Re: [PATCH] gpio: rockchip: mark the GPIO controller as sleeping
From: Heiko Stübner <heiko@sntech.de>
Date: 2026-01-25 13:23:38
Also in:
linux-gpio, linux-rockchip, lkml, stable
Am Samstag, 24. Januar 2026, 22:07:12 Mitteleuropäische Normalzeit schrieb Bartosz Golaszewski:
On Sat, 24 Jan 2026 00:45:25 +0100, Robin Murphy [off-list ref] said:quoted
On 2026-01-23 9:52 pm, Heiko Stübner wrote:
[ snip to get to the core question at the bottom ]
quoted
diff --git a/drivers/pinctrl/pinctrl-rockchip.c b/drivers/pinctrl/pinctrl-rockchip.c index e44ef262beec..2fc67aeafdb3 100644 --- a/drivers/pinctrl/pinctrl-rockchip.c +++ b/drivers/pinctrl/pinctrl-rockchip.c@@ -3545,10 +3545,9 @@ static int rockchip_pmx_set(struct pinctrl_dev *pctldev, unsigned selector, return 0; } -static int rockchip_pmx_gpio_set_direction(struct pinctrl_dev *pctldev, - struct pinctrl_gpio_range *range, - unsigned offset, - bool input) +static int rockchip_pmx_gpio_request_enable(struct pinctrl_dev *pctldev, + struct pinctrl_gpio_range *range, + unsigned int offset) { struct rockchip_pinctrl *info = pinctrl_dev_get_drvdata(pctldev); struct rockchip_pin_bank *bank;@@ -3562,7 +3561,7 @@ static const struct pinmux_ops rockchip_pmx_ops = { .get_function_name = rockchip_pmx_get_func_name, .get_function_groups = rockchip_pmx_get_groups, .set_mux = rockchip_pmx_set, - .gpio_set_direction = rockchip_pmx_gpio_set_direction, + .gpio_request_enable = rockchip_pmx_gpio_request_enable, }; /*I'm not sure what's going on here. You don't really need to call pinctrl_gpio_direction_input/output()?
No we actually don't.
The current _set_direction function:
static int rockchip_pmx_gpio_set_direction(struct pinctrl_dev *pctldev,
struct pinctrl_gpio_range *range,
unsigned offset,
bool input)
{
struct rockchip_pinctrl *info = pinctrl_dev_get_drvdata(pctldev);
struct rockchip_pin_bank *bank;
bank = pin_to_bank(info, offset);
return rockchip_set_mux(bank, offset - bank->pin_base, RK_FUNC_GPIO);
}
really only ever touches the pinmux and has nothing to do with setting
the direction, which is solely the task of the gpio-controller/-driver.
So moving that to the request callback would likely solve all our current
problems?
Heiko