Re: [PATCH v4 03/15] rtc: rzn1: Fix weekday underflow when alarm crosses month boundary

On Wed, Aug 19, 2026 at 2:01 PM Wolfram Sang
<[email protected]> wrote:
>
>
> > [Severity: High]
> > This is a pre-existing issue, but does this time_after() comparison
> > allow bypassing the 1-week hardware limit on 32-bit systems?
>
> This is fixed later.
>
> > [Severity: Medium]
> > This is a pre-existing issue, but what happens if the RZN1_RTC_ALW register
> > is uninitialized and wday is 0?
> >
> > Calling fls(0) - 1 would evaluate to -1. If the current time is a Saturday
> > (tm_now.tm_wday == 6), delta_days evaluates to 0, which bypasses
> > the if (delta_days) block and leaves tm->tm_wday = -1.
> >
> > Since the core function rtc_valid_tm() does not check tm_wday, would this
> > propagate a bogus alarm date to the caller?
>
> This looks worth investigating to me. On RZN1, ALW is default 0.
>
On T2H/N2H too.

I'll create a below patch for it:
diff --git a/drivers/rtc/rtc-rzn1.c b/drivers/rtc/rtc-rzn1.c
index 8c70dbf8c7ec..e0fdb591ba91 100644
--- a/drivers/rtc/rtc-rzn1.c
+++ b/drivers/rtc/rtc-rzn1.c
@@ -234,13 +234,24 @@ static int rzn1_rtc_read_alarm(struct device
*dev, struct rtc_wkalrm *alrm)
        if (ret)
                return ret;

+       ctl1 = readl(rtc->base + RZN1_RTC_CTL1);
+       alrm->enabled = !!(ctl1 & (RZN1_RTC_CTL1_ALME | RZN1_RTC_CTL1_1SE));
+
        min = readl(rtc->base + RZN1_RTC_ALM);
        hour = readl(rtc->base + RZN1_RTC_ALH);
-       wday = readl(rtc->base + RZN1_RTC_ALW);

        tm->tm_sec = 0;
        tm->tm_min = bcd2bin(min);
        tm->tm_hour = bcd2bin(hour);
+
+       /*
+        * If wday is zero, no bit is set in RZN1_RTC_ALW. This is the
+        * register's power-on reset value.
+        */
+       wday = readl(rtc->base + RZN1_RTC_ALW);
+       if (!wday)
+               return 0;
+
        delta_days = ((fls(wday) - 1) - tm->tm_wday + 7) % 7;
        tm->tm_wday = fls(wday) - 1;

@@ -249,9 +260,6 @@ static int rzn1_rtc_read_alarm(struct device *dev,
struct rtc_wkalrm *alrm)
                rtc_time64_to_tm(alarm, tm);
        }

-       ctl1 = readl(rtc->base + RZN1_RTC_CTL1);
-       alrm->enabled = !!(ctl1 & (RZN1_RTC_CTL1_ALME | RZN1_RTC_CTL1_1SE));
-
        return 0;
 }

Logs:
    On a cold-booted board with no alarm configured, the following was
    observed before the fix:

        root@rzn2h-evk:~# cat /proc/driver/rtc
        rtc_time        : 00:00:19
        rtc_date        : 2000-01-01
        alrm_time       : 00:00:00
        alrm_date       : 2000-01-07
        alarm_IRQ       : no
        alrm_pending    : no
        update IRQ enabled      : no
        periodic IRQ enabled    : no
        periodic IRQ frequency  : 1
        max user IRQ frequency  : 64
        24hr            : yes

    After the fix:

        root@rzn2h-evk:~# cat /proc/driver/rtc
        rtc_time        : 00:00:25
        rtc_date        : 2000-01-01
        alrm_time       : 00:00:00
        alrm_date       : 2000-01-01
        alarm_IRQ       : no
        alrm_pending    : no
        update IRQ enabled      : no
        periodic IRQ enabled    : no
        periodic IRQ frequency  : 1
        max user IRQ frequency  : 64
        24hr            : yes


Cheers,
Prabhakar