Hi,
I have the problem, that after running for quite a while, lets say a week, the layer1 seems to crash. This crash is random and only happens on single phones (I have 6 phones connected to one mobile app).
This phone is unusable, even after "shutdown/no shutdown" or complete restart of mobile, this phone stays like this:
OsmocomBB# show ms 3 MS '3' is up, service is limited IMEI: XXXXXXXXXXXX IMEISV: XXXXXXXXXXXXX IMEI generation: fixed automatic network selection state: A1 trying RPLMN MCC=262 MNC=07 (Germany, O2) cell selection state: C1 normal cell selection radio ressource layer state: idle mobility management layer state: MM idle, PLMN search
By restarting the phone (pressing power on the phone, and reloading layer1) its working again. But this need physical access to the phone.
Is there a way to somehow "reboot" or restart the layer1 by software? And if not, could that easily be implemented?
Thanks
Tim
Is there a way to somehow "reboot" or restart the layer1 by software? And if not, could that easily be implemented?
Yes there is the L1CTL_RESET but I think shutdown/no shutdown already sends that command (check the osmocon output) meaning the L1 is no longer responding at all ...
Does the osmocon console does anything at all when it's in that state ? Does that phone have anything special ? (connected to a different network or doing different things ?).
Some people wire-up one of the serial port control line to the enable line of a power supply they use to power the phone so they can do full power off/on, but it requires external hw.
Cheers,
Sylvain
On Mon, 10 Dec 2012, Sylvain Munaut wrote:
Hi Sylvain,
Is there a way to somehow "reboot" or restart the layer1 by software? And if not, could that easily be implemented?
Yes there is the L1CTL_RESET but I think shutdown/no shutdown already sends that command (check the osmocon output) meaning the L1 is no longer responding at all ...
Does the osmocon console does anything at all when it's in that state ? Does that phone have anything special ? (connected to a different network or doing different things ?).
Nothing special and it is not only THIS phone. Randomly any of the phones are crashing.
Unfortunately I start the osmocon processes like this:
( /usr/local/bin/osmocon -s /tmp/osmocom_l2 -l /tmp/osmocom_loader -p /dev/ttyS0 -m c123xor /usr/local/bin/layer1.compalram.bin ) < /dev/null > /dev/null 2>&1 &
I could now start all of them in a screen session, where I could see the latest output afterwards. Since it is so much output coming, I don't want to keep it in a logfile.
If it helps I could do that, too. In any case I have to wait for the next crash...
But when I look at the cpu time used, it seems that the osmocon process is still doing something.
I could now test any command you tell me on the mobile application. Today in the evening, when I am home I will restart it (with osmocon in screen).
Some people wire-up one of the serial port control line to the enable line of a power supply they use to power the phone so they can do full power off/on, but it requires external hw.
Ok, you mean using an unused serial line and raising the control line to control a relais, which virtually presses the power button then?
With such a solution I would have another problem. Often, when I switch off the phone, I can't switch it back on. Even without the connected serial line, I can't switch it on booting the original firmware. I have to remove the power (which is a altered batteriepack, hooked on a powersupply with ~4 Volts [like the battery had before]), wait 30 seconds and put it back on. Then I can start it again.
Is that a known problem, or looks it like to be a problem of my altered powersupply?
Thanks
Tim
On Mon, 10 Dec 2012, Sylvain Munaut wrote:
Hi Sylvain,
Is there a way to somehow "reboot" or restart the layer1 by software? And if not, could that easily be implemented?
Yes there is the L1CTL_RESET but I think shutdown/no shutdown already sends that command (check the osmocon output) meaning the L1 is no longer responding at all ...
Does the osmocon console does anything at all when it's in that state ? Does that phone have anything special ? (connected to a different network or doing different things ?).
now phone number 6 is crashed. I started the osmocon processes in a screen session. A lot of messages are loged now in the session like this:
[...] FB1 (862461:1): TOA= 451, Power= -78dBm, Angle= 3182Hz fn_offset=862460 (fn=862461 + attempt=1 + ntdma = 0) delay=9 (fn_offset=862460 + 11 - fn=862461 - 1 scheduling next FB/SB detection task with delay 9 FB1 (862481:10): TOA=11703, Power= -78dBm, Angle= 3129Hz fn_offset=862480 (fn=862481 + attempt=10 + ntdma = 9) delay=9 (fn_offset=862480 + 11 - fn=862481 - 1 scheduling next FB/SB detection task with delay 9 FB1 (862501:10): TOA=11703, Power= -78dBm, Angle= 3135Hz fn_offset=862500 (fn=862501 + attempt=10 + ntdma = 9) delay=9 (fn_offset=862500 + 11 - fn=862501 - 1 scheduling next FB/SB detection task with delay 9 FB1 (862512:1): TOA= 451, Power= -78dBm, Angle= 3171Hz fn_offset=862511 (fn=862512 + attempt=1 + ntdma = 0) delay=9 (fn_offset=862511 + 11 - fn=862512 - 1 scheduling next FB/SB detection task with delay 9 FB1 (862532:10): TOA=11703, Power= -78dBm, Angle= 3113Hz fn_offset=862531 (fn=862532 + attempt=10 + ntdma = 9) delay=9 (fn_offset=862531 + 11 - fn=862532 - 1 scheduling next FB/SB detection task with delay 9 [...]
And mobile says:
OsmocomBB# show ms 6 MS '6' is up, service is limited IMEI: XXXXXXXXXXXX IMEISV: XXXXXXXXXXXXX IMEI generation: fixed automatic network selection state: A1 trying RPLMN MCC=262 MNC=07 (Germany, O2) cell selection state: C2 stored cell selection radio ressource layer state: idle mobility management layer state: MM idle, PLMN search
I now restated the mobile process and deleted 6.ba. Now it says:
OsmocomBB# show ms 6 MS '6' is up, service is limited IMEI: XXXXXXXXXXXX IMEISV: XXXXXXXXXXXXX IMEI generation: fixed automatic network selection state: A1 trying RPLMN MCC=262 MNC=07 (Germany, O2) cell selection state: C1 normal cell selection radio ressource layer state: idle mobility management layer state: MM idle, PLMN search
shutdown/no shutdown doesn't change a thing.
Now the question again: Could it be because of my hardware situation (special battery simulation with one power supply)? Or in other words has anybody running a steady service with osmocombb and experiences no problem over weeks?
Thanks
Tim
baseband-devel@lists.osmocom.org
-
Sylvain Munaut -
Tim Ehlers