This thread has been locked.
If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.
Hi,
Customer is setting EtherNet/IP Conformance Test tool and using AM243-LP as DUT(Device Under Test) to run test.
In order to reproduce the test environment, They have some questions about setting.
Do they need to check the below checkbox?
Hi Nilabh,
Is there any updated today?
Kindly remined, the "License Expired" is still a problem on version 8.6. According to the last meeting, KUNBUS promise they will update a new version to fixed it. Please help us to make sure KUNBUS will not forget this issue.
Thank you & have a good day.
Best regard
Bo-Cheng
Hi Bo-Cheng,
Kindly remined, the "License Expired" is still a problem on version 8.6.
Is this issue coming in EthernetIP example?
According to the last meeting, KUNBUS promise they will update a new version to fixed it. Please help us to make sure KUNBUS will not forget this issue.
They could find a Siemens SCALANCE X204IRT now, and with that switch between the notebook and the DUT they in fact can reproduce the error the HiWin has seen. This is now under investigation and I will keep you updated. I will also try to get hold of a D-Link switch still to confirm. Below the tests from today.
This is being worked on with high priority.They have now identified more switches that expose the ACD issue reported by HiWin Mikro.
ACD Behavior Test V2.4.13 Run All Generic Device Port 1 LP-AM243 Scalance X204IRT 2023-07-12T15_MCU_PLUS_SDK_08.06.00.43.zipACD Behavior Test V2.4.13 Run All Generic Device Port 1 LP-AM243 TP-Link 2023-07-12T15_MCU_PLUS_SDK_08.06.00.43.zip
Hi Bo-Cheng
we have now a fix for the ACD issue. Kunbus have verified it with three different switches with which we earlier had seen the error. Scalance, Netgear, and 3Com. With all the ACD test passes. however they are struggling with the PHY integration for the 08.06.00.46 release(new release planned). Kunbus is intensively investigating this now.
Hi Nilabh,
May I ask for the update to the 08.06.00.46 release schedule?
Is the issue of the license expired will be fixed in the 08.06.00.46?
Thank your do much for the help.
Best regard
Bo-Cheng
Hi Bo-Cheng,
We are still in progress for debugging the PHY issue. Yes it will fix the license expiry issue.
Hi Nilabh,
May I ask for the update of the PHY issue?
Thank you for the help.
Best regard
Bo-Cheng
Hi Bo-Cheng,
The issue, is not seen with the latest release now. We are planning the release by 3rd week of Aug. The license expiry issue has also been resolved in this release.
Hi Nilabh,
Thank you so much.
To this version, we also want to ask KUNBUS for a help to recheck the two new added LLDP test cases in conformance test to ensure the two functions are okay before the new firmware release.
The test cases are "265-LLDP Management Object" and "266-LLDP Data Table", please refering to the attached figure.
Thanks for your help.
Best regard
Bo-Cheng
I have asked them, Will update here once I get the feedback, latest by early next week.
Below is Update from Kunbus on ACD issue:
During out testing for the 9.0 release, we now see - depending on the switch that is used - additional errors during the ACD tests.
We have now traced that down and believe those errors are not triggered by the KUNBUS software, but by the switch implementation.
It appears that during the initial ACD phase after the link-up phase, where the device under test is supposed to send 4 ARP probes which, if there is no conflict, are followed by 2 ARP announce frames. It looks like that some of the switches we tested drop some of those ARP probes, and the CT19.1 test script complains that it sees the ARP announce frames without the correct number of prior probes.
To verify this, we connected a tap device between the DUT and the switch and recorded the corresponding frame sequences before (bottom trace) and after (upper trace) the switch. Aligning the time stamps, you see the the particular switch used here (D-LINK) drops even 3 of the ARP probes which causes test cases to fail. However it can be seen that our EtherNet/IP stack correctly transmits the 4 probes.
We will take that up into the release notes, but currently I am not sure if we can fix this as it seems to be a problem of the switches used.
Regards,
Hi Nilabh,
Thank you for the update.
ODVA's opinion to this behavior with D-Link switches:
They think the sequence is as below.
1. Plug in the cable.
2. Switch receives the LinkUp event.
3. Switch initializes communication.
4. Switch begins transmitting packets.
5. Device begins sending the ACD ARP sequence (4 Probe, 2 Announce).
Their advise is as follow.
"For some of the switches, the time taken for step 3 is longer than the time your device takes to begin step 5, so the first packets sent are not transmitted. This is a deficiency in the switches. For the D-Link switch, the first 3 probes are lost, so the switch has taken at least 400ms before it begins transmitting."
I hope these message will helpful to the KUNBUS team.
This issue is very important for us since we plan to start the product certification before the end of Q3. We hope this issue could be clearify ASAP, but not just write on the release note.
Best regard
Bo-Cheng
Thanks Bo-Cheng,
I have shared this feedback to Kunbus. I will update here when they share an update, latest by Thursday this week
Hi Nilabh,
I think you are talking about the KUNBUS.
Kindly remind, please do not mention our company in this discussion.
Thank you.
Bo-Cheng
Hi Nilabh,
Please help to foward the following LLDP conformance test problems to KUNBUS team.
The test cases are "LLDP Management Object" and the "LLDP Data Table".
Thank you for the help.
Bo-Cheng
Hi Nilabh,
May I ask for the update of the ARP issue which shell be updated on last Thursday.
Thank you
Best wish
Bo-Cheng
Hi Nilabh,
Could you ask KUNBUS for updating the ARP problems on Aug,4, and the LLDP problems on Aug, 10?
Thank you
Best regard
Bo-Cheng
Hi Nilabh,
We saw the updates of MCU_PLUS_SDK v09.00.00.30 (18 Aug 2023), but we can't find the folder "industrial_comms" in this version.
Does this version have EtherNet/IP
Thanks,
Chia-Hsun
Hi Bo-Cheng,
I will update the Status from Kunbus by tomorrow.
We saw the updates of MCU_PLUS_SDK v09.00.00.30 (18 Aug 2023), but we can't find the folder "industrial_comms" in this version.
Does this version have EtherNet/IP
Hi CH Li, From 9.0 SDK we will be having separate INDUSTRIAL COOMS SDK, it is planned by end of August 2023.
Hi Nilabh,
About the 9.0 SDK, may we ask for a more precise date?
It looks a little delay then the third week of Aug.
Best regard
Bo-Cheng
Hi Bo-Cheng
last week Kunbus performed a few tests with an increased delay (PROBE_WAIT, set in the RFC 5227 to 1 second, shortened by ODVA in Vol 2, F-1.2.1 IPv4 ACD Timing Constants to 200ms)
Cisco SG350-20 |
||||||
|
Seen frames on system start (ARP probes and ARP announcement) |
|||||
Probe Wait Time |
first attempt |
second attempt |
third attempt |
fourth attempt |
fifth attempt |
Average |
200 |
5 |
5 |
5 |
5 |
5 |
5 |
500 |
6 |
6 |
6 |
6 |
6 |
6 |
750 |
6 |
6 |
6 |
6 |
6 |
6 |
1000 |
6 |
6 |
6 |
6 |
6 |
6 |
D-Link DGS-12210-10MP |
||||||
|
Seen frames on system start (ARP probes and ARP announcement) |
|||||
Probe Wait Time |
first attempt |
second attempt |
third attempt |
fourth attempt |
fifth attempt |
Average |
200 |
1 |
1 |
3 |
1 |
1 |
1,4 |
500 |
1 |
1 |
3 |
2 |
3 |
2 |
750 |
6 |
1 |
6 |
2 |
4 |
3,8 |
1000 |
2 |
6 |
6 |
1 |
4 |
3,8 |
So the D-Link switch seems to be really problematic. Whereas our Cisco switch apears to transmits all ARPs with a initial delay of 500ms or more, the D-Link does not even at 1s. They will try to contact the ODVA certification lab in Germany today.
After we replace the D-Link DGS-1210-10MP switches by a PHOENIX CONTACT 2891001 FL switch
The PHOENIX CONTACT 2891001 FL switch is described as an unmanaged switch.
Whereas the D-Link DGS-1210-10MP is a managed switch. The at DGS-1210-10MP user manuals at https://kc.dlink.com/Home/FindBlob?FileName=DGS-1210-10MP describes a "Safeguard Engine" which is enabled by default.
Where the Safeguard Engine "throttles the impact of packet flooding into the switch's itself". Not sure if could be causing ARP packet loss, and if worth investigating the effect of disabling the Safeguard Engine.
Hi Chester,
I think this question is better to discuss with ODVA USA, since the D-Link switch is used by thier certification tests.
If ODVA thinks such behavior is not a problem, we can mark issue as resolved.
Matthew Frazer <mfrazer@odva.org> is my currently contact ODVA expert.
Best regard
Bo-Cheng
Hi Nilabh
I thnik it's better to let Cheter or Kunbus experts to contact ODVA directly because they know their test results more clearly and they have the test data.
They would be easier to explain and answer question to ODVA experts.
Thank you
Bo-Cheng
Response from Kunbus:
I wanted to perform some additional tests. So,, getting back to the discussion that I had last Wednesday with Falko Bendik, who is running the ODVA TSP in Magdeburg in Germany. We essentially discussed the issues that are observed with different switches during the ACD testing. He confirmed that this is essentially a known problem, and as I mentioned in he previous reply, discouraged us from increasing the PROBE_WAIT time (ODVA requires 200 ms +/- 10%). First of all, as you also see from the tables I posted earlier, this does not necessarily make the test perform better, secondly he raised the valid question to which delay one should increase this time. Clearly one cannot increase it too much, as this would contradict the argumentation of ODVA that the timings referred in RFC 5227 are not acceptable for industrial applications, and they need to be shorter.
It seems to be clear that this would exclude quite a few switches from usage within EtherNet/IP installations. He also suggested that a firmware update of the D-Link switch may possibly help, but this did not result in any improvement. To us it apears that the D-Link switch is just to slow. In any case when putting a network sniffer between the DUT and the switch, and then comparing the Wireshark traces from the network sniffer with those recorded on the test PC, we clearly see that the DUT transmits all the required four ARP probes and the two announce frames with the correct timings, but the are not correctly seen after the switch anymore. The only switch device where we currently do not see this issue at all is the little tp-link TL-SG108E office switch.
I suggested to ODVA that they may consider to publish a list of switch devices where they know that those problems due not occur (ODVA knows that there are problems), which may help end-users to assemble their installations. I'm curious what the reaction on part of ODVA will be here.
Hi Nilabh
Thanks for the updating.
Please ask Kunbus for the firmware release date they expect.
We also wanted to know if it's possible can we can get a beta version fimware for rechecking other previous issues, i.e. the license expired one.
Best regard
Bo-Cheng
Hi Bo-Cheng,
We are planning for INDUS COMMS SDK release by 1st week of september, this will have the fixes.
Hi Nilabh,
Thank you for updating.
The following questions KUNBUS haven't replied yet. Please help to ask Kunbus for support.
====== Previous question =========
The test cases are "LLDP Management Object" and the "LLDP Data Table".
5086.Log and Wireshark files.zip
=============================
Best regard
Bo-Cheng
Hi Bo-Cheng,
Could you please create a new thread for this. It is becoming difficult to track this thread, as there a lot of issues. Appreciate your cooperation. I will pick up that thread.
Hi Nilabh,
Please help to update when the firmware is available this week?
I also have created a new thread for the LLDP question.
https://e2e.ti.com/support/microcontrollers/arm-based-microcontrollers-group/arm-based-microcontrollers/f/arm-based-microcontrollers-forum/1267369/am2434-am2434-am2434-ethernet-ip-lldp-test-question
Best regard
Bo-Cheng
Thanks Bo-Cheng,
Lets follow up on the other thread. I will close this. We are planning for a release tomorrow or Monday.
Hi Nilabh,
We haven't seen the new firmware release in this morning.
Could you inform us here when the firmware is released?
Best regard
Bo-Cheng