Zephyr and OpenThread join support

[Louis]

While attempting to join a network formed by the OpenThread border router reference implementation and an external commissioner using the Android app, my zephyr end device (based on the lwm2m client sample) ends up failing an assertion:

assertion "awaited_property == AWAITED_PROPERTY_NONE" failed: file "WEST_TOPDIR/zephyr/modules/hal_nordic/nrf_802154/serialization/platform/nrf_802154_spinel_response_notifier.c", line 63, function: nrf_802154_spinel_response_notifier_lock_before_request

I'm not sure I'm doing this properly. CONFIG_OPENTHREAD_CUSTOM_PARAMETERS has to be set in order to be able to link the binary which I think is a little suspicious and not documented as far as I know:

CONFIG_OPENTHREAD_JOINER=y
CONFIG_OPENTHREAD_CUSTOM_PARAMETERS="OPENTHREAD_CONFIG_JOINER_ENABLE=1"
CONFIG_OPENTHREAD_JOINER_AUTOSTART=y
CONFIG_OPENTHREAD_JOINER_PSKD="J01NU5"
CONFIG_MBEDTLS_HEAP_SIZE=10240

[Carl Richard]

Hello, Louis!

Are you using Zephyr standalone or the nRF Connect SDK, based on Zephyr? In the latter we provide a OpenThread configuration guide and samples that should be more straightforward to setup with our devices. 

I will ask the developers about the issue you're seeing here.

Best regards,
Carl Richard

[Louis]

Thanks,

I'm using the nRF SDK. I tried to add join support to the cli sample. Configuring the application with OPENTHREAD_JOINER_AUTOSTART didin't do anything (no failed assertion but no join either). So I added a "join button" (project here) and ended up with the same failed assertion.

[Carl Richard]

Just want to clarify here; the recommended way to update the nRF Connect SDK is to pull the newest sdk-nrf (nothing else) and then do west update. Manually pulling other repos could result undefined behavior or the SDK not working at all.

Can you confirm that you follow the recommended procedure?

Best regards,
Carl Richard

[Louis]

Hi Carl,

I rebuild my dev environment and only pulled nrfxlib before performing a "west update". I confirm that the pull request you mentioned made it to ZEPHYR_BASE. I also added : 

CONFIG_NRF_RTC_TIMER_USER_CHAN_COUNT=2

to cli sample with added join support. Unfortunately, still the same result.

Louis

[Carl Richard]

Hi again!

Please try to pull in nrf, not nrfxlib and then do the west update. 

Best regards,
Carl Richard

[Louis]

still the same result after:

# cd nrf
# git pull origin master
# cd ../nrfxlib
# git pull ncs master
# cd ..
# west update
# cd nrf
# git rev-parse HEAD
75db44dec9af4f52861759b05bcb3b309bcf3544
# cd ../nrfxlib/
# git rev-parse HEAD
5c6944ecf908d050da93d8755c3441a35ebdfa98

[Carl Richard]

Hello again, Louis!

Apologies for the delay here. I have been awaiting further comments from the developers, but have yet to receive any. In the meantime I would like you to test this a Thread: Cli application compiled on my system. Attached is the hex files for both the Application Core and the Network core. Please flash these to an nRF5340 DK and test whether it works for you or not. 

Network core:GENERATED_CP_NETWORK_merged_domains.hex
Application core: GENERATED_CP_APPLICATION_merged_domains.hex

For the on mesh joining procedure I've been following this documentation. It may be a bit different for external commissioners. Please report back your results!

Best regards,
Carl Richard

[Louis]

Thanks Carl,

Different result but still failed assertion triggering a reboot:

uart:~$ ot joiner start J01NU5
Done
uart:~$ I: State changed! Flags: 0x0800800b Current role: 0
I: State changed! Flags: 0x0800c00b Current role: 0
E: No response within timeout 500
ASSERTION FAIL @ WEST_TOPDIR/zephyr/drivers/ieee802154/ieee802154_nrf5.c:835
E: r0/a1:  0x00000004  r1/a2:  0x00000343  r2/a3:  0x00000001
E: r3/a4:  0x00006705 r12/ip:  0x0000a000 r14/lr:  0x000108d9
E:  xpsr:  0x41000000
E: s[ 0]:  0x00000000  s[ 1]:  0x00000000  s[ 2]:  0x00000000  s[ 3]:  0x00000000
E: s[ 4]:  0x00000000  s[ 5]:  0x00000000  s[ 6]:  0x00000000  s[ 7]:  0x00000000
E: s[ 8]:  0x00000000  s[ 9]:  0x00000000  s[10]:  0x00000000  s[11]:  0x00000000
E: s[12]:  0x00000000  s[13]:  0x00000000  s[14]:  0x00000000  s[15]:  0x00000000
E: fpscr:  0x00061cfd
E: Faulting instruction address (r15/pc): 0x000462ec
E: >>> ZEPHYR FATAL ERROR 4: Kernel panic on CPU 0
E: Current thread: 0x200015c0 (ot_radio_workq)
E: Resetting system

I'll try to re-setup the OpenThread border router from scratch tomorrow. How did you setup your own?

[Carl Richard]

Hi again!

Apologies for the long delay here, but I've finally managed to reproduce your issue using an nRF5340 and external commissioning through a border router setup according to this codelabs tutorial. I will do further investigation, but it's clear that something is not correct here. Thank you for your patience!

Best regards,
Carl Richard

[Louis]

Thanks. No problem for the delay.

[Louis]

Hi Carl,

Just doing a follow-up on this. 

[Øyvind]

Hello, 

I have been assigned this ticket as Carl Richard is currently out of office. 

Since last response, we have release version 1.7.0 of nRF Connect SDK which should have better support for OpenThread and the border router. Could you please try to upgrade to 1.7.0? I fear that the developers will ask the same questions, as I see that there are no updates in the internal ticket.

Kind regards,
Øyvind