Re: [PATCH] rtc: zynqmp: Clear alarm interrupt status before interrupt enable

From: Michal Simek
Date: Mon Feb 10 2020 - 06:48:29 EST


Hi,

Ät 12. 12. 2019 v 14:01 odesÃlatel Srinivas Neeli
<srinivas.neeli@xxxxxxxxxx> napsal:
>
> Fix multiple occurring interrupts for alarm interrupt. RTC module doesn't
> clear the alarm interrupt status bit immediately after the interrupt is
> triggered.This is due to the sticky nature of the alarm interrupt status
> register. The alarm interrupt status register can be cleared only after
> the second counter outruns the set alarm value. To fix multiple spurious
> interrupts, disable alarm interrupt in the handler and clear the status
> bit before enabling the alarm interrupt.
>
> Fixes: 11143c19eb57 ("rtc: add xilinx zynqmp rtc driver")
> Signed-off-by: Srinivas Neeli <srinivas.neeli@xxxxxxxxxx>
> ---
> drivers/rtc/rtc-zynqmp.c | 29 ++++++++++++++++++++++++-----
> 1 file changed, 24 insertions(+), 5 deletions(-)
>
> diff --git a/drivers/rtc/rtc-zynqmp.c b/drivers/rtc/rtc-zynqmp.c
> index 5786866c09e9..d311e3ef1f21 100644
> --- a/drivers/rtc/rtc-zynqmp.c
> +++ b/drivers/rtc/rtc-zynqmp.c
> @@ -38,6 +38,8 @@
>
> #define RTC_CALIB_DEF 0x198233
> #define RTC_CALIB_MASK 0x1FFFFF
> +#define RTC_ALRM_MASK BIT(1)
> +#define RTC_MSEC 1000
>
> struct xlnx_rtc_dev {
> struct rtc_device *rtc;
> @@ -124,11 +126,28 @@ static int xlnx_rtc_alarm_irq_enable(struct device *dev, u32 enabled)
> {
> struct xlnx_rtc_dev *xrtcdev = dev_get_drvdata(dev);
>

here shouldn't be empty line.

> - if (enabled)
> + unsigned int status;
> + ulong timeout;
> +
> + timeout = jiffies + msecs_to_jiffies(RTC_MSEC);
> +
> + if (enabled) {
> + while (1) {
> + status = readl(xrtcdev->reg_base + RTC_INT_STS);
> + if (!((status & RTC_ALRM_MASK) == RTC_ALRM_MASK))
> + break;
> +
> + if (time_after_eq(jiffies, timeout)) {
> + dev_err(dev, "Time out occur, while clearing alarm status bit\n");
> + return -ETIMEDOUT;
> + }
> + writel(RTC_INT_ALRM, xrtcdev->reg_base + RTC_INT_STS);
> + }
> +
> writel(RTC_INT_ALRM, xrtcdev->reg_base + RTC_INT_EN);
> - else
> + } else {
> writel(RTC_INT_ALRM, xrtcdev->reg_base + RTC_INT_DIS);
> -
> + }

And here it was good to have empty line.

> return 0;
> }
>
> @@ -183,8 +202,8 @@ static irqreturn_t xlnx_rtc_interrupt(int irq, void *id)
> if (!(status & (RTC_INT_SEC | RTC_INT_ALRM)))
> return IRQ_NONE;
>
> - /* Clear RTC_INT_ALRM interrupt only */
> - writel(RTC_INT_ALRM, xrtcdev->reg_base + RTC_INT_STS);
> + /* Disable RTC_INT_ALRM interrupt only */
> + writel(RTC_INT_ALRM, xrtcdev->reg_base + RTC_INT_DIS);
>
> if (status & RTC_INT_ALRM)
> rtc_update_irq(xrtcdev->rtc, 1, RTC_IRQF | RTC_AF);
> --
> 2.7.4

Other then these two above things look good.

Alexandre: Any issue with this patch?

Thanks,
Michal

--
Michal Simek, Ing. (M.Eng), OpenPGP -> KeyID: FE3D1F91
w: www.monstr.eu p: +42-0-721842854
Maintainer of Linux kernel - Xilinx Microblaze
Maintainer of Linux kernel - Xilinx Zynq ARM and ZynqMP ARM64 SoCs
U-Boot custodian - Xilinx Microblaze/Zynq/ZynqMP/Versal SoCs