linux.git/drivers/power/supply, branch v6.6

power: supply: qcom_battmgr: fix enable request endianness

2023-09-30T19:00:20+00:00

Add the missing endianness conversion when sending the enable request so
that the driver will work also on a hypothetical big-endian machine.

This issue was reported by sparse.

Fixes: 29e8142b5623 ("power: supply: Introduce Qualcomm PMIC GLINK power supply")
Cc: stable@vger.kernel.org	# 6.3
Cc: Bjorn Andersson 
Signed-off-by: Johan Hovold 
Reviewed-by: Andrew Halaney 
Link: https://lore.kernel.org/r/20230929101649.20206-1-johan+linaro@kernel.org
Signed-off-by: Sebastian Reichel

power: supply: qcom_battmgr: fix battery_id type

2023-09-30T18:58:56+00:00

qcom_battmgr_update_request.battery_id is written to using cpu_to_le32()
and should be of type __le32, just like all other 32bit integer requests
for qcom_battmgr.

Cc: stable@vger.kernel.org	# 6.3
Reported-by: kernel test robot 
Closes: https://lore.kernel.org/oe-kbuild-all/202309162149.4owm9iXc-lkp@intel.com/
Fixes: 29e8142b5623 ("power: supply: Introduce Qualcomm PMIC GLINK power supply")
Reviewed-by: Johan Hovold 
Link: https://lore.kernel.org/r/20230919124222.1155894-1-sebastian.reichel@collabora.com
Signed-off-by: Sebastian Reichel

power: supply: rk817: Fix node refcount leak

2023-09-20T17:37:15+00:00

Dan Carpenter reports that the Smatch static checker warning has found
that there is another refcount leak in the probe function. While
of_node_put() was added in one of the return paths, it should in
fact be added for ALL return paths that return an error and at driver
removal time.

Fixes: 54c03bfd094f ("power: supply: Fix refcount leak in rk817_charger_probe")
Reported-by: Dan Carpenter 
Closes: https://lore.kernel.org/linux-pm/dc0bb0f8-212d-4be7-be69-becd2a3f9a80@kili.mountain/
Signed-off-by: Chris Morgan 
Link: https://lore.kernel.org/r/20230920145644.57964-1-macroalpha82@gmail.com
Signed-off-by: Sebastian Reichel

power: supply: core: fix use after free in uevent

2023-09-19T12:08:34+00:00

power_supply_uevent() which is called to emit a udev event on device
deletion attempts to use the power_supply_battery_info structure,
which is device-managed and has been freed before this point.

Fix this by not generating all battery/charger properties when the
device is about to be removed. This also avoids generating errors
when trying to access the hardware in hot-unplug scenarios.

==================================================================
 BUG: KASAN: slab-use-after-free in power_supply_battery_info_has_prop (power_supply_core.c:872)
 Read of size 4 at addr 0000000062e59028 by task python3/27

 Call Trace:
  power_supply_battery_info_has_prop (power_supply_core.c:872)
  power_supply_uevent (power_supply_sysfs.c:504)
  dev_uevent (drivers/base/core.c:2590)
  kobject_uevent_env (lib/kobject_uevent.c:558)
  kobject_uevent (lib/kobject_uevent.c:643)
  device_del (drivers/base/core.c:3266 drivers/base/core.c:3831)
  device_unregister (drivers/base/core.c:3730 drivers/base/core.c:3854)
  power_supply_unregister (power_supply_core.c:1608)
  devm_power_supply_release (power_supply_core.c:1515)
  release_nodes (drivers/base/devres.c:506)
  devres_release_group (drivers/base/devres.c:669)
  i2c_device_remove (drivers/i2c/i2c-core-base.c:629)
  device_remove (drivers/base/dd.c:570)
  device_release_driver_internal (drivers/base/dd.c:1274 drivers/base/dd.c:1295)
  device_driver_detach (drivers/base/dd.c:1332)
  unbind_store (drivers/base/bus.c:247)
  ...

 Allocated by task 27:
  devm_kmalloc (drivers/base/devres.c:119 drivers/base/devres.c:829)
  power_supply_get_battery_info (include/linux/device.h:316 power_supply_core.c:626)
  __power_supply_register (power_supply_core.c:1408)
  devm_power_supply_register (power_supply_core.c:1544)
  bq256xx_probe (bq256xx_charger.c:1539 bq256xx_charger.c:1727) bq256xx_charger
  i2c_device_probe (drivers/i2c/i2c-core-base.c:584)
  really_probe (drivers/base/dd.c:579 drivers/base/dd.c:658)
  __driver_probe_device (drivers/base/dd.c:800)
  device_driver_attach (drivers/base/dd.c:1128)
  bind_store (drivers/base/bus.c:273)
  ...

 Freed by task 27:
  kfree (mm/slab_common.c:1073)
  release_nodes (drivers/base/devres.c:503)
  devres_release_all (drivers/base/devres.c:536)
  device_del (drivers/base/core.c:3829)
  device_unregister (drivers/base/core.c:3730 drivers/base/core.c:3854)
  power_supply_unregister (power_supply_core.c:1608)
  devm_power_supply_release (power_supply_core.c:1515)
  release_nodes (drivers/base/devres.c:506)
  devres_release_group (drivers/base/devres.c:669)
  i2c_device_remove (drivers/i2c/i2c-core-base.c:629)
  device_remove (drivers/base/dd.c:570)
  device_release_driver_internal (drivers/base/dd.c:1274 drivers/base/dd.c:1295)
  device_driver_detach (drivers/base/dd.c:1332)
  unbind_store (drivers/base/bus.c:247)
  ...
 ==================================================================

Reported-by: Vincent Whitchurch 
Fixes: 27a2195efa8d ("power: supply: core: auto-exposure of simple-battery data")
Tested-by: Vincent Whitchurch 
Signed-off-by: Sebastian Reichel

power: supply: rt9467: Fix rt9467_run_aicl()

2023-09-14T16:09:14+00:00

It is spurious to bail-out on a wait_for_completion_timeout() call that
does NOT timeout.

Reverse the logic to return -ETIMEDOUT instead, in case of tiemout.

Fixes: 6f7f70e3a8dd ("power: supply: rt9467: Add Richtek RT9467 charger driver")
Signed-off-by: Christophe JAILLET 
Reviewed-by: ChiYuan Huang 
Link: https://lore.kernel.org/r/2ed01020fa8a135c36dbaa871095ded47d926507.1676464968.git.christophe.jaillet@wanadoo.fr
Signed-off-by: Sebastian Reichel

power: supply: rk817: Add missing module alias

2023-09-14T15:47:36+00:00

Similar to the rk817 codec alias that was missing, the rk817 charger
driver is missing a module alias as well. This absence prevents the
driver from autoprobing on OF systems when it is built as a module.

Add the right MODULE_ALIAS to fix this.

Fixes: 11cb8da0189b ("power: supply: Add charger driver for Rockchip RK817")
Cc: stable@vger.kernel.org
Signed-off-by: Nicolas Frattaroli 
Reviewed-by: Chris Morgan 
Link: https://lore.kernel.org/r/20230612143651.959646-2-frattaroli.nicolas@gmail.com
Signed-off-by: Sebastian Reichel

power: supply: ucs1002: fix error code in ucs1002_get_property()

2023-09-14T15:38:02+00:00

This function is supposed to return 0 for success instead of returning
the val->intval.  This makes it the same as the other case statements
in this function.

Fixes: 81196e2e57fc ("power: supply: ucs1002: fix some health status issues")
Signed-off-by: Dan Carpenter 
Link: https://lore.kernel.org/r/687f64a4-4c6e-4536-8204-98ad1df934e5@moroto.mountain
Signed-off-by: Sebastian Reichel

power: supply: rt5033_charger: recognize EXTCON setting

2023-09-12T16:30:59+00:00

CHARGER_RT5033 should honor the EXTCON setting to prevent these
build errors:

riscv64-linux-ld: drivers/power/supply/rt5033_charger.o: in function `.L33':
rt5033_charger.c:(.text.rt5033_charger_probe+0x578): undefined reference to `extcon_find_edev_by_node'
riscv64-linux-ld: drivers/power/supply/rt5033_charger.o: in function `.L0 ':
rt5033_charger.c:(.text.rt5033_charger_probe+0x64e): undefined reference to `devm_extcon_register_notifier_all'
riscv64-linux-ld: drivers/power/supply/rt5033_charger.o: in function `.L96':
rt5033_charger.c:(.text.rt5033_charger_extcon_work+0x32): undefined reference to `extcon_get_state'

Fixes: 12cc585f36b8 ("power: supply: rt5033_charger: Add cable detection and USB OTG supply")
Signed-off-by: Randy Dunlap 
Cc: Jakob Hauser 
Cc: Sebastian Reichel 
Cc: Lee Jones 
Cc: linux-pm@vger.kernel.org
Link: https://lore.kernel.org/r/20230828224201.26823-1-rdunlap@infradead.org
Signed-off-by: Sebastian Reichel

power: supply: mt6370: Fix missing error code in mt6370_chg_toggle_cfo()

2023-09-11T18:15:43+00:00

When mt6370_chg_field_get() suceeds, ret is set to zero and returning
zero when flash led is still in strobe mode looks incorrect.

Fixes: 233cb8a47d65 ("power: supply: mt6370: Add MediaTek MT6370 charger driver")
Signed-off-by: Harshit Mogalapalli 
Reviewed-by: AngeloGioacchino Del Regno 
Reviewed-by: ChiaEn Wu 
Link: https://lore.kernel.org/r/20230906084815.2827930-1-harshit.m.mogalapalli@oracle.com
Signed-off-by: Sebastian Reichel

power: supply: ab8500: Set typing and props

2023-09-11T10:51:07+00:00

I had the following weird phenomena on a mobile phone: while
the capacity in /sys/class/power_supply/ab8500_fg/capacity
would reflect the actual charge and capacity of the battery,
only 1/3 of the value was shown on the battery status
indicator and warnings for low battery appeared.

It turns out that UPower, the Freedesktop power daemon,
will average all the power supplies of type "battery" in
/sys/class/power_supply/* if there is more than one battery.

For the AB8500, there was "battery" ab8500_fg, ab8500_btemp
and ab8500_chargalg. The latter two don't know anything
about the battery, and should not be considered. They were
however averaged and with the capacity of 0.

Flag ab8500_btemp and ab8500_chargalg with type "unknown"
so they are not averaged as batteries.

Remove the technology prop from ab8500_btemp as well, all
it does is snoop in on knowledge from another supply.

After this the battery indicator shows the right value.

Cc: Stefan Hansson 
Cc: stable@vger.kernel.org
Signed-off-by: Linus Walleij 
Signed-off-by: Sebastian Reichel