On Fri, May 19, 2023 at 7:17 PM Lizhe <sensor1010@xxxxxxx> wrote: > > Before executing microchip_sgpio_irq_set_type(), > type has already been cleared IRQ_TYPE_SENSE_MASK, see __irq_set_trigger(). > > Signed-off-by: Lizhe <sensor1010@xxxxxxx> > --- > arch/arm/plat-orion/gpio.c | 1 - > drivers/gpio/gpio-mvebu.c | 1 - > 2 files changed, 2 deletions(-) > > diff --git a/arch/arm/plat-orion/gpio.c b/arch/arm/plat-orion/gpio.c > index 595e9cb33c1d..863fa497b1a2 100644 > --- a/arch/arm/plat-orion/gpio.c > +++ b/arch/arm/plat-orion/gpio.c > @@ -364,7 +364,6 @@ static int gpio_irq_set_type(struct irq_data *d, u32 type) > return -EINVAL; > } > > - type &= IRQ_TYPE_SENSE_MASK; > if (type == IRQ_TYPE_NONE) > return -EINVAL; > > diff --git a/drivers/gpio/gpio-mvebu.c b/drivers/gpio/gpio-mvebu.c > index a68f682aec01..34fd007b0308 100644 > --- a/drivers/gpio/gpio-mvebu.c > +++ b/drivers/gpio/gpio-mvebu.c > @@ -505,7 +505,6 @@ static int mvebu_gpio_irq_set_type(struct irq_data *d, unsigned int type) > if ((u & BIT(pin)) == 0) > return -EINVAL; > > - type &= IRQ_TYPE_SENSE_MASK; > if (type == IRQ_TYPE_NONE) > return -EINVAL; > > -- > 2.34.1 > Please split it into two patches. I can apply the second part but arch/arm is beyond my jurisdiction. While at it: is this platform even used at all? If so, then maybe we could migrate this driver to drivers/gpio/? Bart