Re: [PATCH v2 06/11] drm/panel: s6e3ha8: Correct the polarity logic within

[Krzysztof Kozlowski <krzk@kernel.org>]

On 29/09/2026 11:41, David Heidelberg wrote:
> On 29/09/2026 09:45, Krzysztof Kozlowski wrote:
>> On Thu, Sep 24, 2026 at 04:01:34PM +0200, David Heidelberg wrote:
>>> The reset was introduced with wrong polarity. Correct for the future
>>> compatibles and keep current with reverted logic.
>>>
>>> Old DTs keep GPIO_ACTIVE_HIGH and are fixed up via
>>> gpiod_toggle_active_low() on the deprecated compatible.
>>>
>>> Assisted-by: LLM
>>> Reviewed-by: Neil Armstrong <neil.armstrong@linaro.org>
>>> Signed-off-by: David Heidelberg <david@ixit.cz>
>>> ---
>>>   drivers/gpu/drm/panel/panel-samsung-s6e3ha8.c | 23 +++++++++++++++++++----
>>>   1 file changed, 19 insertions(+), 4 deletions(-)
>>>
>>> diff --git a/drivers/gpu/drm/panel/panel-samsung-s6e3ha8.c b/drivers/gpu/drm/panel/panel-samsung-s6e3ha8.c
>>> index 5e1e997b83b36..99290913de69a 100644
>>> --- a/drivers/gpu/drm/panel/panel-samsung-s6e3ha8.c
>>> +++ b/drivers/gpu/drm/panel/panel-samsung-s6e3ha8.c
>>> @@ -20,16 +20,17 @@
>>>   #include "panel-samsung-dsi.h"
>>>   
>>>   struct s6e3ha8_desc {
>>>   	const struct drm_panel_funcs *funcs;
>>>   	const struct drm_display_mode *mode;
>>>   	unsigned long mode_flags;
>>>   	const struct regulator_bulk_data *supplies;
>>>   	unsigned int num_supplies;
>>> +	bool broken_reset_polarity;
>>>   };
>>>   
>>>   struct s6e3ha8 {
>>>   	struct drm_panel panel;
>>>   	struct mipi_dsi_device *dsi;
>>>   	const struct s6e3ha8_desc *desc;
>>>   	struct drm_dsc_config dsc;
>>>   	struct gpio_desc *reset_gpio;
>>> @@ -62,22 +63,22 @@ static int s6e3ha8_unprepare(struct drm_panel *panel)
>>>   {
>>>   	struct s6e3ha8 *priv = to_s6e3ha8(panel);
>>>   
>>>   	return regulator_bulk_disable(priv->desc->num_supplies, priv->supplies);
>>>   }
>>>   
>>>   static void s6e3ha8_amb577px01_wqhd_reset(struct s6e3ha8 *priv)
>>>   {
>>> -	gpiod_set_value_cansleep(priv->reset_gpio, 1);
>>> -	usleep_range(5000, 6000);
>>>   	gpiod_set_value_cansleep(priv->reset_gpio, 0);
>>>   	usleep_range(5000, 6000);
>>>   	gpiod_set_value_cansleep(priv->reset_gpio, 1);
>>>   	usleep_range(5000, 6000);
>>> +	gpiod_set_value_cansleep(priv->reset_gpio, 0);
>>
>> This breaks all users and this usage of ABI was already released.
> 
> See the gpiod_toggle_active_low() usage later in the patch which keep the logic 
> for the original compatible as intended.
> 

OK, I went way too fast, that's correct part. But splitting fix is still
just confusing. Backporting to stable is a different thing than fixing
issues.

Best regards,
Krzysztof