, .pin_config_group_[get|set] and corresponding external calls. - Remove generic semantic meanings of return values from config calls, these belong in the generic config patch. Just pass the return value through instead. - Add a debugfs entry "pinconf-groups" to read status from group configuration only, also slam in a per-group debug callback in the pinconf_ops so custom drivers can display something meaningful for their pins. - Fix some dangling newline. - Drop dangling #else clause. - Update documentation to match the above. ChangeLog v5->v6: - Change to using a pin name as parameter for the [get|set]_config() functions, as suggested by Stephen Warren. This is more natural as names will be what a developer has access to in written documentation etc. ChangeLog v6->v7: - Refactor out by-pin and by-name get/set functions, only expose the by-name functions externally, expose the by-pin functions internally. - Show supported pin control functionality in the debugfs pinctrl-devices file. Acked-by: Stephen Warren <swarren@nvidia.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2011-10-19 20:14:33 +04:00
if (ops && ops->pin_config_group_dbg_show)
ops->pin_config_group_dbg_show(pctldev, s, selector);
}
static int pinconf_groups_show(struct seq_file *s, void *what)
{
struct pinctrl_dev *pctldev = s->private;
const struct pinctrl_ops *pctlops = pctldev->desc->pctlops;
const struct pinconf_ops *ops = pctldev->desc->confops;
unsigned ngroups = pctlops->get_groups_count(pctldev);
pinctrl: add a pin config interface This add per-pin and per-group pin config interfaces for biasing, driving and other such electronic properties. The details of passed configurations are passed in an opaque unsigned long which may be dereferences to integer types, structs or lists on either side of the configuration interface. ChangeLog v1->v2: - Clear split of terminology: we now have pin controllers, and those may support two interfaces using vtables: pin multiplexing and pin configuration. - Break out pin configuration to its own C file, controllers may implement only config without mux, and vice versa, so keep each sub-functionality of pin controllers separate. Introduce CONFIG_PINCONF in Kconfig. - Implement some core logic around pin configuration in the pinconf.c file. - Remove UNKNOWN config states, these were just surplus baggage. - Remove FLOAT config state - HIGH_IMPEDANCE should be enough for everyone. - PIN_CONFIG_POWER_SOURCE added to handle switching the power supply for the pin logic between different sources - Explicit DISABLE config enums to turn schmitt-trigger, wakeup etc OFF. - Update documentation to reflect all the recent reasoning. ChangeLog v2->v3: - Twist API around to pass around arrays of config tuples instead of (param, value) pairs everywhere. - Explicit drive strength semantics for push/pull and similar drive modes, this shall be the number of drive stages vs nominal load impedance, which should match the actual electronics used in push/pull CMOS or TTY totempoles. - Drop load capacitance configuration - I probably don't know what I'm doing here so leave it out. - Drop PIN_CONFIG_INPUT_SCHMITT_OFF, instead the argument zero to PIN_CONFIG_INPUT_SCHMITT turns schmitt trigger off. - Drop PIN_CONFIG_NORMAL_POWER_MODE and have a well defined argument to PIN_CONFIG_LOW_POWER_MODE to get out of it instead. - Drop PIN_CONFIG_WAKEUP_ENABLE/DISABLE and just use PIN_CONFIG_WAKEUP with defined value zero to turn wakeup off. - Add PIN_CONFIG_INPUT_DEBOUNCE for configuring debounce time on input lines. - Fix a bug when we tried to configure pins for pin controllers without pinconf support. - Initialized debugfs properly so it works. - Initialize the mutex properly and lock around config tampering sections. - Check the return value from get_initial_config() properly. ChangeLog v3->v4: - Export the pin_config_get(), pin_config_set() and pin_config_group() functions. - Drop the entire concept of just getting initial config and keeping track of pin states internally, instead ask the pins what state they are in. Previous idea was plain wrong, if the device cannot keep track of its state, the driver should do it. - Drop the generic configuration layout, it seems this impose too much restriction on some pin controllers, so let them do things the way they want and split off support for generic config as an optional add-on. ChangeLog v4->v5: - Introduce two symmetric driver calls for group configuration, .pin_config_group_[get|set] and corresponding external calls. - Remove generic semantic meanings of return values from config calls, these belong in the generic config patch. Just pass the return value through instead. - Add a debugfs entry "pinconf-groups" to read status from group configuration only, also slam in a per-group debug callback in the pinconf_ops so custom drivers can display something meaningful for their pins. - Fix some dangling newline. - Drop dangling #else clause. - Update documentation to match the above. ChangeLog v5->v6: - Change to using a pin name as parameter for the [get|set]_config() functions, as suggested by Stephen Warren. This is more natural as names will be what a developer has access to in written documentation etc. ChangeLog v6->v7: - Refactor out by-pin and by-name get/set functions, only expose the by-name functions externally, expose the by-pin functions internally. - Show supported pin control functionality in the debugfs pinctrl-devices file. Acked-by: Stephen Warren <swarren@nvidia.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2011-10-19 20:14:33 +04:00
unsigned selector = 0;
if (!ops || !ops->pin_config_group_get)
return 0;
seq_puts(s, "Pin config settings per pin group\n");
seq_puts(s, "Format: group (name): configs\n");
pinctrl: add a pin config interface This add per-pin and per-group pin config interfaces for biasing, driving and other such electronic properties. The details of passed configurations are passed in an opaque unsigned long which may be dereferences to integer types, structs or lists on either side of the configuration interface. ChangeLog v1->v2: - Clear split of terminology: we now have pin controllers, and those may support two interfaces using vtables: pin multiplexing and pin configuration. - Break out pin configuration to its own C file, controllers may implement only config without mux, and vice versa, so keep each sub-functionality of pin controllers separate. Introduce CONFIG_PINCONF in Kconfig. - Implement some core logic around pin configuration in the pinconf.c file. - Remove UNKNOWN config states, these were just surplus baggage. - Remove FLOAT config state - HIGH_IMPEDANCE should be enough for everyone. - PIN_CONFIG_POWER_SOURCE added to handle switching the power supply for the pin logic between different sources - Explicit DISABLE config enums to turn schmitt-trigger, wakeup etc OFF. - Update documentation to reflect all the recent reasoning. ChangeLog v2->v3: - Twist API around to pass around arrays of config tuples instead of (param, value) pairs everywhere. - Explicit drive strength semantics for push/pull and similar drive modes, this shall be the number of drive stages vs nominal load impedance, which should match the actual electronics used in push/pull CMOS or TTY totempoles. - Drop load capacitance configuration - I probably don't know what I'm doing here so leave it out. - Drop PIN_CONFIG_INPUT_SCHMITT_OFF, instead the argument zero to PIN_CONFIG_INPUT_SCHMITT turns schmitt trigger off. - Drop PIN_CONFIG_NORMAL_POWER_MODE and have a well defined argument to PIN_CONFIG_LOW_POWER_MODE to get out of it instead. - Drop PIN_CONFIG_WAKEUP_ENABLE/DISABLE and just use PIN_CONFIG_WAKEUP with defined value zero to turn wakeup off. - Add PIN_CONFIG_INPUT_DEBOUNCE for configuring debounce time on input lines. - Fix a bug when we tried to configure pins for pin controllers without pinconf support. - Initialized debugfs properly so it works. - Initialize the mutex properly and lock around config tampering sections. - Check the return value from get_initial_config() properly. ChangeLog v3->v4: - Export the pin_config_get(), pin_config_set() and pin_config_group() functions. - Drop the entire concept of just getting initial config and keeping track of pin states internally, instead ask the pins what state they are in. Previous idea was plain wrong, if the device cannot keep track of its state, the driver should do it. - Drop the generic configuration layout, it seems this impose too much restriction on some pin controllers, so let them do things the way they want and split off support for generic config as an optional add-on. ChangeLog v4->v5: - Introduce two symmetric driver calls for group configuration, .pin_config_group_[get|set] and corresponding external calls. - Remove generic semantic meanings of return values from config calls, these belong in the generic config patch. Just pass the return value through instead. - Add a debugfs entry "pinconf-groups" to read status from group configuration only, also slam in a per-group debug callback in the pinconf_ops so custom drivers can display something meaningful for their pins. - Fix some dangling newline. - Drop dangling #else clause. - Update documentation to match the above. ChangeLog v5->v6: - Change to using a pin name as parameter for the [get|set]_config() functions, as suggested by Stephen Warren. This is more natural as names will be what a developer has access to in written documentation etc. ChangeLog v6->v7: - Refactor out by-pin and by-name get/set functions, only expose the by-name functions externally, expose the by-pin functions internally. - Show supported pin control functionality in the debugfs pinctrl-devices file. Acked-by: Stephen Warren <swarren@nvidia.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2011-10-19 20:14:33 +04:00
while (selector < ngroups) {
pinctrl: add a pin config interface This add per-pin and per-group pin config interfaces for biasing, driving and other such electronic properties. The details of passed configurations are passed in an opaque unsigned long which may be dereferences to integer types, structs or lists on either side of the configuration interface. ChangeLog v1->v2: - Clear split of terminology: we now have pin controllers, and those may support two interfaces using vtables: pin multiplexing and pin configuration. - Break out pin configuration to its own C file, controllers may implement only config without mux, and vice versa, so keep each sub-functionality of pin controllers separate. Introduce CONFIG_PINCONF in Kconfig. - Implement some core logic around pin configuration in the pinconf.c file. - Remove UNKNOWN config states, these were just surplus baggage. - Remove FLOAT config state - HIGH_IMPEDANCE should be enough for everyone. - PIN_CONFIG_POWER_SOURCE added to handle switching the power supply for the pin logic between different sources - Explicit DISABLE config enums to turn schmitt-trigger, wakeup etc OFF. - Update documentation to reflect all the recent reasoning. ChangeLog v2->v3: - Twist API around to pass around arrays of config tuples instead of (param, value) pairs everywhere. - Explicit drive strength semantics for push/pull and similar drive modes, this shall be the number of drive stages vs nominal load impedance, which should match the actual electronics used in push/pull CMOS or TTY totempoles. - Drop load capacitance configuration - I probably don't know what I'm doing here so leave it out. - Drop PIN_CONFIG_INPUT_SCHMITT_OFF, instead the argument zero to PIN_CONFIG_INPUT_SCHMITT turns schmitt trigger off. - Drop PIN_CONFIG_NORMAL_POWER_MODE and have a well defined argument to PIN_CONFIG_LOW_POWER_MODE to get out of it instead. - Drop PIN_CONFIG_WAKEUP_ENABLE/DISABLE and just use PIN_CONFIG_WAKEUP with defined value zero to turn wakeup off. - Add PIN_CONFIG_INPUT_DEBOUNCE for configuring debounce time on input lines. - Fix a bug when we tried to configure pins for pin controllers without pinconf support. - Initialized debugfs properly so it works. - Initialize the mutex properly and lock around config tampering sections. - Check the return value from get_initial_config() properly. ChangeLog v3->v4: - Export the pin_config_get(), pin_config_set() and pin_config_group() functions. - Drop the entire concept of just getting initial config and keeping track of pin states internally, instead ask the pins what state they are in. Previous idea was plain wrong, if the device cannot keep track of its state, the driver should do it. - Drop the generic configuration layout, it seems this impose too much restriction on some pin controllers, so let them do things the way they want and split off support for generic config as an optional add-on. ChangeLog v4->v5: - Introduce two symmetric driver calls for group configuration, .pin_config_group_[get|set] and corresponding external calls. - Remove generic semantic meanings of return values from config calls, these belong in the generic config patch. Just pass the return value through instead. - Add a debugfs entry "pinconf-groups" to read status from group configuration only, also slam in a per-group debug callback in the pinconf_ops so custom drivers can display something meaningful for their pins. - Fix some dangling newline. - Drop dangling #else clause. - Update documentation to match the above. ChangeLog v5->v6: - Change to using a pin name as parameter for the [get|set]_config() functions, as suggested by Stephen Warren. This is more natural as names will be what a developer has access to in written documentation etc. ChangeLog v6->v7: - Refactor out by-pin and by-name get/set functions, only expose the by-name functions externally, expose the by-pin functions internally. - Show supported pin control functionality in the debugfs pinctrl-devices file. Acked-by: Stephen Warren <swarren@nvidia.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2011-10-19 20:14:33 +04:00
const char *gname = pctlops->get_group_name(pctldev, selector);
seq_printf(s, "%u (%s):", selector, gname);
pinconf_dump_group(pctldev, s, selector, gname);
seq_printf(s, "\n");
pinctrl: add a pin config interface This add per-pin and per-group pin config interfaces for biasing, driving and other such electronic properties. The details of passed configurations are passed in an opaque unsigned long which may be dereferences to integer types, structs or lists on either side of the configuration interface. ChangeLog v1->v2: - Clear split of terminology: we now have pin controllers, and those may support two interfaces using vtables: pin multiplexing and pin configuration. - Break out pin configuration to its own C file, controllers may implement only config without mux, and vice versa, so keep each sub-functionality of pin controllers separate. Introduce CONFIG_PINCONF in Kconfig. - Implement some core logic around pin configuration in the pinconf.c file. - Remove UNKNOWN config states, these were just surplus baggage. - Remove FLOAT config state - HIGH_IMPEDANCE should be enough for everyone. - PIN_CONFIG_POWER_SOURCE added to handle switching the power supply for the pin logic between different sources - Explicit DISABLE config enums to turn schmitt-trigger, wakeup etc OFF. - Update documentation to reflect all the recent reasoning. ChangeLog v2->v3: - Twist API around to pass around arrays of config tuples instead of (param, value) pairs everywhere. - Explicit drive strength semantics for push/pull and similar drive modes, this shall be the number of drive stages vs nominal load impedance, which should match the actual electronics used in push/pull CMOS or TTY totempoles. - Drop load capacitance configuration - I probably don't know what I'm doing here so leave it out. - Drop PIN_CONFIG_INPUT_SCHMITT_OFF, instead the argument zero to PIN_CONFIG_INPUT_SCHMITT turns schmitt trigger off. - Drop PIN_CONFIG_NORMAL_POWER_MODE and have a well defined argument to PIN_CONFIG_LOW_POWER_MODE to get out of it instead. - Drop PIN_CONFIG_WAKEUP_ENABLE/DISABLE and just use PIN_CONFIG_WAKEUP with defined value zero to turn wakeup off. - Add PIN_CONFIG_INPUT_DEBOUNCE for configuring debounce time on input lines. - Fix a bug when we tried to configure pins for pin controllers without pinconf support. - Initialized debugfs properly so it works. - Initialize the mutex properly and lock around config tampering sections. - Check the return value from get_initial_config() properly. ChangeLog v3->v4: - Export the pin_config_get(), pin_config_set() and pin_config_group() functions. - Drop the entire concept of just getting initial config and keeping track of pin states internally, instead ask the pins what state they are in. Previous idea was plain wrong, if the device cannot keep track of its state, the driver should do it. - Drop the generic configuration layout, it seems this impose too much restriction on some pin controllers, so let them do things the way they want and split off support for generic config as an optional add-on. ChangeLog v4->v5: - Introduce two symmetric driver calls for group configuration, .pin_config_group_[get|set] and corresponding external calls. - Remove generic semantic meanings of return values from config calls, these belong in the generic config patch. Just pass the return value through instead. - Add a debugfs entry "pinconf-groups" to read status from group configuration only, also slam in a per-group debug callback in the pinconf_ops so custom drivers can display something meaningful for their pins. - Fix some dangling newline. - Drop dangling #else clause. - Update documentation to match the above. ChangeLog v5->v6: - Change to using a pin name as parameter for the [get|set]_config() functions, as suggested by Stephen Warren. This is more natural as names will be what a developer has access to in written documentation etc. ChangeLog v6->v7: - Refactor out by-pin and by-name get/set functions, only expose the by-name functions externally, expose the by-pin functions internally. - Show supported pin control functionality in the debugfs pinctrl-devices file. Acked-by: Stephen Warren <swarren@nvidia.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2011-10-19 20:14:33 +04:00
selector++;
}
return 0;
}
static int pinconf_pins_open(struct inode *inode, struct file *file)
{
return single_open(file, pinconf_pins_show, inode->i_private);
}
static int pinconf_groups_open(struct inode *inode, struct file *file)
{
return single_open(file, pinconf_groups_show, inode->i_private);
}
static const struct file_operations pinconf_pins_ops = {
.open = pinconf_pins_open,
.read = seq_read,
.llseek = seq_lseek,
.release = single_release,
};
static const struct file_operations pinconf_groups_ops = {
.open = pinconf_groups_open,
.read = seq_read,
.llseek = seq_lseek,
.release = single_release,
};
#define MAX_NAME_LEN 15
struct dbg_cfg {
enum pinctrl_map_type map_type;
char dev_name[MAX_NAME_LEN+1];
char state_name[MAX_NAME_LEN+1];
char pin_name[MAX_NAME_LEN+1];
};
/*
* Goal is to keep this structure as global in order to simply read the
* pinconf-config file after a write to check config is as expected
*/
static struct dbg_cfg pinconf_dbg_conf;
/**
* pinconf_dbg_config_print() - display the pinctrl config from the pinctrl
* map, of the dev/pin/state that was last written to pinconf-config file.
* @s: string filled in with config description
* @d: not used
*/
static int pinconf_dbg_config_print(struct seq_file *s, void *d)
{
struct pinctrl_maps *maps_node;
const struct pinctrl_map *map;
struct pinctrl_dev *pctldev = NULL;
const struct pinconf_ops *confops = NULL;
const struct pinctrl_map_configs *configs;
struct dbg_cfg *dbg = &pinconf_dbg_conf;
int i, j;
bool found = false;
unsigned long config;
pinctrl: move subsystem mutex to pinctrl_dev struct This mutex avoids deadlock in case of use of multiple pin controllers. Before this modification, by using a global mutex, deadlock appeared when, for example, a call to pinctrl_pins_show() locked the pinctrl_mutex, called the ops->pin_dbg_show of a particular pin controller. If this pin controller needs I2C access to retrieve configuration information and I2C driver is using pinctrl to drive its pins, a call to pinctrl_select_state() try to lock again pinctrl_mutex which leads to a deadlock. Notice that the mutex grab from the two direction functions was moved into pinctrl_gpio_direction(). For several cases, we can't replace pinctrl_mutex by pctldev->mutex, because at this stage, pctldev is not accessible : - pinctrl_get()/pinctrl_put() - pinctrl_register_maps() So add respectively pinctrl_list_mutex and pinctrl_maps_mutex in order to protect pinctrl_list and pinctrl_maps list instead. Reintroduce pinctrldev_list_mutex in find_pinctrl_by_of_node(), pinctrl_find_and_add_gpio_range() pinctrl_request_gpio(), pinctrl_free_gpio(), pinctrl_gpio_direction(), pinctrl_devices_show(), pinctrl_register() and pinctrl_unregister() to protect pinctrldev_list. Changes v2->v3: - Fix a missing EXPORT_SYMBOL_GPL() for pinctrl_select_state(). Changes v1->v2: - pinctrl_select_state_locked() is removed, all lock mechanism is located inside pinctrl_select_state(). When parsing the state->setting list, take the per-pin-controller driver lock. (Patrice). - Introduce pinctrldev_list_mutex to protect pinctrldev_list in all functions which parse or modify pictrldev_list. (Patrice). - move find_pinctrl_by_of_node() from pinctrl/devicetree.c to pinctrl/core.c in order to protect pinctrldev_list. (Patrice). - Sink mutex:es into some functions and remove some _locked variants down to where the lists are actually accessed to make things simpler. (Linus) - Drop *all* mutexes completely from pinctrl_lookup_state() and pinctrl_select_state() - no relevant mutex was taken and it was unclear what this was protecting against. (Linus) Reported by : Seraphin Bonnaffe <seraphin.bonnaffe@stericsson.com> Signed-off-by: Patrice Chotard <patrice.chotard@st.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2013-04-11 13:01:27 +04:00
mutex_lock(&pctldev->mutex);
/* Parse the pinctrl map and look for the elected pin/state */
for_each_maps(maps_node, i, map) {
if (map->type != dbg->map_type)
continue;
if (strcmp(map->dev_name, dbg->dev_name))
continue;
if (strcmp(map->name, dbg->state_name))
continue;
for (j = 0; j < map->data.configs.num_configs; j++) {
if (!strcmp(map->data.configs.group_or_pin,
dbg->pin_name)) {
/*
* We found the right pin / state, read the
* config and he pctldev for later use
*/
configs = &map->data.configs;
pctldev = get_pinctrl_dev_from_devname
(map->ctrl_dev_name);
found = true;
break;
}
}
}
if (!found) {
seq_printf(s, "No config found for dev/state/pin, expected:\n");
seq_printf(s, "Searched dev:%s\n", dbg->dev_name);
seq_printf(s, "Searched state:%s\n", dbg->state_name);
seq_printf(s, "Searched pin:%s\n", dbg->pin_name);
seq_printf(s, "Use: modify config_pin <devname> "\
"<state> <pinname> <value>\n");
goto exit;
}
config = *(configs->configs);
seq_printf(s, "Dev %s has config of %s in state %s: 0x%08lX\n",
dbg->dev_name, dbg->pin_name,
dbg->state_name, config);
if (pctldev)
confops = pctldev->desc->confops;
if (confops && confops->pin_config_config_dbg_show)
confops->pin_config_config_dbg_show(pctldev, s, config);
exit:
pinctrl: move subsystem mutex to pinctrl_dev struct This mutex avoids deadlock in case of use of multiple pin controllers. Before this modification, by using a global mutex, deadlock appeared when, for example, a call to pinctrl_pins_show() locked the pinctrl_mutex, called the ops->pin_dbg_show of a particular pin controller. If this pin controller needs I2C access to retrieve configuration information and I2C driver is using pinctrl to drive its pins, a call to pinctrl_select_state() try to lock again pinctrl_mutex which leads to a deadlock. Notice that the mutex grab from the two direction functions was moved into pinctrl_gpio_direction(). For several cases, we can't replace pinctrl_mutex by pctldev->mutex, because at this stage, pctldev is not accessible : - pinctrl_get()/pinctrl_put() - pinctrl_register_maps() So add respectively pinctrl_list_mutex and pinctrl_maps_mutex in order to protect pinctrl_list and pinctrl_maps list instead. Reintroduce pinctrldev_list_mutex in find_pinctrl_by_of_node(), pinctrl_find_and_add_gpio_range() pinctrl_request_gpio(), pinctrl_free_gpio(), pinctrl_gpio_direction(), pinctrl_devices_show(), pinctrl_register() and pinctrl_unregister() to protect pinctrldev_list. Changes v2->v3: - Fix a missing EXPORT_SYMBOL_GPL() for pinctrl_select_state(). Changes v1->v2: - pinctrl_select_state_locked() is removed, all lock mechanism is located inside pinctrl_select_state(). When parsing the state->setting list, take the per-pin-controller driver lock. (Patrice). - Introduce pinctrldev_list_mutex to protect pinctrldev_list in all functions which parse or modify pictrldev_list. (Patrice). - move find_pinctrl_by_of_node() from pinctrl/devicetree.c to pinctrl/core.c in order to protect pinctrldev_list. (Patrice). - Sink mutex:es into some functions and remove some _locked variants down to where the lists are actually accessed to make things simpler. (Linus) - Drop *all* mutexes completely from pinctrl_lookup_state() and pinctrl_select_state() - no relevant mutex was taken and it was unclear what this was protecting against. (Linus) Reported by : Seraphin Bonnaffe <seraphin.bonnaffe@stericsson.com> Signed-off-by: Patrice Chotard <patrice.chotard@st.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2013-04-11 13:01:27 +04:00
mutex_unlock(&pctldev->mutex);
return 0;
}
/**
* pinconf_dbg_config_write() - modify the pinctrl config in the pinctrl
* map, of a dev/pin/state entry based on user entries to pinconf-config
* @user_buf: contains the modification request with expected format:
* modify config_pin <devicename> <state> <pinname> <newvalue>
* modify is literal string, alternatives like add/delete not supported yet
* config_pin is literal, alternatives like config_mux not supported yet
* <devicename> <state> <pinname> are values that should match the pinctrl-maps
* <newvalue> reflects the new config and is driver dependant
*/
static int pinconf_dbg_config_write(struct file *file,
const char __user *user_buf, size_t count, loff_t *ppos)
{
struct pinctrl_maps *maps_node;
const struct pinctrl_map *map;
struct pinctrl_dev *pctldev = NULL;
const struct pinconf_ops *confops = NULL;
struct dbg_cfg *dbg = &pinconf_dbg_conf;
const struct pinctrl_map_configs *configs;
char config[MAX_NAME_LEN+1];
bool found = false;
char buf[128];
char *b = &buf[0];
int buf_size;
char *token;
int i;
/* Get userspace string and assure termination */
buf_size = min(count, (sizeof(buf)-1));
if (copy_from_user(buf, user_buf, buf_size))
return -EFAULT;
buf[buf_size] = 0;
/*
* need to parse entry and extract parameters:
* modify configs_pin devicename state pinname newvalue
*/
/* Get arg: 'modify' */
token = strsep(&b, " ");
if (!token)
return -EINVAL;
if (strcmp(token, "modify"))
return -EINVAL;
/* Get arg type: "config_pin" type supported so far */
token = strsep(&b, " ");
if (!token)
return -EINVAL;
if (strcmp(token, "config_pin"))
return -EINVAL;
dbg->map_type = PIN_MAP_TYPE_CONFIGS_PIN;
/* get arg 'device_name' */
token = strsep(&b, " ");
if (token == NULL)
return -EINVAL;
if (strlen(token) >= MAX_NAME_LEN)
return -EINVAL;
strncpy(dbg->dev_name, token, MAX_NAME_LEN);
/* get arg 'state_name' */
token = strsep(&b, " ");
if (token == NULL)
return -EINVAL;
if (strlen(token) >= MAX_NAME_LEN)
return -EINVAL;
strncpy(dbg->state_name, token, MAX_NAME_LEN);
/* get arg 'pin_name' */
token = strsep(&b, " ");
if (token == NULL)
return -EINVAL;
if (strlen(token) >= MAX_NAME_LEN)
return -EINVAL;
strncpy(dbg->pin_name, token, MAX_NAME_LEN);
/* get new_value of config' */
token = strsep(&b, " ");
if (token == NULL)
return -EINVAL;
if (strlen(token) >= MAX_NAME_LEN)
return -EINVAL;
strncpy(config, token, MAX_NAME_LEN);
pinctrl: move subsystem mutex to pinctrl_dev struct This mutex avoids deadlock in case of use of multiple pin controllers. Before this modification, by using a global mutex, deadlock appeared when, for example, a call to pinctrl_pins_show() locked the pinctrl_mutex, called the ops->pin_dbg_show of a particular pin controller. If this pin controller needs I2C access to retrieve configuration information and I2C driver is using pinctrl to drive its pins, a call to pinctrl_select_state() try to lock again pinctrl_mutex which leads to a deadlock. Notice that the mutex grab from the two direction functions was moved into pinctrl_gpio_direction(). For several cases, we can't replace pinctrl_mutex by pctldev->mutex, because at this stage, pctldev is not accessible : - pinctrl_get()/pinctrl_put() - pinctrl_register_maps() So add respectively pinctrl_list_mutex and pinctrl_maps_mutex in order to protect pinctrl_list and pinctrl_maps list instead. Reintroduce pinctrldev_list_mutex in find_pinctrl_by_of_node(), pinctrl_find_and_add_gpio_range() pinctrl_request_gpio(), pinctrl_free_gpio(), pinctrl_gpio_direction(), pinctrl_devices_show(), pinctrl_register() and pinctrl_unregister() to protect pinctrldev_list. Changes v2->v3: - Fix a missing EXPORT_SYMBOL_GPL() for pinctrl_select_state(). Changes v1->v2: - pinctrl_select_state_locked() is removed, all lock mechanism is located inside pinctrl_select_state(). When parsing the state->setting list, take the per-pin-controller driver lock. (Patrice). - Introduce pinctrldev_list_mutex to protect pinctrldev_list in all functions which parse or modify pictrldev_list. (Patrice). - move find_pinctrl_by_of_node() from pinctrl/devicetree.c to pinctrl/core.c in order to protect pinctrldev_list. (Patrice). - Sink mutex:es into some functions and remove some _locked variants down to where the lists are actually accessed to make things simpler. (Linus) - Drop *all* mutexes completely from pinctrl_lookup_state() and pinctrl_select_state() - no relevant mutex was taken and it was unclear what this was protecting against. (Linus) Reported by : Seraphin Bonnaffe <seraphin.bonnaffe@stericsson.com> Signed-off-by: Patrice Chotard <patrice.chotard@st.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2013-04-11 13:01:27 +04:00
mutex_lock(&pinctrl_maps_mutex);
/* Parse the pinctrl map and look for the selected dev/state/pin */
for_each_maps(maps_node, i, map) {
if (strcmp(map->dev_name, dbg->dev_name))
continue;
if (map->type != dbg->map_type)
continue;
if (strcmp(map->name, dbg->state_name))
continue;
/* we found the right pin / state, so overwrite config */
if (!strcmp(map->data.configs.group_or_pin, dbg->pin_name)) {
found = true;
pctldev = get_pinctrl_dev_from_devname(
map->ctrl_dev_name);
configs = &map->data.configs;
break;
}
}
if (!found) {
count = -EINVAL;
goto exit;
}
if (pctldev)
confops = pctldev->desc->confops;
if (confops && confops->pin_config_dbg_parse_modify) {
for (i = 0; i < configs->num_configs; i++) {
confops->pin_config_dbg_parse_modify(pctldev,
config,
&configs->configs[i]);
}
}
exit:
pinctrl: move subsystem mutex to pinctrl_dev struct This mutex avoids deadlock in case of use of multiple pin controllers. Before this modification, by using a global mutex, deadlock appeared when, for example, a call to pinctrl_pins_show() locked the pinctrl_mutex, called the ops->pin_dbg_show of a particular pin controller. If this pin controller needs I2C access to retrieve configuration information and I2C driver is using pinctrl to drive its pins, a call to pinctrl_select_state() try to lock again pinctrl_mutex which leads to a deadlock. Notice that the mutex grab from the two direction functions was moved into pinctrl_gpio_direction(). For several cases, we can't replace pinctrl_mutex by pctldev->mutex, because at this stage, pctldev is not accessible : - pinctrl_get()/pinctrl_put() - pinctrl_register_maps() So add respectively pinctrl_list_mutex and pinctrl_maps_mutex in order to protect pinctrl_list and pinctrl_maps list instead. Reintroduce pinctrldev_list_mutex in find_pinctrl_by_of_node(), pinctrl_find_and_add_gpio_range() pinctrl_request_gpio(), pinctrl_free_gpio(), pinctrl_gpio_direction(), pinctrl_devices_show(), pinctrl_register() and pinctrl_unregister() to protect pinctrldev_list. Changes v2->v3: - Fix a missing EXPORT_SYMBOL_GPL() for pinctrl_select_state(). Changes v1->v2: - pinctrl_select_state_locked() is removed, all lock mechanism is located inside pinctrl_select_state(). When parsing the state->setting list, take the per-pin-controller driver lock. (Patrice). - Introduce pinctrldev_list_mutex to protect pinctrldev_list in all functions which parse or modify pictrldev_list. (Patrice). - move find_pinctrl_by_of_node() from pinctrl/devicetree.c to pinctrl/core.c in order to protect pinctrldev_list. (Patrice). - Sink mutex:es into some functions and remove some _locked variants down to where the lists are actually accessed to make things simpler. (Linus) - Drop *all* mutexes completely from pinctrl_lookup_state() and pinctrl_select_state() - no relevant mutex was taken and it was unclear what this was protecting against. (Linus) Reported by : Seraphin Bonnaffe <seraphin.bonnaffe@stericsson.com> Signed-off-by: Patrice Chotard <patrice.chotard@st.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2013-04-11 13:01:27 +04:00
mutex_unlock(&pinctrl_maps_mutex);
return count;
}
static int pinconf_dbg_config_open(struct inode *inode, struct file *file)
{
return single_open(file, pinconf_dbg_config_print, inode->i_private);
}
static const struct file_operations pinconf_dbg_pinconfig_fops = {
.open = pinconf_dbg_config_open,
.write = pinconf_dbg_config_write,
.read = seq_read,
.llseek = seq_lseek,
.release = single_release,
.owner = THIS_MODULE,
};
pinctrl: add a pin config interface This add per-pin and per-group pin config interfaces for biasing, driving and other such electronic properties. The details of passed configurations are passed in an opaque unsigned long which may be dereferences to integer types, structs or lists on either side of the configuration interface. ChangeLog v1->v2: - Clear split of terminology: we now have pin controllers, and those may support two interfaces using vtables: pin multiplexing and pin configuration. - Break out pin configuration to its own C file, controllers may implement only config without mux, and vice versa, so keep each sub-functionality of pin controllers separate. Introduce CONFIG_PINCONF in Kconfig. - Implement some core logic around pin configuration in the pinconf.c file. - Remove UNKNOWN config states, these were just surplus baggage. - Remove FLOAT config state - HIGH_IMPEDANCE should be enough for everyone. - PIN_CONFIG_POWER_SOURCE added to handle switching the power supply for the pin logic between different sources - Explicit DISABLE config enums to turn schmitt-trigger, wakeup etc OFF. - Update documentation to reflect all the recent reasoning. ChangeLog v2->v3: - Twist API around to pass around arrays of config tuples instead of (param, value) pairs everywhere. - Explicit drive strength semantics for push/pull and similar drive modes, this shall be the number of drive stages vs nominal load impedance, which should match the actual electronics used in push/pull CMOS or TTY totempoles. - Drop load capacitance configuration - I probably don't know what I'm doing here so leave it out. - Drop PIN_CONFIG_INPUT_SCHMITT_OFF, instead the argument zero to PIN_CONFIG_INPUT_SCHMITT turns schmitt trigger off. - Drop PIN_CONFIG_NORMAL_POWER_MODE and have a well defined argument to PIN_CONFIG_LOW_POWER_MODE to get out of it instead. - Drop PIN_CONFIG_WAKEUP_ENABLE/DISABLE and just use PIN_CONFIG_WAKEUP with defined value zero to turn wakeup off. - Add PIN_CONFIG_INPUT_DEBOUNCE for configuring debounce time on input lines. - Fix a bug when we tried to configure pins for pin controllers without pinconf support. - Initialized debugfs properly so it works. - Initialize the mutex properly and lock around config tampering sections. - Check the return value from get_initial_config() properly. ChangeLog v3->v4: - Export the pin_config_get(), pin_config_set() and pin_config_group() functions. - Drop the entire concept of just getting initial config and keeping track of pin states internally, instead ask the pins what state they are in. Previous idea was plain wrong, if the device cannot keep track of its state, the driver should do it. - Drop the generic configuration layout, it seems this impose too much restriction on some pin controllers, so let them do things the way they want and split off support for generic config as an optional add-on. ChangeLog v4->v5: - Introduce two symmetric driver calls for group configuration, .pin_config_group_[get|set] and corresponding external calls. - Remove generic semantic meanings of return values from config calls, these belong in the generic config patch. Just pass the return value through instead. - Add a debugfs entry "pinconf-groups" to read status from group configuration only, also slam in a per-group debug callback in the pinconf_ops so custom drivers can display something meaningful for their pins. - Fix some dangling newline. - Drop dangling #else clause. - Update documentation to match the above. ChangeLog v5->v6: - Change to using a pin name as parameter for the [get|set]_config() functions, as suggested by Stephen Warren. This is more natural as names will be what a developer has access to in written documentation etc. ChangeLog v6->v7: - Refactor out by-pin and by-name get/set functions, only expose the by-name functions externally, expose the by-pin functions internally. - Show supported pin control functionality in the debugfs pinctrl-devices file. Acked-by: Stephen Warren <swarren@nvidia.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2011-10-19 20:14:33 +04:00
void pinconf_init_device_debugfs(struct dentry *devroot,
struct pinctrl_dev *pctldev)
{
debugfs_create_file("pinconf-pins", S_IFREG | S_IRUGO,
devroot, pctldev, &pinconf_pins_ops);
debugfs_create_file("pinconf-groups", S_IFREG | S_IRUGO,
devroot, pctldev, &pinconf_groups_ops);
debugfs_create_file("pinconf-config", (S_IRUGO | S_IWUSR | S_IWGRP),
devroot, pctldev, &pinconf_dbg_pinconfig_fops);
pinctrl: add a pin config interface This add per-pin and per-group pin config interfaces for biasing, driving and other such electronic properties. The details of passed configurations are passed in an opaque unsigned long which may be dereferences to integer types, structs or lists on either side of the configuration interface. ChangeLog v1->v2: - Clear split of terminology: we now have pin controllers, and those may support two interfaces using vtables: pin multiplexing and pin configuration. - Break out pin configuration to its own C file, controllers may implement only config without mux, and vice versa, so keep each sub-functionality of pin controllers separate. Introduce CONFIG_PINCONF in Kconfig. - Implement some core logic around pin configuration in the pinconf.c file. - Remove UNKNOWN config states, these were just surplus baggage. - Remove FLOAT config state - HIGH_IMPEDANCE should be enough for everyone. - PIN_CONFIG_POWER_SOURCE added to handle switching the power supply for the pin logic between different sources - Explicit DISABLE config enums to turn schmitt-trigger, wakeup etc OFF. - Update documentation to reflect all the recent reasoning. ChangeLog v2->v3: - Twist API around to pass around arrays of config tuples instead of (param, value) pairs everywhere. - Explicit drive strength semantics for push/pull and similar drive modes, this shall be the number of drive stages vs nominal load impedance, which should match the actual electronics used in push/pull CMOS or TTY totempoles. - Drop load capacitance configuration - I probably don't know what I'm doing here so leave it out. - Drop PIN_CONFIG_INPUT_SCHMITT_OFF, instead the argument zero to PIN_CONFIG_INPUT_SCHMITT turns schmitt trigger off. - Drop PIN_CONFIG_NORMAL_POWER_MODE and have a well defined argument to PIN_CONFIG_LOW_POWER_MODE to get out of it instead. - Drop PIN_CONFIG_WAKEUP_ENABLE/DISABLE and just use PIN_CONFIG_WAKEUP with defined value zero to turn wakeup off. - Add PIN_CONFIG_INPUT_DEBOUNCE for configuring debounce time on input lines. - Fix a bug when we tried to configure pins for pin controllers without pinconf support. - Initialized debugfs properly so it works. - Initialize the mutex properly and lock around config tampering sections. - Check the return value from get_initial_config() properly. ChangeLog v3->v4: - Export the pin_config_get(), pin_config_set() and pin_config_group() functions. - Drop the entire concept of just getting initial config and keeping track of pin states internally, instead ask the pins what state they are in. Previous idea was plain wrong, if the device cannot keep track of its state, the driver should do it. - Drop the generic configuration layout, it seems this impose too much restriction on some pin controllers, so let them do things the way they want and split off support for generic config as an optional add-on. ChangeLog v4->v5: - Introduce two symmetric driver calls for group configuration, .pin_config_group_[get|set] and corresponding external calls. - Remove generic semantic meanings of return values from config calls, these belong in the generic config patch. Just pass the return value through instead. - Add a debugfs entry "pinconf-groups" to read status from group configuration only, also slam in a per-group debug callback in the pinconf_ops so custom drivers can display something meaningful for their pins. - Fix some dangling newline. - Drop dangling #else clause. - Update documentation to match the above. ChangeLog v5->v6: - Change to using a pin name as parameter for the [get|set]_config() functions, as suggested by Stephen Warren. This is more natural as names will be what a developer has access to in written documentation etc. ChangeLog v6->v7: - Refactor out by-pin and by-name get/set functions, only expose the by-name functions externally, expose the by-pin functions internally. - Show supported pin control functionality in the debugfs pinctrl-devices file. Acked-by: Stephen Warren <swarren@nvidia.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2011-10-19 20:14:33 +04:00
}
#endif