Hard crash on upload of 25k file to /flash over ftp

[bartvandeenen]

I was going to play with installing xmodem.py on the LoPy, but no luck, even ftp-ing the file to /flash on the LoPy fails with a Guru Meditation Error (Interrupt wdt timeout on CPU0).
Reading the documentation of the WDT, it seems I can't disable it, and can't alter its timing. So what do I do? I'm just trying to make working with the LoPy somewhat more pleasant.

On a side node, the current Lora OTAA never returns 'has_joined()' is True, and ABP immediately crashes the LoPy. Is there a way to talk directly to the SX1272 via a serial port? I could try to figure out why your LoRa python module is not working.
Even better, just opensource the firmware.

(sysname='LoPy', nodename='LoPy', release='0.9.3.b2', version='d78a5a3 on 2016-11-05', machine='LoPy with ESP32')

[robert-hh]

Confirmed with a 24k sized file. A 16 k sized file could be loaded. Trying to save another file just wiped the file system.

[Batilan]

Accessing the SX127X probably requires accessing it via the SPI interface. But indeed better wait for the source to become available, or even better a fix ;-)

[daniel]

@bartvandeenen the ftp issue is caused by the WDT and it will be fixed as part of the release tomorrow. The firmware will go open source this week :-)

ABP should not crash the LoPy. Can you share your ABP code?

Cheers,
Daniel

[bartvandeenen]

@daniel

DEV_ADDR = struct.unpack(">l", binascii.unhexlify('F08C9A56'))
NWK_SESSKEY = binascii.unhexlify('30BB4B25282FA6E7C441F7BADC13A223')
APP_SESSKEY = binascii.unhexlify('56D6CB701264DE9836BCEC119FDDBD68')

def setup_lora():                                                                                                   
    lora = LoRa(mode=LoRa.LORAWAN)                                                                                  
    lora.join(activation=LoRa.ABP, auth=(DEV_ADDR, NWK_SESSKEY, APP_SESSKEY))                                       
    return lora                                                                                                     

I don't have the lopy connected here, so I can't test, but there's a hard crash at the moment def function setup_lora() is called.

It's awesome that the firmware is going opensource this week!!

[robert-hh]

In the absence of a documentation of lora.join, I cannot tell wther the code is right or wrong. There is an example in the docs, which looks different. Are you aware, that on the first line DEV_ADDR receives a tuple, and not a single 32biit binary value?
But nevertheless, a hard crash should not happen. If parameters are wrong, either an exception should be raised to a proper return code shall be given.

[robert-hh]

The crash by uploading a larger file is gone with today's build. Thanks a lot.

[bartvandeenen]

@bartvandeenen
Reply to myself!!!!

The ABP lora function call does cause a hard crash, if you provide it the wrong parameter:

DEV_ADDR = struct.unpack(">l", binascii.unhexlify('311...513'))

And then call

lora.join(activation=LoRa.ABP, auth=(DEV_ADDR, NWK_SESSKEY, APP_SESSKEY))

The problem is I forgot to take the first item from the struct.unpack array return!

To be honest, it doesn't help that the documentation shows this in blue on darkgrey