ot-br-posix: ot-br-posix as an on-mesh commissioner fails to accept client hello packets

OOB connections are still working for me, as are DTLS connections soc-to-soc, however the case of soc to BR using DTLS psk is giving me problems. I did have this working previously so this did work before. Originally we had thought this was a thread 1.2/1.1 incompatibility issue but we found that 1.1 images also aren’t connecting properly.

Here is my BR console for one of the various tests I ran:

sudo ot-ctl factoryreset
sleep 3
sudo ot-ctl channel 11
Done
sudo ot-ctl panid 0xdf47
Done
sudo ot-ctl networkname R0GUE7
Done
sudo ot-ctl masterkey 00112233445566778899aabbccddeeff
Done
sudo ot-ctl extpanid 000df47000000000
Done
sudo ot-ctl pskc -p S1LABS6
Done
sudo ot-ctl ifconfig up
Done
sudo ot-ctl thread start
Done
pi@rpi1:~ $ sudo ot-ctl
> commissioner start
Commissioner: petitioning
Done
> Commissioner: active

>
> commissioner joiner add * S1LABS6
Done
> Commissioner: Joiner remove

And soc console:

> factoryreset
channel 11
Done
>
> panid 0xdf47
Done
>
> ifconfig up
Done
>
> joiner start S1LABS6
Done
>
> Join failed [Security]

Version information:

pi@rpi1:~/ot-br-posix $ git show
commit 274c11a371df6ffe0f7562359f31dd43af0fe871
Author: Jiacheng Guo <gjc@google.com>
Date:   Fri Sep 11 23:45:04 2020 +0800
> version
SL-OPENTHREAD/1.1.0.0 GitHub-babb66d9c; EFR32; Sep 14 2020 22:01:04
Done

I would attach the actual trace file but the format isn’t supported here and the tool I used was Silabs Simplicity Studio which you may not have installed. But here is a picture so you can see the point of failure: image

I’m using 2x brd4161a but I don’t think that matters. My topology is just the simple 2 node network BR to SoC.

Please let me know if I’ve done something incorrectly and/or what other information I can provide. Thanks.

About this issue

  • Original URL
  • State: closed
  • Created 4 years ago
  • Comments: 23 (15 by maintainers)

Most upvoted comments

@jwhui Can you please update the ot-br-posix pointer again? It looks like it was updated to openthread/master, but the fix on the stack seems to have gone in after this pointer update.

@suveshpratapa , see #596.