Re: [PATCH v4 2/7] pinctrl: ingenic: Fix hardware dual-edge interrupts on X1600
From: Paul Cercueil
Date: Thu Nov 19 2026 - 04:34:14 EST
Hi Nikolaus,
Le mercredi 18 novembre 2026 à 18:25 +0100, H. Nikolaus Schaller a écrit :
> Use the hardware dual-edge detector on X1600 and X2000.
>
> Set PAT1 when configuring both edges and program EDG on X1600.
> Keep software edge switching for older SoCs without EDG.
>
> Tested on X1600 and X2000 using GPIO buttons and register readback.
>
> Fixes: 4da56f9c6f14 ("pinctrl: ingenic: add x1600 support")
> Cc: stable@xxxxxxxxxxxxxxx
> Signed-off-by: H. Nikolaus Schaller <hns@xxxxxxxxxxxxx>
Reviewed-by: Paul Cercueil <paul@xxxxxxxxxxxxxxx>
Cheers,
-Paul
> ---
>
> Notes:
> Changes V4:
> - Fix Fixes: and stable Cc trailers.
> (suggested by Paul Cercueil and Sashiko-Bot)
> - Merge the common shadow-register paths and simplify EDG handling.
> (suggested by Paul Cercueil)
> - Retested hardware dual-edge interrupts on X1600/X2000 after the
> shadow-register refactoring. Verified both button transitions
> with evtest and confirmed via devmem2 that PAT1, PAT0 and EDG
> remain unchanged across GPIO input transitions.
>
> drivers/pinctrl/pinctrl-ingenic.c | 17 ++++++++---------
> 1 file changed, 8 insertions(+), 9 deletions(-)
>
> diff --git a/drivers/pinctrl/pinctrl-ingenic.c b/drivers/pinctrl/pinctrl-
> ingenic.c
> index 88c5edeaaddcc..7f65f3abf05f5 100644
> --- a/drivers/pinctrl/pinctrl-ingenic.c
> +++ b/drivers/pinctrl/pinctrl-ingenic.c
> @@ -3601,7 +3601,8 @@ static void irq_set_type(struct ingenic_gpio_chip *jzgc,
>
> switch (type) {
> case IRQ_TYPE_EDGE_BOTH:
> - val1 = val2 = false;
> + val1 = false;
> + val2 = is_soc_or_above(jzgc->jzpc, ID_X1600);
> val3 = true;
> break;
> case IRQ_TYPE_EDGE_RISING:
> @@ -3635,15 +3636,13 @@ static void irq_set_type(struct ingenic_gpio_chip
> *jzgc,
> return;
> }
>
> - if (is_soc_or_above(jzgc->jzpc, ID_X2000)) {
> - ingenic_gpio_shadow_set_bit(jzgc, reg2, offset, val1);
> - ingenic_gpio_shadow_set_bit(jzgc, reg1, offset, val2);
> - ingenic_gpio_shadow_set_bit_load(jzgc);
> - ingenic_gpio_set_bit(jzgc, X2000_GPIO_EDG, offset, val3);
> - } else if (is_soc_or_above(jzgc->jzpc, ID_X1000)) {
> + if (is_soc_or_above(jzgc->jzpc, ID_X1000)) {
> ingenic_gpio_shadow_set_bit(jzgc, reg2, offset, val1);
> ingenic_gpio_shadow_set_bit(jzgc, reg1, offset, val2);
> ingenic_gpio_shadow_set_bit_load(jzgc);
> +
> + if (is_soc_or_above(jzgc->jzpc, ID_X1600))
> + ingenic_gpio_set_bit(jzgc, X2000_GPIO_EDG, offset,
> val3);
> } else {
> ingenic_gpio_set_bit(jzgc, reg2, offset, val1);
> ingenic_gpio_set_bit(jzgc, reg1, offset, val2);
> @@ -3718,7 +3717,7 @@ static void ingenic_gpio_irq_ack(struct irq_data *irqd)
> bool high;
>
> if ((irqd_get_trigger_type(irqd) == IRQ_TYPE_EDGE_BOTH) &&
> - !is_soc_or_above(jzgc->jzpc, ID_X2000)) {
> + !is_soc_or_above(jzgc->jzpc, ID_X1600)) {
> /*
> * Switch to an interrupt for the opposite edge to the one
> that
> * triggered the interrupt being ACKed.
> @@ -3758,7 +3757,7 @@ static int ingenic_gpio_irq_set_type(struct irq_data
> *irqd, unsigned int type)
> irq_set_handler_locked(irqd, handle_bad_irq);
> }
>
> - if ((type == IRQ_TYPE_EDGE_BOTH) && !is_soc_or_above(jzgc->jzpc,
> ID_X2000)) {
> + if (type == IRQ_TYPE_EDGE_BOTH && !is_soc_or_above(jzgc->jzpc,
> ID_X1600)) {
> /*
> * The hardware does not support interrupts on both edges.
> The
> * best we can do is to set up a single-edge interrupt and
> then