WL1837MOD: WL1837MOD: Messages are not shown on AP (receiver side) nr2

And Error come back every 20 minutes....

Hi, 

Can you please let me know your setup details? This will help to replicate the setup here. 

Regards, 

Sudharshan K N 

   
• Description of use case
We have an AP with a web server, the ECU are called AU.
For fault tracing, we have a minimal app, for each iteration:
1. Connects to server
2. From TC20 we are fetching one json file (192.168.242.1:3001/.../planogram)
3. Closing connection

******
* AU *
******
AU has a imx6 and TI chip - WL1837MOD.
It has a Yocto project - linux-variscite

chip.id = 0x6030111
chip.fw_ver_str = Rev 8.9.0.0.86
chip.phy_fw_ver_str = Rev 8.2.0.0.245

We are not running BT.

Qualcomm MSM8937

- Linux kernel version
It has a Yocto project - linux-variscite
Version: thud
uname -a gave: Linux var-som-mx6 4.9.88-mx6+gec7d7d8ccb49 #1 SMP PREEMPT Mon Oct 5 11:35:30 UTC 2020 armv7l GNU/Linux


**************
* Zebra TC20 *
**************

TC20 are zebra devices. Mobile phones/scanners
TC20 - 8.1.0


•What is the problem description?
With TC20 we always get stuck after a while, i.e. communication between TC20 and AU are not working. Re-transmissions are not handled => ARP messages send out from zebra, sometimes AU send out ARP request and communication starts again. Sometimes AU doesn't send out ARP request or doesn't respond on zebra ARP requests => not coming back.

Error come back every 20 minutes....

Please come back if you need something else.

Hi, 

Thanks for the details. We will take a look into it. Does the Ping work during this session? Also if two TI devices are connected to the AP does it behave the same way? Please let us know. 

Regards, 

Sudharshan K N 

Hi, 

Can you please send us the packet trace for this session where Wilink is used as AP? 

Regards, 

Sudharshan K N 

Hi, 

I have setup 2 TI units one set to AP mode and another as STA. I can see them pining each other without any connection loss for more than 8hrs now. I will continue the tests and let you know if i am bale to reproduce the failure. 

Regards, 

Sudharshan K N 

Hi,

Ping also stop to work if it is running at the same time. Seems to work lite better if ping are run from AP -> Zebra (not included in attached pcapng file)

AP works good if another Zebra with higher Android version (10) are used. The one that we see the problem in conjunction with the AP, does have android 8.1.0
If both Zebra units TC20 - Android 8.1 and TC21 - Android 10 are run at the same time, against same AP, we get problem with ARP with zebra TC20 and TC21 are working fine. Only TC20 communication are attached in pcapng file.

Attach link to monitor mode capture:

files.elysics.se/.../z2ZWSSsriNsFdMk

MonitorMode.pcapng - Logfile
MonitorMode.txt - description of errors. Please look at error 3, here problem continue (ARP) until we reset AU/AP unit.

Please let me know if you need other log files.

Hi, 

Thanks for the logs!! Does the logs contain the traces from TC21 as well? Also i see the retransmissions initiated from higher layers. Need to check if there is something missing in TCP layer as well. 

The setup i have is doing pings in both directions from AP to STA and STA to AP. Both are running linux unlike the one in this case where STA is running Android. We will take a look into the logs again and see if there is anything abnormal. This may need a few days to complete. 

Regards, 

Sudharshan K N

Hi,

TC21 and ping are not included in log file.

TC21 Android 10 are using something called Gratuious ARP, Android 8 didn't according to Zebra support. They think it's a problem with the AP because it isn't responding on ARP request from TC20.

Best regards,

Fredric and Johan

Hi, 

Does the IP address change in your testing? Does static IP configuration also show this issue? Please let me know. 

Regards, 

Sudharshan K N 

Hi, 

Also is any power save mode enabled on WiLink device? Also please let me know if you have provision to log the FW debug logs. This needs a UART connection to the WL_UART_DBG pin and GND. Please let us know. 

Regards, 

Sudharshan K N

Hi, 

No, same IP address are used during the same test. Only changes if I use multiple devices and don't remember to start them in correct order, to get same IP address from AP ;)

I have not tested static IP configuration. Will test it next week.

We have power save mode enabled. Tested to disable it but doesn't remember the result. Will test it next week.

From HW person I got this "the TI chip are already mounted on a SOM, so we don't have access to those pins". How can I test that it is avalable and what should I get out?

Best regards,
Johan

Hi, 

Fredric Westerlund1 said:

From HW person I got this "the TI chip are already mounted on a SOM, so we don't have access to those pins". How can I test that it is avalable and what should I get out?

Please refer to the schematics to check if the pin is available. 

Regards, 

Sudharshan K N

Hi,

We don't have a schematic, we bought it completely from Variscite

Best regards,

Johan

Hi, 

Ok. then we may have to see how we can recreate the issue at our end. so far my setups are still running after ~4days. Currently the ping is working both directions. if there is a specific way to reproduce the issue I would like to know. 

Regards, 

Sudharshan K N