linux-stable.git/drivers/power, branch v6.9.2

power: supply: mt6360_charger: Fix of_match for usb-otg-vbus regulator

2024-04-15T11:31:37+00:00

The of_match shall correspond to the name of the regulator subnode,
or the deprecated `regulator-compatible` property must be used:
failing to do so, the regulator won't probe (and the driver will
as well not probe).

Since the devicetree binding for this driver is actually correct
and wants DTs to use the "usb-otg-vbus-regulator" subnode name,
fix this driver by aligning the `of_match` string to what the DT
binding wants.

Fixes: 0402e8ebb8b8 ("power: supply: mt6360_charger: add MT6360 charger support")
Signed-off-by: AngeloGioacchino Del Regno 
Reviewed-by: Chen-Yu Tsai 
Link: https://lore.kernel.org/r/20240410084405.1389378-1-angelogioacchino.delregno@collabora.com
Signed-off-by: Sebastian Reichel

power: rt9455: hide unused rt9455_boost_voltage_values

2024-04-10T07:57:29+00:00

The rt9455_boost_voltage_values[] array is only used when USB PHY
support is enabled, causing a W=1 warning otherwise:

drivers/power/supply/rt9455_charger.c:200:18: error: 'rt9455_boost_voltage_values' defined but not used [-Werror=unused-const-variable=]

Enclose the definition in the same #ifdef as the references to it.

Fixes: e86d69dd786e ("power_supply: Add support for Richtek RT9455 battery charger")
Signed-off-by: Arnd Bergmann 
Link: https://lore.kernel.org/r/20240403080702.3509288-10-arnd@kernel.org
Signed-off-by: Sebastian Reichel

Merge tag 'for-v6.9' of git://git.kernel.org/pub/scm/linux/kernel/git/sre/linux-power-supply

2024-03-14T17:19:48+00:00

Pull power supply and reset updates from Sebastian Reichel:
 "New features:
   - axp20x_usb_power: report USB type

  Cleanups:
   - convert lots of drivers to use devm_power_supply_register()
   - convert lots of reset drivers to use devm_register_sys_off_handler()
   - constify device_type and power_supply_class
   - axp20x_usb_power: use correct property to report input current limit
   - mm8013: correct handling of "not charging" status register
   - core: fix charge_behaviour formatting
   - minor fixes cleanups"

* tag 'for-v6.9' of git://git.kernel.org/pub/scm/linux/kernel/git/sre/linux-power-supply: (66 commits)
  power: supply: core: fix charge_behaviour formatting
  power: supply: core: ease special formatting implementations
  power: supply: mm8013: fix "not charging" detection
  power: supply: move power_supply_attr_groups definition back to sysfs
  power: supply: core: simplify power_supply_class_init
  power: supply: core: add power_supply_for_each_device()
  power: supply: core: make power_supply_class constant
  power: supply: bq2415x_charger: report online status
  power: supply: core: move power_supply_attr_group into #ifdef block
  power: supply: core: Fix power_supply_init_attrs() stub
  power: supply: bq27xxx: Report charge full state correctly
  power: reset: rmobile-reset: Make sysc_base2 local
  power: supply: core: constify the struct device_type usage
  power: supply: axp288_fuel_gauge: Deny ROCK Pi X
  power: reset: rmobile-reset: Map correct MMIO resource
  power: reset: xgene-reboot: Fix a NULL vs IS_ERR() test
  power: supply: axp288_fuel_gauge: Add STCK1A* Intel Compute Sticks to the deny-list
  power: reset: syscon-poweroff: Use devm_register_sys_off_handler(POWER_OFF)
  power: reset: syscon-poweroff: Move device data into a struct
  power: reset: restart-poweroff: Use devm_register_sys_off_handler(POWER_OFF)
  ...

power: supply: core: fix charge_behaviour formatting

2024-03-05T23:39:45+00:00

This property is documented to have a special format which exposes all
available behaviours and the currently active one at the same time. For
this special format some helpers are provided.

When the charge_behaviour property was added in
1b0b6cc8030d ("power: supply: add charge_behaviour attributes"), it did
not update the default logic in in power_supply_sysfs.c to use the
format helpers. Thus by default only the currently active behaviour
is printed. This fixes the default logic to follow the documented
format.

There is currently only one in-tree drivers exposing charge behaviours -
thinkpad_acpi, which is not affected by the change, as it directly uses
the helpers and does not use the power_supply_sysfs.c logic.

Signed-off-by: Thomas Weißschuh 
Link: https://lore.kernel.org/r/20240303-power_supply-charge_behaviour_prop-v2-3-8ebb0a7c2409@weissschuh.net
Signed-off-by: Sebastian Reichel

power: supply: core: ease special formatting implementations

2024-03-05T23:37:36+00:00

By moving the conditional into the default-branch of the switch new
additions to the switch won't have to bypass the conditional.

This makes it easier to implement those special cases like the upcoming
change to the formatting of "charge_behaviour".

Suggested-by: Hans de Goede 
Link: https://lore.kernel.org/lkml/53082075-852f-4698-b354-ed30e7fd2683@redhat.com/
Signed-off-by: Thomas Weißschuh 
Link: https://lore.kernel.org/r/20240303-power_supply-charge_behaviour_prop-v2-2-8ebb0a7c2409@weissschuh.net
Signed-off-by: Sebastian Reichel

power: supply: mm8013: fix "not charging" detection

2024-03-05T23:34:29+00:00

The charge_behaviours property is meant as a control-knob that can be
changed by the user.

Page 23 of [0] which documents the flag CHG_INH as follows:

  CHG_INH : Charge Inhibit      When the current is more than or equal to charge
                                threshold current,
                                charge inhibit temperature (upper/lower limit) ：1
                                charge permission temperature or the current is
                                less than charge threshold current ：0

So this is pure read-only information which is better represented as
POWER_SUPPLY_STATUS_NOT_CHARGING.

[0] https://product.minebeamitsumi.com/en/product/category/ics/battery/fuel_gauge/parts/download/__icsFiles/afieldfile/2023/07/12/1_download_01_12.pdf

Signed-off-by: Thomas Weißschuh 
Link: https://lore.kernel.org/r/20240303-power_supply-charge_behaviour_prop-v2-1-8ebb0a7c2409@weissschuh.net
Fixes: e39257cde7e8 ("power: supply: mm8013: Add more properties")
Signed-off-by: Sebastian Reichel

power: supply: move power_supply_attr_groups definition back to sysfs

2024-03-05T23:24:26+00:00

As reported by the kernel test robot, 'power_supply_attr_group' is defined
but not used when CONFIG_SYSFS is not set. Sebastian suggested that the
correct fix implemented by this patch, instead of my attempt in commit
ea4367c40c79 ("power: supply: core: move power_supply_attr_group into #ifdef
block"), is to define power_supply_attr_groups in power_supply_sysfs.c and
expose it in the power_supply.h header. For the case where CONFIG_SYSFS=n,
define it as NULL.

Suggested-by: Sebastian Reichel 
Fixes: ea4367c40c79 ("power: supply: core: move power_supply_attr_group into #ifdef block")
Reported-by: kernel test robot 
Closes: https://lore.kernel.org/oe-kbuild-all/202403021518.SUQzk3oA-lkp@intel.com/
Signed-off-by: Ricardo B. Marliere 
Link: https://lore.kernel.org/r/20240303-class_cleanup-power-v2-1-e248b7128519@marliere.net
Signed-off-by: Sebastian Reichel

power: supply: core: simplify power_supply_class_init

2024-03-05T23:21:34+00:00

Technically the sysfs attributes should be initialized
before the class is registered, since that will use them.
As a nice side effect this nicely simplifies the code,
since it allows dropping the helper variable.

Reviewed-by: Ricardo B. Marliere 
Link: https://lore.kernel.org/r/20240301-psy-class-cleanup-v1-2-aebe8c4b6b08@collabora.com
Signed-off-by: Sebastian Reichel

power: supply: core: add power_supply_for_each_device()

2024-03-05T23:21:34+00:00

Introduce power_supply_for_each_device(), which is a wrapper
for class_for_each_device() using the power_supply_class and
going through all devices.

This allows making the power_supply_class itself a local
variable, so that drivers cannot mess with it and simplifies
the code slightly.

Reviewed-by: Ricardo B. Marliere 
Link: https://lore.kernel.org/r/20240301-psy-class-cleanup-v1-1-aebe8c4b6b08@collabora.com
Signed-off-by: Sebastian Reichel

power: supply: core: make power_supply_class constant

2024-03-01T20:16:01+00:00

Since commit 43a7206b0963 ("driver core: class: make class_register() take
a const *"), the driver core allows for struct class to be in read-only
memory, so move the power_supply_class structure to be declared at build
time placing it into read-only memory, instead of having to be dynamically
allocated at boot time.

Cc: Greg Kroah-Hartman 
Suggested-by: Greg Kroah-Hartman 
Signed-off-by: Ricardo B. Marliere 
Reviewed-by: Linus Walleij 
Link: https://lore.kernel.org/r/20240301-class_cleanup-power-v1-1-97e0b7bf9c94@marliere.net
Signed-off-by: Sebastian Reichel