Skip to contentSkip to conversation

power_supply: Introduce power supply charging driver

PATCH Superseded Jenny TC 4 participants 14 messages Mar 7, 2014 – Mar 13, 2014

Subsystems (2)

v8 of 12 Superseded Latest is v12 →

lore
Share & tools
lore

Jenny TC

author
#0 lore raw

v8 cover power_supply: Introduce power supply charging driver

v1: introduced feature as a framework within power supply class driver with
separate files for battid framework and charging framework
v2: fixed review comments, moved macros and inline functions to power_supply.h
v3: moved the feature as a separate driver, combined battid framework and
charging framework inside the power supply charging driver. Moved charger specific properties to power_supply_charger.h and plugged the driver with power supply subsystem using power_supply_notifier introduced in my previous patch. Also a sample charger chip driver (bq24261) patch added to give more idea on the psy charging driver usage
v4: Fixed review comments, no major design changes.
v5: Fixed makefile inconsistencies, removed unused pdata callbacks
v6: Fixed nested loops, commenting style
v7: added kerneldocs for structs and minor fixes
v8: used msecs_to_jiffies instead of HZ directly, modified Kconfig help text
for POWER_SUPPLY_CHARGING_ALGO_PSE

The Power Supply charging driver connects multiple subsystems
to do charging in a generic way. The subsystems involves power_supply, thermal and battery communication subsystems (1wire).With this the charging is handled in a generic way.

The driver makes use of different new features - Battery Identification interfaces, pluggable charging algorithms, charger cable arbitrations etc. The patch also introduces generic interface for charger cable notifications. Charger cable events and capabilities can be notified using the generic power_supply_notifier chain.

Overall this driver removes the charging logic out of the charger chip driver and the charger chip driver can just listen to the request from the power supply charging driver to set the charger properties. This can be implemented by exposing get_property and set property callbacks.

Jenny TC (4):
power_supply: Add inlmt,iterm, min/max temp props
power_supply: Introduce generic psy charging driver
power_supply: Introduce PSE compliant algorithm
power_supply: bq24261 charger driver

 Documentation/power/power_supply_charger.txt |  353 +++++++
 Documentation/power/power_supply_class.txt   |    6 +
 drivers/power/Kconfig                        |   31 +
 drivers/power/Makefile                       |    3 +
 drivers/power/bq24261-charger.c              | 1350 ++++++++++++++++++++++++++
 drivers/power/charging_algo_pse.c            |  204 ++++
 drivers/power/power_supply_charger.c         | 1186 ++++++++++++++++++++++
 drivers/power/power_supply_charger.h         |  218 +++++
 drivers/power/power_supply_core.c            |    3 +
 drivers/power/power_supply_sysfs.c           |    4 +
 include/linux/power/bq24261-charger.h        |   25 +
 include/linux/power/power_supply_charger.h   |  341 +++++++
 include/linux/power_supply.h                 |  164 ++++
 13 files changed, 3888 insertions(+)
 create mode 100644 Documentation/power/power_supply_charger.txt
 create mode 100644 drivers/power/bq24261-charger.c
 create mode 100644 drivers/power/charging_algo_pse.c
 create mode 100644 drivers/power/power_supply_charger.c
 create mode 100644 drivers/power/power_supply_charger.h
 create mode 100644 include/linux/power/bq24261-charger.h
 create mode 100644 include/linux/power/power_supply_charger.h
Signature · 1 line
1.7.9.5
author
#1 lore raw

PATCH v8 1/4 power_supply: Add inlmt,iterm, min/max temp props

Add new power supply properties for input current, charge termination current, min and max temperature

POWER_SUPPLY_PROP_TEMP_MIN - minimum operatable temperature POWER_SUPPLY_PROP_TEMP_MAX - maximum operatable temperature

POWER_SUPPLY_PROP_INLMT - input current limit programmed by charger. Indicates the input current for a charging source.

POWER_SUPPLY_PROP_CHARGE_TERM_CUR - Charge termination current used to detect the end of charge condition

---
 Documentation/power/power_supply_class.txt |    6 ++++++
 drivers/power/power_supply_sysfs.c         |    4 ++++
 include/linux/power_supply.h               |    4 ++++
 3 files changed, 14 insertions(+)
Documentation/power/power_supply_class.txt +6 -0
index 89a8816..48cff88 100644
@@ -118,6 +118,10 @@ relative, time-based measurements.
 CONSTANT_CHARGE_CURRENT - constant charge current programmed by charger.
 CONSTANT_CHARGE_CURRENT_MAX - maximum charge current supported by the
 power supply object.
+INPUT_CURRENT_LIMIT - input current limit programmed by charger. Indicates
+the current drawn from a charging source.
+CHARGE_TERM_CURRENT - Charge termination current used to detect the end of charge
+condition.
 
 CONSTANT_CHARGE_VOLTAGE - constant charge voltage programmed by charger.
 CONSTANT_CHARGE_VOLTAGE_MAX - maximum charge voltage supported by the
@@ -140,6 +144,8 @@ TEMP_ALERT_MAX - maximum battery temperature alert.
 TEMP_AMBIENT - ambient temperature.
 TEMP_AMBIENT_ALERT_MIN - minimum ambient temperature alert.
 TEMP_AMBIENT_ALERT_MAX - maximum ambient temperature alert.
+TEMP_MIN - minimum operatable temperature
+TEMP_MAX - maximum operatable temperature
 
 TIME_TO_EMPTY - seconds left for battery to be considered empty (i.e.
 while battery powers a load)
drivers/power/power_supply_sysfs.c +4 -0
index 44420d1..750a202 100644
@@ -167,6 +167,7 @@ static struct device_attribute power_supply_attrs[] = {
 	POWER_SUPPLY_ATTR(constant_charge_voltage_max),
 	POWER_SUPPLY_ATTR(charge_control_limit),
 	POWER_SUPPLY_ATTR(charge_control_limit_max),
+	POWER_SUPPLY_ATTR(input_current_limit),
 	POWER_SUPPLY_ATTR(energy_full_design),
 	POWER_SUPPLY_ATTR(energy_empty_design),
 	POWER_SUPPLY_ATTR(energy_full),
@@ -178,6 +179,8 @@ static struct device_attribute power_supply_attrs[] = {
 	POWER_SUPPLY_ATTR(capacity_alert_max),
 	POWER_SUPPLY_ATTR(capacity_level),
 	POWER_SUPPLY_ATTR(temp),
+	POWER_SUPPLY_ATTR(temp_max),
+	POWER_SUPPLY_ATTR(temp_min),
 	POWER_SUPPLY_ATTR(temp_alert_min),
 	POWER_SUPPLY_ATTR(temp_alert_max),
 	POWER_SUPPLY_ATTR(temp_ambient),
@@ -189,6 +192,7 @@ static struct device_attribute power_supply_attrs[] = {
 	POWER_SUPPLY_ATTR(time_to_full_avg),
 	POWER_SUPPLY_ATTR(type),
 	POWER_SUPPLY_ATTR(scope),
+	POWER_SUPPLY_ATTR(charge_term_current),
 	/* Properties of type `const char *' */
 	POWER_SUPPLY_ATTR(model_name),
 	POWER_SUPPLY_ATTR(manufacturer),
include/linux/power_supply.h +4 -0
index c9dc4e0..0278600 100644
@@ -120,6 +120,7 @@ enum power_supply_property {
 	POWER_SUPPLY_PROP_CONSTANT_CHARGE_VOLTAGE_MAX,
 	POWER_SUPPLY_PROP_CHARGE_CONTROL_LIMIT,
 	POWER_SUPPLY_PROP_CHARGE_CONTROL_LIMIT_MAX,
+	POWER_SUPPLY_PROP_INPUT_CURRENT_LIMIT,
 	POWER_SUPPLY_PROP_ENERGY_FULL_DESIGN,
 	POWER_SUPPLY_PROP_ENERGY_EMPTY_DESIGN,
 	POWER_SUPPLY_PROP_ENERGY_FULL,
@@ -131,6 +132,8 @@ enum power_supply_property {
 	POWER_SUPPLY_PROP_CAPACITY_ALERT_MAX, /* in percents! */
 	POWER_SUPPLY_PROP_CAPACITY_LEVEL,
 	POWER_SUPPLY_PROP_TEMP,
+	POWER_SUPPLY_PROP_TEMP_MAX,
+	POWER_SUPPLY_PROP_TEMP_MIN,
 	POWER_SUPPLY_PROP_TEMP_ALERT_MIN,
 	POWER_SUPPLY_PROP_TEMP_ALERT_MAX,
 	POWER_SUPPLY_PROP_TEMP_AMBIENT,
@@ -142,6 +145,7 @@ enum power_supply_property {
 	POWER_SUPPLY_PROP_TIME_TO_FULL_AVG,
 	POWER_SUPPLY_PROP_TYPE, /* use power_supply.type instead */
 	POWER_SUPPLY_PROP_SCOPE,
+	POWER_SUPPLY_PROP_CHARGE_TERM_CURRENT,
 	/* Properties of type `const char *' */
 	POWER_SUPPLY_PROP_MODEL_NAME,
 	POWER_SUPPLY_PROP_MANUFACTURER,
Signature · 1 line
1.7.9.5
Reviewed-by on v8 1/4
#2 lore raw

Re: [PATCHv8 1/4] power_supply: Add inlmt,iterm, min/max temp props

Quoted 6 lines · Jenny TC

Add new power supply properties for input current, charge termination current, min and max temperature

POWER_SUPPLY_PROP_TEMP_MIN - minimum operatable temperature POWER_SUPPLY_PROP_TEMP_MAX - maximum operatable temperature

Quoted 7 lines

POWER_SUPPLY_PROP_INLMT - input current limit programmed by charger. Indicates the input current for a charging source.

POWER_SUPPLY_PROP_CHARGE_TERM_CUR - Charge termination current used to detect the end of charge condition

The patch no longer matches the description, otherwise it looks good.

Signature · 2 lines
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
author
#3 lore raw

Re: [PATCHv8 1/4] power_supply: Add inlmt,iterm, min/max temp props

Quoted 10 lines · Pavel Machek
Quoted 9 lines · Jenny TC

Add new power supply properties for input current, charge termination current, min and max temperature

POWER_SUPPLY_PROP_TEMP_MIN - minimum operatable temperature POWER_SUPPLY_PROP_TEMP_MAX - maximum operatable temperature

POWER_SUPPLY_PROP_INLMT - input current limit programmed by charger. Indicates the input current for a charging source.

Quoted 6 lines
Quoted 4 lines

POWER_SUPPLY_PROP_CHARGE_TERM_CUR - Charge termination current used to detect the end of charge condition

The patch no longer matches the description, otherwise it looks good.

Will fix it in next patch set

author
#4 lore raw

PATCH v8 2/4 power_supply: Introduce generic psy charging driver

In reply to Jenny TC

The Power Supply charging driver connects multiple subsystems
to do charging in a generic way. The subsystems involves power_supply, thermal and battery communication subsystems (1wire).With this the charging is handled in a generic way.

The driver makes use of different new features - Battery Identification interfaces, pluggable charging algorithms, charger cable arbitrations etc. The patch also introduces generic interface for charger cable notifications. Charger cable events and capabilities can be notified using the generic power_supply_notifier chain.

Overall this driver removes the charging logic out of the charger chip driver and the charger chip driver can just listen to the request from the power supply charging driver to set the charger properties. This can be implemented by exposing get_property and set property callbacks.

---
 Documentation/power/power_supply_charger.txt |  353 ++++++++
 drivers/power/Kconfig                        |    8 +
 drivers/power/Makefile                       |    1 +
 drivers/power/power_supply_charger.c         | 1186 ++++++++++++++++++++++++++
 drivers/power/power_supply_charger.h         |  218 +++++
 drivers/power/power_supply_core.c            |    3 +
 include/linux/power/power_supply_charger.h   |  280 ++++++
 include/linux/power_supply.h                 |  160 ++++
 8 files changed, 2209 insertions(+)
 create mode 100644 Documentation/power/power_supply_charger.txt
 create mode 100644 drivers/power/power_supply_charger.c
 create mode 100644 drivers/power/power_supply_charger.h
 create mode 100644 include/linux/power/power_supply_charger.h
Documentation/power/power_supply_charger.txt new file +353 -0
new file mode 100644
index 0000000..4a2e0e3
@@ -0,0 +1,353 @@
+1. Introduction
+===============
+
+The Power Supply charging driver connects multiple subsystems
+to do charging in a generic way. The subsystems involves power_supply,
+thermal and battery communication subsystems (1wire).With this the charging is
+handled in a generic way by plugging the driver to power supply subsystem.
+
+The driver introduces different new features - Battery Identification
+interfaces, pluggable charging algorithms, charger cable arbitrations etc.
+
+In existing driver implementations the charging is done based on the static
+battery characteristics. This is done at the boot time by passing the battery
+properties (max_voltage, capacity) etc. as a platform data to the
+charger/battery driver. But new generation high volt batteries needs to be
+identified dynamically to do charging in a safe manner. The batteries are
+coming with different communication protocols. It become necessary to
+communicate with battery and identify it's charging profiles before setup
+charging.
+
+Also the charging algorithms can vary based on the battery characteristics
+and the platform characteristics. To handle charging in a generic way it's
+necessary to support pluggable charging algorithms. Power Supply Charging
+driver selects algorithms based on the type of battery charging profile.
+This is a simple binding and can be improved later. This may be improved to
+select the algorithms based on the platform requirements. Also we can extend
+this driver to plug algorithms from the user space.
+
+The driver also introduces the charger cable arbitration. A charger may
+supports multiple cables, but it may not be able to charge with multiple
+cables at a time (USB/AC/Wireless etc.). The arbitration logic inside the
+driver selects the cable based on it's capabilities and the maximum
+charge current the platform can support.
+
+Also the driver exposes features to control charging on different platform
+states. One such feature is thermal. The driver handles the thermal
+throttling requests for charging and control charging based on the thermal
+subsystem requirements.
+
+Overall this driver removes the charging logic out of the charger chip driver
+and the charger chip driver can just listen to the request from the power
+supply charging driver to set the charger properties. This can be implemented
+by exposing get_property and set property callbacks.
+
+2. Reading Battery charging profile
+===================================
+
+Power Supply charging driver expose APIs to retrieve battery profile of a
+battery. The battery profile can be read by battery identification driver which
+may be 1wire/I2C/SFI driver. Battery identification driver can register the
+battery profile with the power supply charging driver using the API
+psy_battery_prop_changed(). The driver also exposes API
+psy_get_battery_prop() to retrieve the battery profile which can be
+used by power supply drivers to setup the charging. Also drivers
+can register for battery removal/insertion notifications using
+power_supply_reg_notifier()
+
+3. Use Charging Driver to setup charging
+===========================================
+
+* Register the driver with the power_supply class
+* Register the driver with the power supply charging driver
+* Expose set_property and get_property functions so that charging
+  framework can control the charging
+
+3.1  Registering charger chip driver with power supply charging driver
+================================================================
+
+struct power_supply_class psy_usb;
+struct power_supply_charger psyc_usb;
+
+/* register with power supply subsystem */
+psy_usb.name = DEV_NAME;
+psy_usb.type = POWER_SUPPLY_TYPE_USB;
+
+/* pointer to power supply property structure */
+psy_usb.properties =  &psy_usb_props;
+psy_usb.num_properties = ARRAY_SIZE(psy_usb_props);
+
+/* pointer to power supply get property function */
+psy_usb.get_property = psy_usb_get_property;
+
+/* pointer to power supply set property function */
+psy_usb.set_property = psy_usb_set_property;
+
+/* pointer to the supplied_to argument to indicate the batteries
+   to which this charger is supplying power  */
+psy_usb.supplied_to = supplied_to;
+psy_usb.num_supplicants = num_supplicants;
+
+/* register with power_supply subsystem */
+power_supply_register(device, &psy_usb);
+
+/* Register with power supply charging driver */
+
+/* Assign psy pointer to power supply charger */
+psyc_usb.psy = &chip->psy_usb;
+
+/* define supported cable types for this driver */
+psyc_usb.supported_cables = POWER_SUPPLY_CHARGER_TYPE_USB;
+
+/* pointer to power supply charger get property function */
+psyc_usb.get_property = bq24261_usb_psyc_get_property;
+
+/* pointer to power supply charger set property function */
+psyc_usb.set_property = bq24261_usb_psyc_set_property;
+
+/* pointer to throttle states */
+psyc_usb.throttle_states = chip->pdata->throttle_states;
+
+/* Number of throttle states */
+psyc_usb.num_throttle_states = chip->pdata->num_throttle_states;
+
+/*register with power supply charging */
+power_supply_register_charger(&chip->psyc_usb);
+
+3.2 Properties exposed to power supply class driver
+==================================================
+* POWER_SUPPLY_PROP_ONLINE
+	* Read access using get_property_function
+		* Returns the online property set using the set_property
+		  function
+	* Write access using set_property function
+		* Expose the value through get_property_function.
+* POWER_SUPPLY_PROP_PRESENT
+	* Read access using get_property_function
+		* Returns the present property set using the set_property
+		  function
+	* Write access using set_property function
+		* Expose the value through get_property_function.
+* POWER_SUPPLY_PROP_HEALTH
+	* Read access using get_property_function
+		* Returns charger health
+* POWER_SUPPLY_PROP_TYPE
+	* Read access using power_supply structure
+		* Returns charger type
+* POWER_SUPPLY_PROP_CHARGE_CONTROL_LIMIT
+	* Read access using get property
+		* Returns the present charge control limit
+	* Write access using set property
+		* Set charge control limit.
+		* Action: Driver is not expected to take any actions on this.
+			  Instead need to expose this on the read interface.
+			  Charging framework process this and notifies the
+			  charging parameters accordingly.
+* POWER_SUPPLY_PROP_CHARGE_CONTROL_LIMIT_MAX
+	* Read access using get property
+		* Returns the maximum charge control limit
+* POWER_SUPPLY_PROP_INPUT_CURRENT_LIMIT
+	* Read access using get property
+		* Returns the input current limit passed by set_property
+		  function
+		* Default value : 0
+	* Write access using set property
+		* Set input current limit
+		* Action : Write h/w register to set the input current limit
+			   based on the value
+* POWER_SUPPLY_PROP_CHARGE_CURRENT
+	* Read access using get property
+		* Returns charge current passed by set_property function
+		* Default value : 0
+	* Write access using set property
+		* Set charge current
+		* Action : Write h/w register to set the charge current based
+			   on the value
+* POWER_SUPPLY_PROP_CHARGE_VOLTAGE
+	* Read access using get property
+		* Returns charge voltage passed by set_property function
+		* Default value: 0
+	* Write access using set property
+		* Set charge voltage
+		* Action : Write h/w register to set the charge voltage based
+			   on the value
+* POWER_SUPPLY_PROP_CHARGE_TERM_CURRENT
+	* Read access using get property
+		* Returns charge termination current passed by set_property
+		  function
+		* Default value : 0
+	* Write access using set property
+		* Set charge termination current
+		* Action : Write h/w register to set the charge termination
+			   current
+			   based on the value
+* POWER_SUPPLY_PROP_TEMP_MIN
+	* Read access using get property
+		* Returns minimum charging  temperature passed by set_property
+		* Default value : Platform dependant
+	* Write access using set_property
+		* Set Minimum charging temperature
+		* Action : Write h/w register to set the minimum charging
+			   temperature based on the value
+* POWER_SUPPLY_PROP_TEMP_MAX
+	* Write access using set_property
+		* Set Maximum charging temperature
+		* Action : Write h/w register to set the maximum charging
+			   temperature based on the value
+	* Read access using get property
+		* Returns maximum charging  temperature passed by set_property
+		* Default value : Platform dependent
+* POWER_SUPPLY_PROP_MAX_CHARGE_CURRENT
+	* Read access using get property
+		* Returns maximum charging  current passed by set_property
+		* Default value : Platform dependent
+	* Write access using set_property
+		* Set Maximum charging current
+		* Action: Configure safety charging registers if any. If no h/w
+			  actions expected, report the value on the
+			  get_property interface.
+* POWER_SUPPLY_PROP_MAX_CHARGE_VOLTAGE
+	* Read access using get property
+		* Returns maximum charging  current passed by set_property
+		* Default value : Platform dependent
+	* Write access using set_property
+		* Set Maximum charging voltage
+		* Action: Configure safety charging registers if any. If not,
+			  no actions expected for this.
+
+3.3 Properties exposed to power supply charging driver
+=====================================================
+
+* POWER_SUPPLY_CHARGER_PROP_ENABLE_CHARGING
+	* Write access using set_property
+		* Enable/Disable charging. Charger supplies power to platform,
+		  but charging is disabled
+		* Action: Configure charger chip registers to enable/disable
+		  charging. Writing 0, disables charging, writing 1 enables
+		  charging
+* POWER_SUPPLY_CHARGER_PROP_ENABLE_CHARGER
+	* Write access using set_property
+		* Enable/Disable charger. Charger doesn't supply power to
+		  platform or battery and battery start to discharge.
+		* Action: Configure charger chip registers to enable/disable
+		  charger. Writing 0 disables charger, writing 1 enables
+		  charger
+* POWER_SUPPLY_CHARGER_PROP_CABLE_TYPE
+	* Write access using set_property
+		* Set charger cable type.
+		* Action: Report power supply type based on cable type
+* POWER_SUPPLY_CHARGER_PROP_RESET_WDT
+	* Write access using set_property
+		* Reset charger Watch Dog Timer
+		* Action: Reset charger Watch Dog Timer.
+* POWER_SUPPLY_CHARGER_PROP_PRIORITY
+	* Read access using get property
+		* Expose charger driver/chips priority if platform has multiple
+		  charger chip/drivers.
+
+3.4 Throttling data configuration
+=============================
+Power supply charging driver can take actions for Thermal throttling requests.
+Power supply core sends event PSY_EVENT_THROTTLE  on power_supply_notifier
+chain. Power supply charging driver handle this event and takes throttling
+actions as in the throttling configuration structure. The throttling
+configuration structure has the following format:
+
+struct psy_throttle_state {
+	enum psy_throttle_action throttle_action;
+	unsigned throttle_val;
+};
+
+enum psy_throttle_action {
+	/* Disable charger, charger doesn't supply VSYS */
+	PSY_THROTTLE_DISABLE_CHARGER = 0,
+
+	/* Disable charging, charger would supply VSYS */
+	PSY_THROTTLE_DISABLE_CHARGING,
+
+	/* Throttle charge current */
+	PSY_THROTTLE_CC_LIMIT,
+
+	/* Throttle input current limit */
+	PSY_THROTTLE_INPUT_LIMIT,
+};
+
+Throttling configuration example:
+
+struct psy_throttle_state my_throttle_states[] = {
+
+	/* Level 0:  Limit charge current to 1500mA. Normal Level */
+	{
+		.throttle_action = PSY_THROTTLE_CC_LIMIT,
+		.throttle_val = 1500,
+	},
+
+	/* Level 1: Limit charge current to 500mA */
+	{
+		.throttle_action = PSY_THROTTLE_CC_LIMIT,
+		.throttle_val = 500,
+	},
+
+	/*
+	* Level 2: Disable charging: Stop charging, charger supply power to
+	* platform.
+	*/
+	{
+		.throttle_action = PSY_THROTTLE_DISABLE_CHARGING,
+	},
+
+	/* Level 3: Disable charger: Battery start discharging */
+	{
+		.throttle_action = PSY_THROTTLE_DISABLE_CHARGER,
+	},
+
+};
+
+4. Charger Cable notifications
+==============================
+
+Charger cables can have different charging capabilities (current) depending on
+platform and cable type. The cable provider drivers detect and identify the
+cable types and notifies the power supply subsystem by posting an event
+PSY_EVENT_CABLE on the power_supply_notifier chain. The event notification
+should have a property structure pointer which indicates the cable type, event
+and capability.
+
+struct psy_cable_props cable;
+cable.chrg_evt = PSY_CHARGER_CABLE_EVENT_CONNECT;
+cable.chrg_type = PSY_CHARGER_CABLE_TYPE_USB_DCP;
+cable.ma = 1500;
+
+atomic_notifier_call_chain(&power_supply_notifier,
+				PSY_EVENT_CABLE, &cable);
+
+Power supply charging driver process this event and takes actions to setup
+charging.
+
+5. Registering new charging algorithm
+===================================
+Power supply charging driver supports pluggable charging algorithms. Charging
+algorithms processes charging profile and/or applies different logic (pulse
+charging/fast charging/relaxed charging) to decide the charging parameters (CC
+and CV).
+
+* Populate algorithm structure
+	struct psy_charging_algo new_algo;
+	/* populate the type charging profile the algorithm handles */
+	pse_algo.chrg_prof_type = PSY_CHRG_PROF_PSE;
+	pse_algo.name = "my_algo";
+	/* callback function to retrieve CC and CV */
+	pse_algo.get_next_cc_cv = my_algo_get_next_cc_cv;
+	/* callback function to retreive battery thresholds */
+	pse_algo.get_batt_thresholds = my_algo_get_bat_thresholds;
+	/* register charging algorithm */
+	power_supply_register_charging_algo(&pse_algo);
+
+When the type of charging profile reported by battery and algorithm matches,
+the algorithm will be invoked to get the charging parameters.
+
+6. TODO
+=======
+* Replace static cable array with dynamic list
+* Implement safety timer and watch dog timer features with more monitoring
+* Move charge full detection logic to psy charging driver from algorithm driver
drivers/power/Kconfig +8 -0
index 0196acf..f679f82 100644
@@ -14,6 +14,14 @@ config POWER_SUPPLY_DEBUG
 	  Say Y here to enable debugging messages for power supply class
 	  and drivers.
 
+config POWER_SUPPLY_CHARGER
+	bool "Power Supply Charger"
+	help
+	  Say Y here to enable the power supply charging control driver. Charging
+	  control supports charging in a generic way. This allows the charger
+	  drivers to keep the charging logic outside and the charger driver
+	  just need to abstract the charger hardware.
+
 config PDA_POWER
 	tristate "Generic PDA/phone power driver"
 	depends on !S390
drivers/power/Makefile +1 -0
index ee54a3e..405f0f4 100644
@@ -7,6 +7,7 @@ power_supply-$(CONFIG_LEDS_TRIGGERS)	+= power_supply_leds.o
 obj-$(CONFIG_POWER_SUPPLY)	+= power_supply.o
 obj-$(CONFIG_GENERIC_ADC_BATTERY)	+= generic-adc-battery.o
 
+obj-$(CONFIG_POWER_SUPPLY_CHARGER) += power_supply_charger.o
 obj-$(CONFIG_PDA_POWER)		+= pda_power.o
 obj-$(CONFIG_APM_POWER)		+= apm_power.o
 obj-$(CONFIG_MAX8925_POWER)	+= max8925_power.o
drivers/power/power_supply_charger.c new file +1186 -0

1,186 lines not on this page · raw message

drivers/power/power_supply_charger.h new file +218 -0

218 lines not on this page · raw message

drivers/power/power_supply_core.c +3 -0
index 2660664..1daa5c2 100644
@@ -483,6 +483,9 @@ static int ps_set_cur_charge_cntl_limit(struct thermal_cooling_device *tcd,
 	ret = psy->set_property(psy,
 		POWER_SUPPLY_PROP_CHARGE_CONTROL_LIMIT, &val);
 
+	atomic_notifier_call_chain(&power_supply_notifier,
+		PSY_EVENT_THROTTLE, psy);
+
 	return ret;
 }
 
include/linux/power/power_supply_charger.h new file +280 -0

280 lines not on this page · raw message

include/linux/power_supply.h +160 -0

170 lines not on this page · raw message

Signature · 1 line
1.7.9.5
#5 lore raw

Re: [PATCHv8 2/4] power_supply: Introduce generic psy charging driver

Hi!

Quoted 4 lines

The Power Supply charging driver connects multiple subsystems
to do charging in a generic way. The subsystems involves power_supply, thermal and battery communication subsystems (1wire).With this the charging is handled in a generic way.

" " after ".", please.

Documentation/power/power_supply_charger.txt line 6

Quoted 4 lines
Documentation/power/power_supply_charger.txt new file +4 -0
+
+The Power Supply charging driver connects multiple subsystems
+to do charging in a generic way. The subsystems involves power_supply,
+thermal and battery communication subsystems (1wire).With this the charging is

Here too.

Documentation/power/power_supply_charger.txt line 14

Quoted 7 lines
Documentation/power/power_supply_charger.txt new file +7 -0
+
+The driver introduces different new features - Battery Identification
+interfaces, pluggable charging algorithms, charger cable arbitrations etc.
+
+In existing driver implementations the charging is done based on the static
+battery characteristics. This is done at the boot time by passing the battery
+properties (max_voltage, capacity) etc. as a platform data to the

-> (max_voltage, capacity, etc.)

Quoted 4 lines
drivers/power/Kconfig +0 -0
@@ -14,6 +14,14 @@ config POWER_SUPPLY_DEBUG
 	  Say Y here to enable debugging messages for power supply class

drivers/power/Kconfig config POWER_SUPPLY_DEBUG · line 24

Quoted 10 lines
drivers/power/Kconfig +8 -0
 	  and drivers.
 
+config POWER_SUPPLY_CHARGER
+	bool "Power Supply Charger"
+	help
+	  Say Y here to enable the power supply charging control driver. Charging
+	  control supports charging in a generic way. This allows the charger
+	  drivers to keep the charging logic outside and the charger driver
+	  just need to abstract the charger hardware.
+

Umm. This is not too helpful for our users.

Quoted 12 lines
drivers/power/Kconfig +12 -0
+struct psy_charger_context {
+	bool is_usb_cable_evt_reg;
+	int psyc_cnt;
+	int batt_status;
+	/* cache battery and charger properties */
+	struct list_head chrgr_cache_lst;
+	struct list_head batt_cache_lst;
+	struct mutex event_lock;
+	struct list_head event_queue;
+	struct psy_batt_chrg_prof batt_property;
+	wait_queue_head_t wait_chrg_enable;
+	spinlock_t battid_spinlock;

drivers/power/Kconfig config POWER_SUPPLY_DEBUG

Quoted 10 lines
drivers/power/Kconfig +10 -0
+	spinlock_t event_queue_lock;
+	struct work_struct event_work;
+};
+
+struct charger_cable {
+	struct psy_cable_props cable_props;
+	enum psy_charger_cable_type psy_cable_type;
+};
+
+static struct psy_charger_context psy_chrgr;

You still miss some wovels here. Sometimes it imakes it unlear:
chrg is charge? charger?

drivers/power/Kconfig

Quoted 4 lines
drivers/power/Kconfig +4 -0
+static inline bool psy_is_charger_prop_changed(struct psy_charger_props prop,
+		struct psy_charger_props cache_prop)
+{
+	/* if online/prsent/health/is_charging is changed, then return true */

Typo - present.

drivers/power/Kconfig

Quoted 8 lines
drivers/power/Kconfig +8 -0
+static inline void cache_chrgr_prop(struct psy_charger_props *chrgr_prop_new)
+{
+	struct psy_charger_props *chrgr_cache;
+
+	list_for_each_entry(chrgr_cache, &psy_chrgr.chrgr_cache_lst, node) {
+		if (!strcmp(chrgr_cache->name, chrgr_prop_new->name))
+			goto update_props;
+	}

Interesting use of goto. Maybe update_properties should be separate function?

Quoted 6 lines
drivers/power/Kconfig +6 -0
+update_props:
+	chrgr_cache->is_charging = chrgr_prop_new->is_charging;
+	chrgr_cache->online = chrgr_prop_new->online;
+	chrgr_cache->health = chrgr_prop_new->health;
+	chrgr_cache->present = chrgr_prop_new->present;
+	chrgr_cache->cable = chrgr_prop_new->cable;

drivers/power/Kconfig

Quoted 10 lines
drivers/power/Kconfig +10 -0
+	chrgr_cache->tstamp = chrgr_prop_new->tstamp;
+	chrgr_cache->psyc = chrgr_prop_new->psyc;
+}
+
+	chrgr_prop.psyc = chrgr_prop_cache.psyc;
+	cache_chrgr_prop(&chrgr_prop);
+	return true;
+}
+static void cache_successive_samples(long *sample_array, long new_sample)
+{

Add empty line between the functions.

Quoted 4 lines
drivers/power/Kconfig +4 -0
+static inline void cache_bat_prop(struct psy_batt_props *bat_prop_new)
+{
+	struct psy_batt_props *bat_cache;
+

drivers/power/Kconfig

Quoted 10 lines
drivers/power/Kconfig +10 -0
+	/*
+	*  Find entry in cache list. If an entry is located update
+	*  the existing entry else create new entry in the list
+	*/
+	list_for_each_entry(bat_cache, &psy_chrgr.batt_cache_lst, node) {
+		if (!strcmp(bat_cache->name, bat_prop_new->name))
+			goto update_props;
+	}
+
+	bat_cache = kzalloc(sizeof(*bat_cache), GFP_KERNEL);

What is it with all the caching? Is it good idea to have caches indexed by strings? Can't you go without caching, or attach caches to some structure? Interesting goto again.

drivers/power/Kconfig

Quoted 8 lines
drivers/power/Kconfig +8 -0
+static inline void get_cur_bat_prop(struct power_supply *psy,
+				    struct psy_batt_props *bat_prop)
+{
+	struct psy_batt_props bat_prop_cache;
+	int ret;
+
+	bat_prop->name = psy->name;
+	bat_prop->voltage_now = PSY_VOLTAGE_OCV(psy) / 1000;

Not sure what OCV means...

drivers/power/Kconfig

Quoted 10 lines
drivers/power/Kconfig +10 -0
+static inline bool is_supplied_to_has_ext_pwr_changed(struct power_supply *psy)
+{
+	int i;
+	struct power_supply *psb;
+	bool is_pwr_changed_defined = true;
+
+	for (i = 0; i < psy->num_supplicants; i++) {
+		psb =
+		    power_supply_get_by_name(psy->
+					     supplied_to[i]);

You can do it on one line, right?

drivers/power/Kconfig

Quoted 2 lines
drivers/power/Kconfig +2 -0
+		if (psb && !psb->external_power_changed)
+			is_pwr_changed_defined &= false;

WTF?

Quoted 11 lines
drivers/power/Kconfig +11 -0
+static int get_supplied_by_list(struct power_supply *psy,
+				struct power_supply *psy_lst[])
+{
+	struct class_dev_iter iter;
+	struct device *dev;
+	struct power_supply *pst;
+	int cnt = 0, i, j;
+
+	if (!psy_is_battery(psy))
+		return 0;
+

drivers/power/Kconfig

Quoted 10 lines
drivers/power/Kconfig +10 -0
+	/* Identify chargers which are supplying power to the battery */
+	class_dev_iter_init(&iter, power_supply_class, NULL, NULL);
+	while ((dev = class_dev_iter_next(&iter))) {
+		pst = (struct power_supply *)dev_get_drvdata(dev);
+		if (!psy_is_charger(pst))
+			continue;
+		for (i = 0; i < pst->num_supplicants; i++) {
+			if (!strcmp(pst->supplied_to[i], psy->name))
+				psy_lst[cnt++] = pst;
+		}

Awful lot of string compares around.

drivers/power/Kconfig

Quoted 6 lines
drivers/power/Kconfig +6 -0
+	/* sort based on priority. 0 has the highest priority  */
+	for (i = 0; i < cnt; ++i)
+		for (j = 0; j < cnt; ++j)
+			if (psy_prioirty(psy_lst[j]) > psy_prioirty(psy_lst[i]))
+				swap(psy_lst[j], psy_lst[i]);
+

WTF? Bubble sort in kernel?

priority is misspelled...

Does it really be so complex? Dynamically allocated caches, name compares everywhere?

Pavel

Signature · 2 lines
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
author
#6 lore raw

Re: [PATCHv8 2/4] power_supply: Introduce generic psy charging driver

On Fri, Mar 07, 2014 at 09:25:20PM +0100, Pavel Machek wrote:

Hi,

Quoted 6 lines
Quoted 4 lines

The Power Supply charging driver connects multiple subsystems
to do charging in a generic way. The subsystems involves power_supply, thermal and battery communication subsystems (1wire).With this the charging is handled in a generic way.

" " after ".", please.

Will fix in next patch set

Quoted 6 lines

Documentation/power/power_supply_charger.txt line 6

Quoted 4 lines
Documentation/power/power_supply_charger.txt new file +4 -0
+
+The Power Supply charging driver connects multiple subsystems
+to do charging in a generic way. The subsystems involves power_supply,
+thermal and battery communication subsystems (1wire).With this the charging is

Here too.

Same here

Quoted 12 lines

Documentation/power/power_supply_charger.txt line 14

Quoted 10 lines
Documentation/power/power_supply_charger.txt new file +7 -0
+
+The driver introduces different new features - Battery Identification
+interfaces, pluggable charging algorithms, charger cable arbitrations etc.
+
+In existing driver implementations the charging is done based on the static
+battery characteristics. This is done at the boot time by passing the battery
+properties (max_voltage, capacity) etc. as a platform data to the

-> (max_voltage, capacity, etc.)

Same here

Quoted 7 lines
Quoted 7 lines
drivers/power/Kconfig +0 -0
@@ -14,6 +14,14 @@ config POWER_SUPPLY_DEBUG
 	  Say Y here to enable debugging messages for power supply class
 	  and drivers.
 
Quoted 10 lines

drivers/power/Kconfig config POWER_SUPPLY_DEBUG · line 24

Quoted 8 lines
drivers/power/Kconfig +8 -0
+config POWER_SUPPLY_CHARGER
+	bool "Power Supply Charger"
+	help
+	  Say Y here to enable the power supply charging control driver. Charging
+	  control supports charging in a generic way. This allows the charger
+	  drivers to keep the charging logic outside and the charger driver
+	  just need to abstract the charger hardware.
+

Umm. This is not too helpful for our users.

Will add more text to help users.

Quoted 18 lines
Quoted 18 lines
drivers/power/Kconfig +17 -0
+struct psy_charger_context {
+	bool is_usb_cable_evt_reg;
+	int psyc_cnt;
+	int batt_status;
+	/* cache battery and charger properties */
+	struct list_head chrgr_cache_lst;
+	struct list_head batt_cache_lst;
+	struct mutex event_lock;
+	struct list_head event_queue;
+	struct psy_batt_chrg_prof batt_property;
+	wait_queue_head_t wait_chrg_enable;
+	spinlock_t battid_spinlock;
+	spinlock_t event_queue_lock;
+	struct work_struct event_work;
+};
+
+struct charger_cable {
Quoted 8 lines

drivers/power/Kconfig

Quoted 5 lines
drivers/power/Kconfig +5 -0
+	struct psy_cable_props cable_props;
+	enum psy_charger_cable_type psy_cable_type;
+};
+
+static struct psy_charger_context psy_chrgr;

You still miss some wovels here. Sometimes it imakes it unlear:
chrg is charge? charger?

chrgr means charger, chrg means charge. Isn't it used consistently?. Can fix it if it's really annoying. Please suggest.

Quoted 8 lines

drivers/power/Kconfig

Quoted 6 lines
drivers/power/Kconfig +4 -0
+static inline bool psy_is_charger_prop_changed(struct psy_charger_props prop,
+		struct psy_charger_props cache_prop)
+{
+	/* if online/prsent/health/is_charging is changed, then return true */

Typo - present.

Will fix in next patch set

Quoted 11 lines

drivers/power/Kconfig

Quoted 9 lines
drivers/power/Kconfig +8 -0
+static inline void cache_chrgr_prop(struct psy_charger_props *chrgr_prop_new)
+{
+	struct psy_charger_props *chrgr_cache;
+
+	list_for_each_entry(chrgr_cache, &psy_chrgr.chrgr_cache_lst, node) {
+		if (!strcmp(chrgr_cache->name, chrgr_prop_new->name))
+			goto update_props;
+	}

Interesting use of goto. Maybe update_properties should be separate function?

I feel, having a function just for few assignments may not be a good idea. What about having memcpy?

Quoted 11 lines
Quoted 11 lines
drivers/power/Kconfig +10 -0
+update_props:
+	chrgr_cache->is_charging = chrgr_prop_new->is_charging;
+	chrgr_cache->online = chrgr_prop_new->online;
+	chrgr_cache->health = chrgr_prop_new->health;
+	chrgr_cache->present = chrgr_prop_new->present;
+	chrgr_cache->cable = chrgr_prop_new->cable;
+	chrgr_cache->tstamp = chrgr_prop_new->tstamp;
+	chrgr_cache->psyc = chrgr_prop_new->psyc;
+}
+
Quoted 8 lines

drivers/power/Kconfig

Quoted 6 lines
drivers/power/Kconfig +6 -0
+	chrgr_prop.psyc = chrgr_prop_cache.psyc;
+	cache_chrgr_prop(&chrgr_prop);
+	return true;
+}
+static void cache_successive_samples(long *sample_array, long new_sample)
+{

Add empty line between the functions.

Ok..Will fix in next patch set

Quoted 11 lines
Quoted 11 lines
drivers/power/Kconfig +10 -0
+static inline void cache_bat_prop(struct psy_batt_props *bat_prop_new)
+{
+	struct psy_batt_props *bat_cache;
+
+	/*
+	*  Find entry in cache list. If an entry is located update
+	*  the existing entry else create new entry in the list
+	*/
+	list_for_each_entry(bat_cache, &psy_chrgr.batt_cache_lst, node) {
+		if (!strcmp(bat_cache->name, bat_prop_new->name))
Quoted 8 lines

drivers/power/Kconfig

Quoted 4 lines
drivers/power/Kconfig +4 -0
+			goto update_props;
+	}
+
+	bat_cache = kzalloc(sizeof(*bat_cache), GFP_KERNEL);

What is it with all the caching? Is it good idea to have caches indexed by strings? Can't you go without caching, or attach caches to some structure? Interesting goto again.

Cache is to store previous battery properties. On receiving a new event compare the properties with cached value to decide charging state (charging/not charging/full etc.) and to control charging. There is added saving with caching. If the previous and current battery properties doesn't differ much, ignore the event instead of continuing (as implemented in is_trigger_charging_algo) with invoking charging algorithms. Also caching helps to take few critical charging decisions - like charge termination which need charge current and voltage over a period of time.

Since a power_supply object (driver) is identified by name, it's the only index available.

goto can be fixed by a memcpy, but just wanted to keep the assignments to make it more meaningful than having a dumb copy.

Quoted 11 lines

drivers/power/Kconfig

Quoted 9 lines
drivers/power/Kconfig +8 -0
+static inline void get_cur_bat_prop(struct power_supply *psy,
+				    struct psy_batt_props *bat_prop)
+{
+	struct psy_batt_props bat_prop_cache;
+	int ret;
+
+	bat_prop->name = psy->name;
+	bat_prop->voltage_now = PSY_VOLTAGE_OCV(psy) / 1000;

Not sure what OCV means...

Open Circuit voltage which is a standard battery property and is defined in power_supply.h

Quoted 13 lines

drivers/power/Kconfig

Quoted 11 lines
drivers/power/Kconfig +10 -0
+static inline bool is_supplied_to_has_ext_pwr_changed(struct power_supply *psy)
+{
+	int i;
+	struct power_supply *psb;
+	bool is_pwr_changed_defined = true;
+
+	for (i = 0; i < psy->num_supplicants; i++) {
+		psb =
+		    power_supply_get_by_name(psy->
+					     supplied_to[i]);

You can do it on one line, right?

Agreed. Will fix it in next patch set

Quoted 5 lines

drivers/power/Kconfig

Quoted 3 lines
drivers/power/Kconfig +2 -0
+		if (psb && !psb->external_power_changed)
+			is_pwr_changed_defined &= false;

WTF?

The is_supplied_to_has_ext_pwr_changed() return true if any of the power_supply objects in supplied_to list has the external_power_changed() call back defined. This to optimize the notifications and to reduce charging algo invocations. This is used in is_trigger_charging_algo() to decide charging algorithms should be invoked or not.

Quoted 16 lines
Quoted 16 lines
drivers/power/Kconfig +15 -0
+static int get_supplied_by_list(struct power_supply *psy,
+				struct power_supply *psy_lst[])
+{
+	struct class_dev_iter iter;
+	struct device *dev;
+	struct power_supply *pst;
+	int cnt = 0, i, j;
+
+	if (!psy_is_battery(psy))
+		return 0;
+
+	/* Identify chargers which are supplying power to the battery */
+	class_dev_iter_init(&iter, power_supply_class, NULL, NULL);
+	while ((dev = class_dev_iter_next(&iter))) {
+		pst = (struct power_supply *)dev_get_drvdata(dev);
Quoted 8 lines

drivers/power/Kconfig

Quoted 6 lines
drivers/power/Kconfig +6 -0
+		if (!psy_is_charger(pst))
+			continue;
+		for (i = 0; i < pst->num_supplicants; i++) {
+			if (!strcmp(pst->supplied_to[i], psy->name))
+				psy_lst[cnt++] = pst;
+		}

Awful lot of string compares around.

In reality this would be one/two, just because the number of chargers supplying power to a battery would be limited.

Quoted 10 lines

drivers/power/Kconfig

Quoted 8 lines
drivers/power/Kconfig +6 -0
+	/* sort based on priority. 0 has the highest priority  */
+	for (i = 0; i < cnt; ++i)
+		for (j = 0; j < cnt; ++j)
+			if (psy_prioirty(psy_lst[j]) > psy_prioirty(psy_lst[i]))
+				swap(psy_lst[j], psy_lst[i]);
+

WTF? Bubble sort in kernel?

Yes, it's bubble sort. Since the number of power supply objects in real systems (max 4) are limited, I feel the complexity would be as same as any other sorting algorithms. Any suggestions?

Quoted 2 lines

priority is misspelled...

Will fix it in next patch set.

Quoted 3 lines

Does it really be so complex? Dynamically allocated caches, name compares everywhere?

It's really not so complex. The caches are allocated dynamically only when a new charger/battery power supply object gets registered - basically when a charger/battery driver gets loaded. At runtime no dynamic allocation is needed. There is not too many string comparisons just because the number of power supply objects are limited. Power supply subsystem has only name as unique identifier (psy->name). I couldn't find a better way without string comparisons to identify a power supply object.

Thanks
Jenny

#7 lore raw

Re: [PATCHv8 2/4] power_supply: Introduce generic psy charging driver

Hi!

Quoted 5 lines
Quoted 2 lines

You still miss some wovels here. Sometimes it imakes it unlear:
chrg is charge? charger?

chrgr means charger, chrg means charge. Isn't it used consistently?. Can fix it if it's really annoying. Please suggest.

Well... with all the missing letters, it is not clear if letter is missing because you just made it short, or it is real difference.

Please just use full words. ... but read below.

Quoted 20 lines
Quoted 10 lines

drivers/power/power_supply_charger.c line 296

Quoted 6 lines
drivers/power/power_supply_charger.c new file +6 -0
+	list_for_each_entry(bat_cache, &psy_chrgr.batt_cache_lst, node) {
+		if (!strcmp(bat_cache->name, bat_prop_new->name))
+			goto update_props;
+	}
+
+	bat_cache = kzalloc(sizeof(*bat_cache), GFP_KERNEL);

What is it with all the caching? Is it good idea to have caches indexed by strings? Can't you go without caching, or attach caches to some structure? Interesting goto again.

Cache is to store previous battery properties. On receiving a new event compare the properties with cached value to decide charging state (charging/not charging/full etc.) and to control charging. There is added saving with caching. If the previous and current battery properties doesn't differ much, ignore the event instead of continuing (as implemented in is_trigger_charging_algo) with invoking charging algorithms. Also caching helps to take few critical charging decisions - like charge termination which need charge current and voltage over a period of time.

Quoted 2 lines

Since a power_supply object (driver) is identified by name, it's the only index available.

Why can't you just use address of struct power_supply? There should be no need to work with names.

Feel free to extend struct power_supply, wrap it in another struct, anything.

Quoted 10 lines
Quoted 4 lines

drivers/power/power_supply_charger.c line 400

Quoted 2 lines
drivers/power/power_supply_charger.c new file +2 -0
+		if (psb && !psb->external_power_changed)
+			is_pwr_changed_defined &= false;

WTF?

The is_supplied_to_has_ext_pwr_changed() return true if any of the power_supply objects in supplied_to list has the external_power_changed() call back defined. This to optimize the notifications and to reduce charging algo invocations. This is used in is_trigger_charging_algo() to decide charging algorithms should be invoked or not.

No, I mean... using "&=" operator in case where plain assignment should work is evil.

Quoted 13 lines
Quoted 4 lines
Quoted 4 lines
drivers/power/power_supply_charger.c new file +4 -0
+	/* Identify chargers which are supplying power to the battery */
+	class_dev_iter_init(&iter, power_supply_class, NULL, NULL);
+	while ((dev = class_dev_iter_next(&iter))) {
+		pst = (struct power_supply *)dev_get_drvdata(dev);
Quoted 8 lines

drivers/power/power_supply_charger.c line 605

Quoted 6 lines
drivers/power/power_supply_charger.c new file +6 -0
+		if (!psy_is_charger(pst))
+			continue;
+		for (i = 0; i < pst->num_supplicants; i++) {
+			if (!strcmp(pst->supplied_to[i], psy->name))
+				psy_lst[cnt++] = pst;
+		}

Awful lot of string compares around.

Quoted 2 lines

In reality this would be one/two, just because the number of chargers supplying power to a battery would be limited.

This whole file is nothing but string compares, bubble sorts, and weird caches.

Please find a way to simplify a design. Rafael works for same company, can he help?

Quoted 10 lines
Quoted 2 lines

Does it really be so complex? Dynamically allocated caches, name compares everywhere?

It's really not so complex. The caches are allocated dynamically only when a new charger/battery power supply object gets registered - basically when a charger/battery driver gets loaded. At runtime no dynamic allocation is needed. There is not too many string comparisons just because the number of power supply objects are limited. Power supply subsystem has only name as unique identifier (psy->name). I couldn't find a better way without string comparisons to identify a power supply object.

(void *) psy is also an unique identifier. Plus, you can add new fields to psy.

Pavel

Signature · 2 lines
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
#8 lore raw

Re: [PATCHv8 2/4] power_supply: Introduce generic psy charging driver

In reply to Jenny Tc
Quoted 9 lines · Jenny Tc
Quoted 7 lines · Pavel Machek
Quoted 6 lines
drivers/power/power_supply_charger.c new file +6 -0
+   /* sort based on priority. 0 has the highest priority  */
+   for (i = 0; i < cnt; ++i)
+           for (j = 0; j < cnt; ++j)
+                   if (psy_prioirty(psy_lst[j]) > psy_prioirty(psy_lst[i]))
+                           swap(psy_lst[j], psy_lst[i]);
+
Quoted 5 lines
Quoted 1 line

WTF? Bubble sort in kernel?

Yes, it's bubble sort. Since the number of power supply objects in real systems (max 4) are limited, I feel the complexity would be as same as any other sorting algorithms. Any suggestions?

You already have a kernel quicksort implementation in lib/sort.c.

Please restructure the code to make use of this as it is already compiled into every kernel.

Yours,
Linus Walleij

#9 lore raw

Re: [PATCHv8 2/4] power_supply: Introduce generic psy charging driver

Quoted 12 lines · Linus Walleij
Quoted 11 lines · Jenny Tc
Quoted 8 lines · Pavel Machek

drivers/power/power_supply_charger.c line 617

Quoted 6 lines
drivers/power/power_supply_charger.c new file +6 -0
+   /* sort based on priority. 0 has the highest priority  */
+   for (i = 0; i < cnt; ++i)
+           for (j = 0; j < cnt; ++j)
+                   if (psy_prioirty(psy_lst[j]) > psy_prioirty(psy_lst[i]))
+                           swap(psy_lst[j], psy_lst[i]);
+

WTF? Bubble sort in kernel?

Quoted 8 lines
Quoted 3 lines

Yes, it's bubble sort. Since the number of power supply objects in real systems (max 4) are limited, I feel the complexity would be as same as any other sorting algorithms. Any suggestions?

You already have a kernel quicksort implementation in lib/sort.c.

Please restructure the code to make use of this as it is already compiled into every kernel.

Actually.. I believe the code needs more surgery than that. It should not need those string lookups -- just attach data directly to struct psy. And yes, if sort remains, it needs to use library function, but I strongly believe sorting should not be neccessary here.

Pavel

Signature · 2 lines
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
#10 lore raw

Re: [PATCHv8 2/4] power_supply: Introduce generic psy charging driver

In reply to Jenny TC
Quoted 5 lines · Jenny TC
include/linux/power/power_supply_charger.h new file +5 -0
+enum psy_charger_cable_type {
+       PSY_CHARGER_CABLE_TYPE_NONE = 0,
+       PSY_CHARGER_CABLE_TYPE_USB_SDP = 1 << 0,
+       PSY_CHARGER_CABLE_TYPE_USB_DCP = 1 << 1,
+       PSY_CHARGER_CABLE_TYPE_USB_CDP = 1 << 2,

include/linux/power/power_supply_charger.h line 41

Quoted 10 lines
include/linux/power/power_supply_charger.h new file +10 -0
+       PSY_CHARGER_CABLE_TYPE_USB_ACA = 1 << 3,
+       PSY_CHARGER_CABLE_TYPE_AC = 1 << 4,
+       PSY_CHARGER_CABLE_TYPE_ACA_DOCK = 1 << 5,
+       PSY_CHARGER_CABLE_TYPE_ACA_A = 1 << 6,
+       PSY_CHARGER_CABLE_TYPE_ACA_B = 1 << 7,
+       PSY_CHARGER_CABLE_TYPE_ACA_C = 1 << 8,
+       PSY_CHARGER_CABLE_TYPE_SE1 = 1 << 9,
+       PSY_CHARGER_CABLE_TYPE_MHL = 1 << 10,
+       PSY_CHARGER_CABLE_TYPE_B_DEVICE = 1 << 11,
+};

I still disagree with using an enum as bitfield.

Atleast
#include <linux/bitops.h>
and use BIT(0), BIT(1) etc to define the bits.

Yours,
Linus Walleij

author
#11 lore raw

PATCH v8 3/4 power_supply: Introduce PSE compliant algorithm

In reply to Jenny TC

As per Product Safety Engineering (PSE) specification for battery charging, the battery characteristics and thereby the charging rates can vary on different temperature zones. This patch introduces a PSE compliant charging algorithm with maintenance charging support. The algorithm can be selected by the power supply charging driver based on the type of the battery charging profile.

---
 drivers/power/Kconfig                      |   15 ++
 drivers/power/Makefile                     |    1 +
 drivers/power/charging_algo_pse.c          |  204 ++++++++++++++++++++++++++++
 include/linux/power/power_supply_charger.h |   63 +++++++++
 4 files changed, 283 insertions(+)
 create mode 100644 drivers/power/charging_algo_pse.c
drivers/power/Kconfig +15 -0
index f679f82..54a0321 100644
@@ -22,6 +22,21 @@ config POWER_SUPPLY_CHARGER
 	  drivers to keep the charging logic outside and the charger driver
 	  just need to abstract the charger hardware.
 
+config POWER_SUPPLY_CHARGING_ALGO_PSE
+	bool "PSE compliant charging algorithm"
+	depends on POWER_SUPPLY_CHARGER
+	help
+	  Say Y here to select Product Safety Engineering (PSE) compliant
+	  charging algorithm. As per PSE standard the battery characteristics
+	  and thereby the charging rates can vary on different temperature
+	  zones. Select this if your charging algorithm need to change the
+	  charging parameters based on the battery temperature and the battery
+	  charging profile follows the struct psy_pse_chrg_prof definition.
+	  This  config will enable PSE compliant charging algorithm with
+	  maintenance charging support. At runtime the algorithm will be
+	  selected by the psy charger driver based on the type of the battery
+	  charging profile.
+
 config PDA_POWER
 	tristate "Generic PDA/phone power driver"
 	depends on !S390
drivers/power/Makefile +1 -0
index 405f0f4..77535fd 100644
@@ -8,6 +8,7 @@ obj-$(CONFIG_POWER_SUPPLY)	+= power_supply.o
 obj-$(CONFIG_GENERIC_ADC_BATTERY)	+= generic-adc-battery.o
 
 obj-$(CONFIG_POWER_SUPPLY_CHARGER) += power_supply_charger.o
+obj-$(CONFIG_POWER_SUPPLY_CHARGING_ALGO_PSE) += charging_algo_pse.o
 obj-$(CONFIG_PDA_POWER)		+= pda_power.o
 obj-$(CONFIG_APM_POWER)		+= apm_power.o
 obj-$(CONFIG_MAX8925_POWER)	+= max8925_power.o
drivers/power/charging_algo_pse.c new file +204 -0

204 lines not on this page · raw message

include/linux/power/power_supply_charger.h +63 -0

71 lines not on this page · raw message

Signature · 1 line
1.7.9.5
author
#12 lore raw

PATCH v8 4/4 power_supply: bq24261 charger driver

In reply to Jenny TC

This patch introduces BQ24261 charger driver. The driver makes use of power supply charging driver to setup charging. So the driver does hardware abstraction and handles h/w specific corner cases. The charging logic resides with power supply charging driver

---
 drivers/power/Kconfig                 |   10 +
 drivers/power/Makefile                |    1 +
 drivers/power/bq24261-charger.c       | 1350 +++++++++++++++++++++++++++++++++
 include/linux/power/bq24261-charger.h |   25 +
 4 files changed, 1386 insertions(+)
 create mode 100644 drivers/power/bq24261-charger.c
 create mode 100644 include/linux/power/bq24261-charger.h
drivers/power/Kconfig +10 -0
index 54a0321..4ff080c 100644
@@ -411,6 +411,16 @@ config BATTERY_GOLDFISH
 	  Say Y to enable support for the battery and AC power in the
 	  Goldfish emulator.
 
+config CHARGER_BQ24261
+	tristate "BQ24261 charger driver"
+	select POWER_SUPPLY_CHARGER
+	depends on I2C
+	help
+	  Say Y to include support for BQ24261 Charger driver. This driver
+	  makes use of power supply charging driver. So the driver gives
+	  the charger hardware abstraction only. Charging logic is abstracted
+	  in the power supply charging driver.
+
 source "drivers/power/reset/Kconfig"
 
 endif # POWER_SUPPLY
drivers/power/Makefile +1 -0
index 77535fd..9dde895 100644
@@ -59,4 +59,5 @@ obj-$(CONFIG_CHARGER_BQ24735)	+= bq24735-charger.o
 obj-$(CONFIG_POWER_AVS)		+= avs/
 obj-$(CONFIG_CHARGER_SMB347)	+= smb347-charger.o
 obj-$(CONFIG_CHARGER_TPS65090)	+= tps65090-charger.o
+obj-$(CONFIG_CHARGER_BQ24261)	+= bq24261-charger.o
 obj-$(CONFIG_POWER_RESET)	+= reset/
drivers/power/bq24261-charger.c new file +1350 -0

1,350 lines not on this page · raw message

include/linux/power/bq24261-charger.h new file +25 -0
new file mode 100644
index 0000000..4ff02af
@@ -0,0 +1,25 @@
+/*
+ * bq24261_charger.h: platform data structure for bq24261 driver
+ *
+ * (C) Copyright 2012 Intel Corporation
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; version 2
+ * of the License.
+ */
+
+#ifndef __BQ24261_CHARGER_H__
+#define __BQ24261_CHARGER_H__
+
+struct bq24261_plat_data {
+	char **supplied_to;
+	size_t num_supplicants;
+	struct psy_throttle_state *throttle_states;
+	size_t num_throttle_states;
+	int safety_timer;
+	bool is_ts_enabled;
+
+};
+
+#endif
Signature · 1 line
1.7.9.5
#13 lore raw
In reply to Jenny TC

Hi Jenny,

Would you mind _not_ CC'ing me on all of your emails please?

I have enough _relevant_ email to deal with already without seeing this churn constantly.

Thanks in advance.

Quoted 61 lines

v1: introduced feature as a framework within power supply class driver with
separate files for battid framework and charging framework
v2: fixed review comments, moved macros and inline functions to power_supply.h
v3: moved the feature as a separate driver, combined battid framework and
charging framework inside the power supply charging driver. Moved charger specific properties to power_supply_charger.h and plugged the driver with power supply subsystem using power_supply_notifier introduced in my previous patch. Also a sample charger chip driver (bq24261) patch added to give more idea on the psy charging driver usage
v4: Fixed review comments, no major design changes.
v5: Fixed makefile inconsistencies, removed unused pdata callbacks
v6: Fixed nested loops, commenting style
v7: added kerneldocs for structs and minor fixes
v8: used msecs_to_jiffies instead of HZ directly, modified Kconfig help text
for POWER_SUPPLY_CHARGING_ALGO_PSE

The Power Supply charging driver connects multiple subsystems
to do charging in a generic way. The subsystems involves power_supply, thermal and battery communication subsystems (1wire).With this the charging is handled in a generic way.

The driver makes use of different new features - Battery Identification interfaces, pluggable charging algorithms, charger cable arbitrations etc. The patch also introduces generic interface for charger cable notifications. Charger cable events and capabilities can be notified using the generic power_supply_notifier chain.

Overall this driver removes the charging logic out of the charger chip driver and the charger chip driver can just listen to the request from the power supply charging driver to set the charger properties. This can be implemented by exposing get_property and set property callbacks.

Jenny TC (4):
power_supply: Add inlmt,iterm, min/max temp props
power_supply: Introduce generic psy charging driver
power_supply: Introduce PSE compliant algorithm
power_supply: bq24261 charger driver

 Documentation/power/power_supply_charger.txt |  353 +++++++
 Documentation/power/power_supply_class.txt   |    6 +
 drivers/power/Kconfig                        |   31 +
 drivers/power/Makefile                       |    3 +
 drivers/power/bq24261-charger.c              | 1350 ++++++++++++++++++++++++++
 drivers/power/charging_algo_pse.c            |  204 ++++
 drivers/power/power_supply_charger.c         | 1186 ++++++++++++++++++++++
 drivers/power/power_supply_charger.h         |  218 +++++
 drivers/power/power_supply_core.c            |    3 +
 drivers/power/power_supply_sysfs.c           |    4 +
 include/linux/power/bq24261-charger.h        |   25 +
 include/linux/power/power_supply_charger.h   |  341 +++++++
 include/linux/power_supply.h                 |  164 ++++
 13 files changed, 3888 insertions(+)
 create mode 100644 Documentation/power/power_supply_charger.txt
 create mode 100644 drivers/power/bq24261-charger.c
 create mode 100644 drivers/power/charging_algo_pse.c
 create mode 100644 drivers/power/power_supply_charger.c
 create mode 100644 drivers/power/power_supply_charger.h
 create mode 100644 include/linux/power/bq24261-charger.h
 create mode 100644 include/linux/power/power_supply_charger.h
Signature · 4 lines
Lee Jones
Linaro STMicroelectronics Landing Team Lead
Linaro.org │ Open source software for ARM SoCs
Follow Linaro: Facebook | Twitter | Blog