Multiple issues with LoPy TTN Nano Gateway

[Thosch42]

@jmarcelino
Currently I try to setup a firmware build environment on Win10 in order to debug the thing, but it works not yet.
Does anybody have a complete guidance or a docker image?

[gertjanvanhethof]

@Thosch42 I don't thing it has anything to do with you build environment.
I don't have guru meditation errors but still facing problems with acceptance of the gateway connection by TTN.
See: TTN Join accept not handled properly

Do you use the code from github?

[Thosch42]

@gertjanvanhethof
Yes, I use the original gitlab source, and also the original binary with Pycom Firmware Updater. But I run always into this Guru Meditation error. I tried it with the tutorial scripts from from gitlab and also with modified versions. Always the same.
Offline debugging tells me only, the bad code line is inside the byte code (what means, inside the python scripts, isn't it?).
Cheers,
Thomas

[jmarcelino]

@Thosch42

I’ve never seen anything quite that bad and I’ve seen a fair share of nano gateways.

Which nodes are you using? just LoPys?

It’s possible to get the offending line from the address given in the guru medication error, if you have the tool chain and a compiled image you can use addr2line included in the compiler tool chain

[Thosch42]

@jmarcelino Yes, just LoPys with Expansionboards V2.1A.
Good to know, with the addr2line command for gdb. I check it, as soon as I'm back in the office.
Cheers,
Thomas

[Thosch42]

@jmarcelino Ok, it took a while. Currently pretty busy:
The addr2line tool told me the following:

ts@Vrumfondell MSYS ~/pycom-micropython-sigfox/esp32/build/LOPY_868/release
$ xtensa-esp32-elf-addr2line -e application.elf -a 0x400f7b66
0x400f7b66
C:\msys32\home\ts\pycom-micropython-sigfox\esp32/../py/vm.c:1206

what leads me directly to

(gdb) info symbol 0x400f7b66
mp_execute_bytecode + 4166 in section .flash.text
(gdb) list *0x400f7b66
0x400f7b66 is in mp_execute_bytecode (../py/vm.c:1206).
1201                    ENTRY(MP_BC_BINARY_OP_MULTI): {
1202                        MARK_EXC_IP_SELECTIVE();
1203                        mp_obj_t rhs = POP();
1204                        mp_obj_t lhs = TOP();
1205                        SET_TOP(mp_binary_op(ip[-1] - P_BC_BINARY_OP_MULTI, lhs, rhs));
1206                        DISPATCH();
1207                    }
1208
1209                    ENTRY_DEFAULT:
1210                        MARK_EXC_IP_SELECTIVE();

but it does (me) not really help.

Cheers,
Thomas

[Thosch42]

@jmarcelino
Hi!
I still have this problem.Another question: Should there really not boot.py? Because I have one.

import os
from machine import UART
uart = UART(0, 115200)
os.dupterm(uart)

Or should I better delete the boot.py and connect the LoPy via Wi-Fi?

Cheers,
Thomas

[rm]

@jmarcelino said in Multiple issues with LoPy TTN Nano Gateway:

https://forum.pycom.io/topic/1361/firmware-release-1-7-3-b1

After my update from 1.6.2 to 1.7.x the nano gateway stopped functioning. Is there anyone who has working code for a nano gateway functioning on 1.7.3-b1 firmware?

[Thosch42]

OK, it seems running now. But I don't know why.
I updated the device with the newest 1.7.5.b1 firmware via update tool - no success.
I flashed my own build from the latest pull - get a red light after update
I updated the device again with the update tool - no red light, but also no success
I synced the lopy directly from the git working directory, just changed the three lines in config.py, leaving the otaa_node.py where it is (so it copied it also to the gateway) and - Tataa - the gateway runs now for more than an houre.
Ok, the node is not joining, but I#ve seen, there are some threads about this issue.

And I have to find out, why I get a red light and a freeze, when I flash my own build from the original source.

Cheers,
Thomas

[jmarcelino]

@Thosch42
This tip from Robert should fix your red light:

https://forum.pycom.io/topic/1343/solved-lopy-bricked-red-led/3