Vybrid Standby Mode Wakeup

I am using following script to test VF61 standby suspend mode with SNVS RTC as wakeup source on Colibri Eva Board V3.2.

#!/bin/sh

n=0

while true;do

echo  +3600  > /sys/class/rtc/rtc0/wakealarm

while [ $? -ne 0 ]
do
sleep 5
echo retry to write wakealarm
echo  0  > /sys/class/rtc/rtc0/wakealarm
echo  +3600  > /sys/class/rtc/rtc0/wakealarm

done

#echo mem  > /sys/power/state
echo "before suspend"
timedatectl
sleep 10
echo "enter into standy mode\n"
temp1="$(cat /sys/class/rtc/rtc0/wakealarm)"
temp2="$(cat /sys/class/rtc/rtc0/since_epoch)"
echo -e "wakealarm: $temp1\t since_epoch: $temp2\t\n" 

echo standby > /sys/power/state


n=`expr $n + 1`
echo "################# Wake up $n ######################"
#date "+%Y-%m-%d %H:%M:%S"
echo "wake from suspend"
timedatectl
sleep 1

done

Only internal RTC is used.

root@colibri-vf:~# dmesg|grep rtc
[    1.301121] rtc-ds1307: probe of 0-0068 failed with error -5
[    1.315359] snvs_rtc 400a7000.snvs:snvs-rtc-lp: rtc core: registered 400a7000.snvs:snvs- as rtc0
[    2.590597] snvs_rtc 400a7000.snvs:snvs-rtc-lp: setting system clock to 2018-02-01 14:26:04 UTC (1517495164)

NTP is also enabled. Suspend gap is 3600 seconds. Sometimes, it is not waken up by snvs-rtc. As the link textlog shows, 8th wakeup should be activated in one hour later(wakealarm is 1517556813). VF61 is waken up manually by GPIO after two and half hours later, since_epoch value already beyond 1517556813.

Hi
During Several Tests it was seen that this issue can occur 1 over 100 tests. How often did the error occur in your case?

Hi Jaski, I have seen it once over 8 tests with above script. In the reality, this customer will put Vybrid into standby mode for even longer period that is 12 hours. They report it will definitely happen in 2 to 3 days. It really takes time to reproduce such kind of issue.

I test again, by now 19 trials have been successful. It is 19 hours. I don’t see the issue.

HIi Jask, do you have any update about this issue? Thanks very much!