On Wed, Dec 29, 2021 at 08:44:56AM -0300, Fabio Estevam wrote:
> From: Fabio Estevam <festevam@xxxxxxx>
>
> Currently, the reset_gpio polarity handling is done backwards.
>
> The gpiod API takes the logic value of the GPIO, not the physical one.
>
> As per the CS4265 datasheet:
>
> "When RESET is low, the CS4265 enters a low-power mode and all
> internal states are reset"
>
> If a devicetree describes reset_gpio as active-low, the correct behaviour
> would be to retrieve the GPIO and put in its active state (GPIOD_OUT_HIGH)
> and then move it to its inactive state so that it can be operational
> (logic level 0).
>
> Fix it accordingly.
>
> Fixes: fb6f806967f6 ("ASoC: Add support for the CS4265 CODEC")
> Signed-off-by: Fabio Estevam <festevam@xxxxxxx>
> ---
> Changes since v1:
> - Newly introduced patch.
>
> sound/soc/codecs/cs4265.c | 4 ++--
> 1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/sound/soc/codecs/cs4265.c b/sound/soc/codecs/cs4265.c
> index b89002189a2b..294fa7ac16cb 100644
> --- a/sound/soc/codecs/cs4265.c
> +++ b/sound/soc/codecs/cs4265.c
> @@ -590,13 +590,13 @@ static int cs4265_i2c_probe(struct i2c_client *i2c_client,
> }
>
> cs4265->reset_gpio = devm_gpiod_get_optional(&i2c_client->dev,
> - "reset", GPIOD_OUT_LOW);
> + "reset", GPIOD_OUT_HIGH);
> if (IS_ERR(cs4265->reset_gpio))
> return PTR_ERR(cs4265->reset_gpio);
>
> if (cs4265->reset_gpio) {
> mdelay(1);
> - gpiod_set_value_cansleep(cs4265->reset_gpio, 1);
> + gpiod_set_value_cansleep(cs4265->reset_gpio, 0);
Hmm... I might defer to Mark on this one. I totally agree your
new code is more correct, however, I would have a slight worry
about existing device trees not correctly specifying the GPIO. As
in if existing systems had been specifying the GPIO correctly
they are presumably currently broken. But I am not sure what the
obvious solution would be, and I don't really have a good feel
for how widely used this part is.
Thanks,
Charles
