Unstable MTCDP-H5

[Herraiz Stéphane]

Some our devices (SN starts at 18259xxx) are unstable.
Before freezing we get on the serial debug console this kind of message :
Unable to handle kernel NULL pointer dereference at virtual address 00000000
pgd = c3328000
[00000000] *pgd=23f44031, *pte=00000000, *ppte=00000000
Internal error: Oops: 817 [#66] PREEMPT
last sysfs file: /sys/module/ohci_hcd/initstate
Modules linked in: ohci_hcd(-) g_serial rfcomm hidp l2cap bluetooth mts_io cdc_acm at91_mci spidev [last unloaded: ohci_hcd]
CPU: 0 Tainted: G D W (2.6.35.14 #1)
PC is at __bug+0x1c/0x28
LR is at __bug+0x18/0x28
pc : [<c002bb64>] lr : [<c002bb60>] psr: 60000093
sp : c3fa1ed0 ip : c03ac69c fp : 000080d0
r10: 000000d0 r9 : 00000010 r8 : 00000000
r7 : 000080d0 r6 : c3c05600 r5 : c3c02620 r4 : c3c00120
r3 : 00000000 r2 : c3fa1ec4 r1 : c032c351 r0 : 00000024
Flags: nZCv IRQs off FIQs on Mode SVC_32 ISA ARM Segment user
Control: 0005317f Table: 23328000 DAC: 00000015
…

It seems that the modem is the problem.
We have multiple CRON task running.
One of the CRON task bash script (3min period) try pinging google. If the ping fails and if we cannot perform a get RSSI command we reboot the radio.
Here is the script:
if ! ping -c 1 google.com
then
mts-io-sysfs store led-sdk-c 0
logger -t NOMAD “Internet lost -> reseting ppp connexion”
if pidof pppd
then
$PPP stop
fi
sleep 5
rssi=$(echo -e “AT+CSQ15″|microcom -t300 -d300 -X /dev/modem_at1|grep CSQ:|cut -c7-|cut -d, -f1|tr -d ‘” 111213141532’)
logger -t NOMAD “Radio signal: “$rssi
if [[ $rssi == 99 || $rssi == “” ]]
then
logger -t NOMAD “No signal detected -> reseting radio”
/usr/sbin/radio-reset-h5
fi
$PPP start
logger -t NOMAD “ppp connexion (re)started”
fi

We have another CRON task script that checking SMS on the modem.

Could someone give us an idea to fix it.
Thanks a lot

[Darrik Spaude]

I recommend you create a Support Portal case and include all of the affected serial numbers.

https://support.multitech.com/

Thanks,
Darrik

[Author]

Posts