4.19-stable review patch. If anyone has any objections, please let me know.
------------------
From: Nuno Sa <nuno.sa@xxxxxxxxxx>
commit c684771630e64bc39bddffeb65dd8a6612a6b249 upstream.
The adp5589 seems to have the same behavior as similar devices as
explained in commit 910a9f5636f5 ("Input: adp5588-keys - get value from
data out when dir is out").
Basically, when the gpio is set as output we need to get the value from
ADP5589_GPO_DATA_OUT_A register instead of ADP5589_GPI_STATUS_A.
Fixes: 9d2e173644bb ("Input: ADP5589 - new driver for I2C Keypad Decoder and I/O Expander")
Signed-off-by: Nuno Sa <nuno.sa@xxxxxxxxxx>
Link: https://lore.kernel.org/r/20241001-b4-dev-adp5589-fw-conversion-v1-2-fca0149dfc47@xxxxxxxxxx
Cc: stable@xxxxxxxxxxxxxxx
Signed-off-by: Dmitry Torokhov <dmitry.torokhov@xxxxxxxxx>
Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>
---
drivers/input/keyboard/adp5589-keys.c | 13 ++++++++++---
1 file changed, 10 insertions(+), 3 deletions(-)
--- a/drivers/input/keyboard/adp5589-keys.c
+++ b/drivers/input/keyboard/adp5589-keys.c
@@ -390,10 +390,17 @@ static int adp5589_gpio_get_value(struct
struct adp5589_kpad *kpad = gpiochip_get_data(chip);
unsigned int bank = kpad->var->bank(kpad->gpiomap[off]);
unsigned int bit = kpad->var->bit(kpad->gpiomap[off]);
+ int val;
- return !!(adp5589_read(kpad->client,
- kpad->var->reg(ADP5589_GPI_STATUS_A) + bank) &
- bit);
+ mutex_lock(&kpad->gpio_lock);
+ if (kpad->dir[bank] & bit)
+ val = kpad->dat_out[bank];
+ else
+ val = adp5589_read(kpad->client,
+ kpad->var->reg(ADP5589_GPI_STATUS_A) + bank);
+ mutex_unlock(&kpad->gpio_lock);
+
+ return !!(val & bit);
}
static void adp5589_gpio_set_value(struct gpio_chip *chip,
