CoAP client application can't reach nrf52840 CoAP server whereas IPv6 pings work

[AKYR]

Hello all ! 

So here is my problem, I have a computer as a CoAP client running the Python script from :https://infocenter.nordicsemi.com/index.jsp?topic=%2Fcom.nordic.infocenter.sdk5.v15.3.0%2Fiot_sdk_user_guides_linux_commands.html&cp=5_1_5_5_0_5

I connect a nrf52840 DK running the CoAP server example through a Raspberry Pi used as border router. 

Pings from my computer to the nrf52840  works fine and I can see the CoAP Get requests directed to my CoAP server from my computer on Wireshark.

However the nrf52840 doesn't seem to receive the request and nothing happens on the DK (and the computer doesn't receive a response). 

I have the folowing logs on Segger: 

<info> app: Physical layer in connectable mode.
<info> app: Physical layer: connected.
<info> app: Got IP Application Handler Event on interface 0x0x20004F9C
<info> app: New interface added!
<info> app: Sending Router Solicitation to all routers!

Here I launch the Python script and get the Wireshark capture.

<info> app: Got IP Application Handler Event on interface 0x0x20004F9C
<info> app: Interface removed!
<info> app: Physical layer: disconnected.
<info> app: Physical layer in connectable mode.

Note: After 25 seconds, the interface is removed. This refers to an other problem I have with my setup that a connection between the border router and the nrf52840 never last more than ~30 seconds, I have another bug about it is someone wants to help: 

https://devzone.nordicsemi.com/f/nordic-q-a/48277/ble-over-ipv6-between-nrf52840-dev-kit-and-rapberry-3-b-border-router-connection-interface-is-removed-after-30-sec

Does anyone have an idea why this happens or some tips to keep trying to resolve this issue ? 

Thanks in advance,

Aloïs KYROU

[Håkon Alseth]

Hi,

 

The server does not seem to reply.

The IP seems to be 2001:db8::d5:f7ff:fed1:70f6. This amounts to 5 byte BLE address, d5:f7:d1:70:f6, looks like you're missing a byte?

what does "hcitool con" show?

 

Kind regards,

Håkon

[AKYR]

Hello and thank you for answering :) 

pi@raspberrypi:~ $ hcitool con
Connections:
< LE 00:D5:F7:D1:70:F6 handle 64 state 1 lm MASTER

In my understanding, since the MAC starts with 00 we can concatenate 00:D5 into D5. 

[Håkon Alseth]

Ah, yes. You're right. The coap example reads out the BLE address, and sets it as public (upper bits set to '00'), but it was a bit suspicious that the remainder of the byte was all zero's, but then I noticed that the FW actually sets the whole byte to 0..

I just tested the example, with #define BLE_6LOWPAN_LEGACY_MODE 0, in ubuntu 19.10. Which coap-client are you using? If you're using libcoap-2-0-bin, I couldn't even get anything out (wireshark concluded the same, req was not sent from PC -> device)

installing libcoap-1-0-bin, then use coap-client to get, it started working:

$ coap-client -m get coap://[fe80::6c:30FF:fe79:30b5]/lights/led3
v:1 t:CON c:GET i:6c25 {} [ ]
0

Note: I set this up on my local PC, so I just use FE80:: prefix, but you still need to route bt0 and eth0.

Since I'm using FE80:: for testing, it looks like this:

sudo route -A inet6 add fe80::/64 dev bt0

 

You can test this by pinging directly, without specifying the interface, i.e. ping6 ABCD::MY:IP:ADDR. If this does not work, then the route is not OK.

 

Here's my wireshark log btw: coap_ipv6.pcapng

 

Could it be either the libcoap-1-0, or the route that is the problem on your side?

  

Cheers,

Håkon

[AKYR]

Thanks a lot for taking the time to realise these tests ! 

I did as you suggested and made work the example from border router with libcoap-1-0 with local-link address. 

I Set a second Rpi as Client for test that successefully pings the nRF52840 through the border router (replacing the PC in my previous example). I downloaded the same libcoap as the border router and sadly it doesn't work. 

On my PC I used the Python script from https://infocenter.nordicsemi.com/index.jsp?topic=%2Fcom.nordic.infocenter.sdk5.v15.3.0%2Fiot_sdk_app_coap_server.html&cp=5_1_4_2_4_1_0_4&anchor=iot_sdk_app_coap_server_python_sample

I guess my routes are the problem, but if it is the case how come I'm able to Ping from the Client ??

Global address of the CoAP Server : 2001:db8::/64 fc00::dcba:a54e:1868:4107

Global address of the RPi border router :  fc00::dcba:a54e:1868:4107 

Global address of the CoAP Client  (RPi)  :  fc00::5b16:ec6f:61a8:3e0c

Routes on the Rpi used as CoAP Client: 

pi@raspberrypi:~ $ route -A inet6
Table de routage IPv6 du noyau
Destination Next Hop Flag Met Ref Use If
2001:db8::/64 fc00::dcba:a54e:1868:4107 UG 1024 4 21 eth0
fc00::/64 [::] U 202 3 8 eth0
fe80::/64 [::] U 256 1 0 eth0
[::]/0 [::] !n -1 1 99 lo
localhost/128 [::] Un 0 4 3 lo
fc00::5b16:ec6f:61a8:3e0c/128 [::] Un 0 3 23 eth0
fe80::9bfa:892c:6e1c:a50c/128 [::] Un 0 3 18 eth0
ff00::/8 [::] U 256 5 815 eth0
ff00::/8 [::] U 256 1 0 wlan0
[::]/0 [::] !n -1 1 99 lo

Routes RPi used as border router: 

route -A inet6
Kernel IPv6 routing table
Destination Next Hop Flag Met Ref Use If
2001:db8::/64 [::] U 256 1 0 bt0
fc00::/64 [::] U 202 1 0 eth0
fe80::/64 [::] U 256 2 0 eth0
fe80::/64 [::] U 256 1 0 bt0
[::]/0 [::] !n -1 1 0 lo
localhost/128 [::] Un 0 3 0 lo
2001:db8::/128 [::] Un 0 3 0 bt0
2001:db8::1/128 [::] Un 0 3 0 bt0
fc00::/128 [::] Un 0 3 0 eth0
fc00::dcba:a54e:1868:4107/128 [::] Un 0 3 0 eth0
fe80::/128 [::] Un 0 3 0 eth0
fe80::/128 [::] Un 0 3 0 bt0
fe80::233e:c816:c332:c668/128 [::] Un 0 5 0 eth0
fe80::b827:ebff:fe76:ebd8/128 [::] Un 0 2 0 bt0
ff00::/8 [::] U 256 6 0 eth0
ff00::/8 [::] U 256 1 0 wlan0
ff00::/8 [::] U 256 4 0 bt0
[::]/0 [::] !n -1 1 0 lo

Ifconfig of the border router : 

ifconfig
bt0: flags=4161<UP,RUNNING,MULTICAST> mtu 1280
inet6 fe80::b827:ebff:fe76:ebd8 prefixlen 64 scopeid 0x20<link>
inet6 2001:db8::1 prefixlen 64 scopeid 0x0<global>
unspec B8-27-EB-76-EB-D8-18-67-00-00-00-00-00-00-00-00 txqueuelen 1000 (UNSPEC)
RX packets 7 bytes 230 (230.0 B)
RX errors 0 dropped 0 overruns 0 frame 0
TX packets 25 bytes 1891 (1.8 KiB)
TX errors 0 dropped 0 overruns 0 carrier 0 collisions 0

eth0: flags=4163<UP,BROADCAST,RUNNING,MULTICAST> mtu 1500
inet 192.168.1.209 netmask 255.255.128.0 broadcast 192.168.127.255
inet6 fe80::233e:c816:c332:c668 prefixlen 64 scopeid 0x20<link>
inet6 fc00::dcba:a54e:1868:4107 prefixlen 64 scopeid 0x0<global>
ether b8:27:eb:dc:41:72 txqueuelen 1000 (Ethernet)
RX packets 2299 bytes 215909 (210.8 KiB)
RX errors 2 dropped 0 overruns 0 frame 0
TX packets 344 bytes 34727 (33.9 KiB)
TX errors 0 dropped 0 overruns 0 carrier 0 collisions 0

If you see something that looks suspicious, because according to me this looks ok but I'm not very experimented... 

Thank you !

Regards,

Aloïs KYROU.

[Håkon Alseth]

Do you have any firewall running? Is ipv6 forwarding enabled?

There's a thread on this subject here: https://devzone.nordicsemi.com/f/nordic-q-a/5481/making-ipv6-device-available-on-the-internet/19137#19137

Could you check if any of these steps are configured incorrectly on your side?

 

Kind regards,

Håkon

[Seb-labs]

Hello Håkon ! 

I think I've found the source of the problem. 

Here is a capture I did on the border router :

On the top line CoAP client requests from its global address the global address of the CoAP client.

On the second line from the top, the CoAP client responds with its local-link address. 

On the third line the border router tells the CoAP client that it can't reach this address as it is out of scope from a link-local address. 

For the example to work, on the second line the CoAP client should use its global address (2001:db8::xxx). 

We can verify it by looking at the capture from a ping between the client and the server that works:

Here we clearly see that the nrf52 client responds using its global address. 

So do you have an idea why the nrf52840 DK would answer using its local-link address instead of its global one ? 

Thanks in advance,

Aloïs KYROU.

[Håkon Alseth]

Hi,

 

The server seems to work locally on the RPi, but accessing it on the network seems to be the show-stopper.

Its likely because there's routing missing somewhere in the chain. I am not very skilled in the network domain, but try googling "static route between two subnets" or similar.

Most of the threads are ipv4 related, but the sequence is very similar for ipv6.

 

Kind regards,

Håkon