Question on Simulation build: why OT_REFERENCE_DEVICE=ON and impact on Ping Response

[Alami-Amine]

Hello,

I am wondering why we are activating flag OT_REFERENCE_DEVICE=ON in the OT_simulation build? isn't it supposed to be OFF for a DUT?
the reason i am asking is: I am investigating an issue I have with MATN_02_MLRFirstUse.py and I see that the behavior of this test changes when we have OT_REFERENCE_DEVICE=ON when using Posix platforms, with regards to Ping Replies:

#if OPENTHREAD_CONFIG_REFERENCE_DEVICE_ENABLE otIcmp6SetEchoMode(gInstance, OT_ICMP6_ECHO_HANDLER_ALL); #else otIcmp6SetEchoMode(gInstance, OT_ICMP6_ECHO_HANDLER_DISABLED); #endif

Basically my issue is that when running the equivalent of [MATN_02_MLRFirstUse.py ] on thread Harness, and using an ot-br-posix as a DUT with Router role, the test fails when OT_REFERENCE_DEVICE=OFF. and the symptom is :
I have is exactly the same as that described in this issue .
in the table summarized by @simonlingoogle, I am in R2, Certification Column. Basically I can only get the test to pass when i manually add a route for the ping response in my TUN interface (ip -6 route add 910b::/64 dev wpan0)

Any help or clarification on issue described above?

[abtink]

The chnage was added in

• [posix] enable all ICMPv6 echo handling for reference device #7288.

There seems to some related earlier discussion on this:

• enable BORDER_ROUTING ot-reference-release#20
• BBR not replying to multicast ping for own registered address #7175

I don't recall details, @edmont do you happen to remember any points related to this.

[edmont]

Hi:

• [posix] enable all ICMPv6 echo handling for reference device #7288 is needed for the reference devices to be able to handle ICMPv6 echo and pass the certification tests as DUT.
• Regular border router devices (non-reference) should enable BORDER_ROUTING instead.

@aalami90 is your device failing to pass the test even with BORDER_ROUTING enabled?

[EskoDijk]

@edmont @abtink Thanks for clarifying. This will be further discussed in the TSTBED-275 JIRA.

[EskoDijk]

Final note, just to record what I think is the outcome on having a "reference device" exhibit different behavior than a "normal" non-reference device:

• the intention of OT / OTBR in general is that having a device into “reference device mode” should never by default change its behavior, compared to a regular non-reference device.
• if that does happen, it’s a bug or oversight of the implementer
• the reference device will expose additional (CLI) commands or options (available for reference devices only) that will allow a test harness or an operator to configure changes in the default behavior, including behaviors that go against the spec.

@edmont @abtink Is this correct?

[jwhui]

I believe this is all correct.

[abtink]

Yes. Correct.