On Wed, Jul 29, 2020 at 5:35 PM Chunyan Zhang <zhang.l...@gmail.com> wrote:
>
> From: Taiping Lai <taiping....@unisoc.com>
>
> The raw interrupt status of GPIO maybe set before the interrupt is enabled,
> which would trigger the interrupt event once enabled it from user side.
> This is the case for edge interrupts only. Adding a clear operation when
> setting interrupt type can avoid that.
>
> Fixes: 9a3821c2bb47 ("gpio: Add GPIO driver for Spreadtrum SC9860 platform")
> Signed-off-by: Taiping Lai <taiping....@unisoc.com>
> Signed-off-by: Chunyan Zhang <chunyan.zh...@unisoc.com>
> ---
> drivers/gpio/gpio-sprd.c | 3 +++
> 1 file changed, 3 insertions(+)
>
> diff --git a/drivers/gpio/gpio-sprd.c b/drivers/gpio/gpio-sprd.c
> index d7314d39ab65..36ea8a3bd451 100644
> --- a/drivers/gpio/gpio-sprd.c
> +++ b/drivers/gpio/gpio-sprd.c
> @@ -149,17 +149,20 @@ static int sprd_gpio_irq_set_type(struct irq_data *data,
> sprd_gpio_update(chip, offset, SPRD_GPIO_IS, 0);
> sprd_gpio_update(chip, offset, SPRD_GPIO_IBE, 0);
> sprd_gpio_update(chip, offset, SPRD_GPIO_IEV, 1);
> + sprd_gpio_update(chip, offset, SPRD_GPIO_IC, 1);
I think you should move this abonormal interrupt clearing operation to
sprd_gpio_request(), when users request a GPIO.
> irq_set_handler_locked(data, handle_edge_irq);
> break;
> case IRQ_TYPE_EDGE_FALLING:
> sprd_gpio_update(chip, offset, SPRD_GPIO_IS, 0);
> sprd_gpio_update(chip, offset, SPRD_GPIO_IBE, 0);
> sprd_gpio_update(chip, offset, SPRD_GPIO_IEV, 0);
> + sprd_gpio_update(chip, offset, SPRD_GPIO_IC, 1);
> irq_set_handler_locked(data, handle_edge_irq);
> break;
> case IRQ_TYPE_EDGE_BOTH:
> sprd_gpio_update(chip, offset, SPRD_GPIO_IS, 0);
> sprd_gpio_update(chip, offset, SPRD_GPIO_IBE, 1);
> + sprd_gpio_update(chip, offset, SPRD_GPIO_IC, 1);
> irq_set_handler_locked(data, handle_edge_irq);
> break;
> case IRQ_TYPE_LEVEL_HIGH:
> --
> 2.20.1
>
--
Baolin Wang
