On Thu, 4 Jan 2024 at 10:43, Wenhua Lin <Wenhua.Lin@xxxxxxxxxx> wrote: > > The eic debounce does not have a clock of rtc_1k in the sleep state, > but the eic debounce will be used to wake up the system, therefore the > clock of rtc_1k needs to be kept open. It seems that this issue is not in the latest SoCs. I would suggest not changing for the time being. Thanks, Chunyan > > Signed-off-by: Wenhua Lin <Wenhua.Lin@xxxxxxxxxx> > --- > drivers/gpio/gpio-eic-sprd.c | 2 ++ > 1 file changed, 2 insertions(+) > > diff --git a/drivers/gpio/gpio-eic-sprd.c b/drivers/gpio/gpio-eic-sprd.c > index be7f2fa5aa7b..bdcb3510a208 100644 > --- a/drivers/gpio/gpio-eic-sprd.c > +++ b/drivers/gpio/gpio-eic-sprd.c > @@ -24,6 +24,7 @@ > #define SPRD_EIC_DBNC_IC 0x24 > #define SPRD_EIC_DBNC_TRIG 0x28 > #define SPRD_EIC_DBNC_CTRL0 0x40 > +#define SPRD_EIC_DBNC_FORCE_CLK 0x8000 > > #define SPRD_EIC_LATCH_INTEN 0x0 > #define SPRD_EIC_LATCH_INTRAW 0x4 > @@ -223,6 +224,7 @@ static int sprd_eic_set_debounce(struct gpio_chip *chip, unsigned int offset, > u32 value = readl_relaxed(base + reg) & ~SPRD_EIC_DBNC_MASK; > > value |= (debounce / 1000) & SPRD_EIC_DBNC_MASK; > + value |= SPRD_EIC_DBNC_FORCE_CLK; > writel_relaxed(value, base + reg); > > return 0; > -- > 2.17.1 >