On Wed, Jul 09, 2025 at 05:01:38PM +0200, Michael Walle wrote:

Hi Ioana,

great to see another user of gpio-regmap.

On Wed Jul 9, 2025 at 1:26 PM CEST, Ioana Ciornei wrote:

quoted

There are GPIO controllers such as the one present in the LX2160ARDB
QIXIS CPLD which are single register fixed-direction. This cannot be
modeled using the gpio-regmap as-is since there is no way to
present the true direction of a GPIO line.

You mean input and output mixed together in one register? At least
to me, that wasn't so obvious by the commit message, I had to look
at the actual driver.

Yes, that is right. I will update the commit message to make it more
clear for everybody.

quoted

In order to make this use case possible, add a new callback to the
gpio_config structure - .get_direction() - which can be used by user
drivers to provide the fixed direction per GPIO line.

Signed-off-by: Ioana Ciornei <ioana.ciornei@nxp.com>
---
 drivers/gpio/gpio-regmap.c  | 17 ++++++++++++++++-
 include/linux/gpio/regmap.h |  3 +++
 2 files changed, 19 insertions(+), 1 deletion(-)

diff --git a/drivers/gpio/gpio-regmap.c b/drivers/gpio/gpio-regmap.c
index 87c4225784cf..dac2acb26655 100644
--- a/drivers/gpio/gpio-regmap.c
+++ b/drivers/gpio/gpio-regmap.c

@@ -32,6 +32,8 @@ struct gpio_regmap {
 	unsigned int reg_dir_in_base;
 	unsigned int reg_dir_out_base;
 
+	int (*get_direction)(struct gpio_regmap *gpio, unsigned int offset);
+
 	int (*reg_mask_xlate)(struct gpio_regmap *gpio, unsigned int base,
 			      unsigned int offset, unsigned int *reg,
 			      unsigned int *mask);

@@ -129,6 +131,9 @@ static int gpio_regmap_get_direction(struct gpio_chip *chip,
 	unsigned int base, val, reg, mask;
 	int invert, ret;
 
+	if (gpio->get_direction)
+		return gpio->get_direction(gpio, offset);
+
 	if (gpio->reg_dat_base && !gpio->reg_set_base)
 		return GPIO_LINE_DIRECTION_IN;
 	if (gpio->reg_set_base && !gpio->reg_dat_base)

@@ -163,7 +168,16 @@ static int gpio_regmap_set_direction(struct gpio_chip *chip,
 {
 	struct gpio_regmap *gpio = gpiochip_get_data(chip);
 	unsigned int base, val, reg, mask;
-	int invert, ret;
+	int invert, ret, dir;
+
+	if (gpio->get_direction) {
+		dir = gpio->get_direction(gpio, offset);
+		if (dir == GPIO_LINE_DIRECTION_IN && output)
+			return -EOPNOTSUPP;
+		if (dir == GPIO_LINE_DIRECTION_OUT && !output)
+			return -EOPNOTSUPP;
+		return 0;
+	}

What is the intention here? Shouldn't there be just a .set_direction
op and if there isn't one, return EOPNOTSUPP?

In any case, that is unused code for your driver as far as I see,
because you neither set .reg_dir_in_base nor .reg_dir_out_base and
thus, .direction_input nor .direction_output are set within the
gpio_chip struct (see gpio_regmap_register()).

Yes, you are right. I did want to return ealier an EOPNOTSUPP to make
sure that in my specific case no directions would be changed. But that
is not needed indeed since, as you said, I do not use .reg_dir_in_base
nor .reg_dir_out_base.

I will remove this in v2.

And since we are on the subject, I will try out Andrew's suggestion with
giving gpio-regmap a bitmap to use directly.

Thanks!
Ioana