rtc: ab-eoz9: fix abeoz9_rtc_read_alarm

abeoz9_rtc_read_alarm assumes we always read the alarm in 12-hour mode
while abeoz9_rtc_set_alarm will always set it in 24-hour mode.

We could support 12-hour mode in both functions but it seems very unlikely
that the RTC would be set to 12-hour mode now as the driver has been
setting it to 24-hour mode for a while now. The setting is undefined at
power-up and unchanged by subsequent resets which doesn't help us.

Link: https://lore.kernel.org/r/[e-mail]
Signed-off-by: Alexandre Belloni <[e-mail]>

drivers/rtc/rtc-ab-eoz9.c

···
 #define ABEOZ9_BIT_ALARM_MIN		GENMASK(6, 0)
 #define ABEOZ9_REG_ALARM_HOURS		0x12
 #define ABEOZ9_BIT_ALARM_HOURS_PM	BIT(5)
-#define ABEOZ9_BIT_ALARM_HOURS		GENMASK(4, 0)
+#define ABEOZ9_BIT_ALARM_HOURS		GENMASK(5, 0)
 #define ABEOZ9_REG_ALARM_DAYS		0x13
 #define ABEOZ9_BIT_ALARM_DAYS		GENMASK(5, 0)
 #define ABEOZ9_REG_ALARM_WEEKDAYS	0x14
···
 	alarm->time.tm_sec = bcd2bin(FIELD_GET(ABEOZ9_BIT_ALARM_SEC, regs[0]));
 	alarm->time.tm_min = bcd2bin(FIELD_GET(ABEOZ9_BIT_ALARM_MIN, regs[1]));
 	alarm->time.tm_hour = bcd2bin(FIELD_GET(ABEOZ9_BIT_ALARM_HOURS, regs[2]));
-	if (FIELD_GET(ABEOZ9_BIT_ALARM_HOURS_PM, regs[2]))
-		alarm->time.tm_hour += 12;
 
 	alarm->time.tm_mday = bcd2bin(FIELD_GET(ABEOZ9_BIT_ALARM_DAYS, regs[3]));