Issue Summary
We are encountering a failure during Matter multi-admin commissioning using OTBR with the JN5189 RCP. The primary commissioning using the CHIPTool Android APK completes successfully. However, the secondary commissioning using the CHIP Tool CLI fails.
During this process, the otbr-agent logs indicate repeated timeouts when communicating with the RCP, which causes the agent to become unresponsive. The only way to recover is by manually resetting the agent or replugging the RCP.
Setup Details
Component: Version / Description
-----------------------------|--------------------------------------------------------
Primary Commissioner : CHIPTool Android APK
Secondary Commissioner : CHIP Tool CLI (Linux)
Thread Device : Nanoleaf A27 Bulb
RCP Platform : NXP JN5189 over UART
RCP Firmware Commit: dfbe12280af4a87cb9ae0c0c1d90d71de3372d82 (from ot-nxp)
OTBR Platform: Ubuntu 22.04 (x86_64)
Kernel Version: Linux 6.8.0-60-generic ##63~22.04.1-Ubuntu SMP PREEMPT_DYNAMI
OTBR Commit: b8bab1babcb040251b51a69ccb4008fb4b9e2d3f
CHIP Tool CLI Commit: 9af93fd3e9eab4315e1097b19112ebac01d1ad94
Steps to Reproduce
Boot the OTBR with the JN5189 RCP connected and initialized.
Perform primary commissioning using the CHIPTool Android APK (successful).
Use the APK to open a multi-admin commissioning window.
Attempt secondary commissioning using the CHIP Tool CLI.
Observe otbr-agent behavior and log output.
Hello Vikash
The baud rate will depend on the one the host stablish. Just as an advice maybe you could try "1000000"; as for reference that baud rate is the default in an example for K32W.
The Software flow control is supported in JN5189 using USART0 and USART1, however the Hardware flow control is more reliable and doesn’t interfere with data, The decision would be up to every customer. Described in Chapter 23.6.5.2 UM11138
Best Regards
Luis
Hello,
The UART clock impact timing accuracy, if the UART clock is unstable, it can cause loss of synchronization between the host and RCP.
Match baud rates carefully between devices and test for tolerance margins, you could try a lower baud rate.
Or as other option; Instead of using the FRO32K, you could switch to the XTAL32K oscillator, which offers better clock frequency accuracy; You can find more information about it in the User Manual [ UM11138] Chapter 4.4.4 UM11138
Best Regards
Luis
Hello,
Sorry for the late response, The JN5189 supports Hardware Flow control in the USART0, you can find more information about it in the User Manual [ UM11138] Chapter 23.6.5.1 UM11138
The baud rate will depend on the one the host stablish.
Best Regards
Luis
Hello,
I review the file nxp_vendor_hook.cpp; This file is destined for both JN5189 and K32W061.
Just for testing purposes, I will recommend you use an i.MX8M and try to isolate the issue to the root cause.
This following link redirect to examples running Open Thread Border Router with i.MX for consulting
How to setup OpenThread Border Router and Openther Deamon on the target
Best Regards
Luis
Hi Luis,
Thank you for your prompt response.
I’ve tested the setup on the following two platforms and encountered the same issue on both:
i) Laptop (Ubuntu)
ii) Router (OpenWRT)
1) From my observations during the commissioning process, the Thread end device sends a rapid sequence of messages to the chip-tool commissioner in response to a request. This quick burst of traffic may be causing the RCP to drop some packets, which the OTBR is likely expecting and waiting for2).
2) For the external commissioning, I believe the main constraint is the PSKd, which is why I’ve raised the concern with the Thread group and am currently awaiting their response.
3) Could you also take a look at the following issue?
There is an nxp_vendor_hook.cpp file used when building for the K32W061 platform.
I'm wondering whether a similar file exists for the JN5189 platform or if it's possible to reuse the existing nxp_vendor_hook.cpp for JN5189 by enabling the DOT_NCP_VENDOR_HOOK_SOURCE macro during compilation.
Hello,
Have you tested running OTBR with an MCU/MPU instead of a computer? as the examples and guidelines that NXP support are destined for different Host's testings and have more information about it.
The compilation for the K32W061 could apply to the JN5189, both are very similar.
Also, could you take a look in External Commissioning | OpenThread if this apply to your additional commissioning issue?
Best Regards
Luis