pinctrl: qcom: spmi-gpio: make direction changes exclusive

Neither direction callback clears the opposite buffer: .direction_output()
only packs PIN_CONFIG_LEVEL and .direction_input() only packs
PIN_CONFIG_INPUT_ENABLE.  As pmic_gpio_populate() seeds input_enabled and
output_enabled from MODE_CTL, a pad left in DIGITAL_INPUT[_OUTPUT] by the
bootloader ends up in DIGITAL_INPUT_OUTPUT whichever direction is
requested.  That leaves an input still driving the line, and makes
pmic_gpio_get_direction() report an output as an input, which breaks
consumers that read the direction back:

  reg-fixed-voltage regulator-wcn-core-vm-1p35: setup of GPIO (default) failed: -1
  reg-fixed-voltage regulator-wcn-core-vm-1p35: error -EPERM: can't get GPIO

Pack the opposite buffer's PIN_CONFIG_*_ENABLE along with the requested
direction, so MODE_CTL becomes DIGITAL_INPUT or DIGITAL_OUTPUT, never
both.  pmic_gpio_config_set() writes the registers once after walking all
configs, so this stays a single write.

An open-drain or open-source output is the exception: it only drives one
rail, so pmic_gpio_get() needs the input buffer to sample the line rather
than return the value last written.  Program the input buffer from
pad->buffer_type, which gpiolib has already updated via
PIN_CONFIG_DRIVE_OPEN_DRAIN, and key pmic_gpio_get_direction() off
output_enabled so such pads still read back as outputs.

Genuinely bidirectional pads can still be described that way through
pinconf; the gpiolib callbacks now mean what gpiolib says they mean.

Fixes: 2634475324 ("pinctrl: qcom: spmi-gpio: implement .get_direction()")
Assisted-by: LLM
Reviewed-by: Linus Walleij <linusw@kernel.org>
Signed-off-by: Shawn Guo <shengchao.guo@oss.qualcomm.com>
Link: https://patch.msgid.link/20260927010913.1212600-1-shengchao.guo@oss.qualcomm.com
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@oss.qualcomm.com>
This commit is contained in:
Shawn Guo 2026-09-27 09:09:13 +08:00 • committed by Bartosz Golaszewski
commit 19fc424035