LAUNCHXL-CC1352P: Queue Full Issue in sensor collector example code in FH mode

Sameena Shaikh

Prodigy 145 points

Part Number: LAUNCHXL-CC1352P
Other Parts Discussed in Thread: CC1352P

Hi,

SDK: 5.20

Example: Sensor & Collector example code. Minimum changes added

Code with changes added to the below link

drive.google.com/.../1Q8hyxEBf9Tj7gdFULDHLR57tZ_t1rHVG

Mode: FH mode, 200kbps

Hardware used: 2 launchpads CC1352P

We are using the sensor collector example projects with very minimal changes on the launchpad. Once the status changes to join, sensor send send 192 bytes of data every 32ms.

Below is a snapshot which shows the stats

the Sensor_sendMsg function returns status 0 when the data is transmitted successfully.

We see that when the status is not success and the Queue gets full and the send is not successful.

The status value returned at this time is 241 which is 0xF1 which shows that the queue is full and hence the data was not transmitted.

I was able to reproduce this in every attempt.

Can you let me know how the Queue being filled up be avoided?

Although the retries haven't been too many and still we see the queue being full.

The queue size is set as below which should be sufficient

-DMAC_CFG_TX_DATA_MAX=100
-DMAC_CFG_TX_MAX=100

Let me know if you need any other information.

Thank you in advance.

Regards,

Sameena Shaikh

over 3 years ago

0 Sameena Shaikh over 3 years ago

Prodigy 145 points

0 Sid over 3 years ago in reply to Sameena Shaikh

TI__Mastermind 20412 points

Hi Sameena,

Thank you for the code. I will try to reproduce the issue and get back to you with feedback soon.

Regards,

Siddanth

0 Sid over 3 years ago in reply to Sid

TI__Mastermind 20412 points

Hi Sameena,

I am working on this case. But I still don't have any feedback to give. I will get back to you with feedback in a couple of days.
Sorry for the delay.

Regards,

Siddanth

0 Sameena Shaikh over 3 years ago in reply to Sid

Prodigy 145 points

Hi Siddanth,

Thank you for looking into it.

Were you able to reproduce it?

Regards,

Sameena Shaikh

0 Sid over 3 years ago in reply to Sameena Shaikh

TI__Mastermind 20412 points

Hi Sameena,

I was able to try your code today, I was not able to reproduce the issue that you are seeing. I simply built the projects and flashed the firmware on two CC1352P launchpads.

These are the outputs I got during my attempts.

Regards,

Siddanth

0 Sameena Shaikh over 3 years ago in reply to Sid

Prodigy 145 points

Hi Siddanth,

Could you please try with 2 sets of launchpads running tha same code.

I mean Set 1: 2 CC1352P :a sensor and collector.

Set 2: 2 CC1352P :a sensor and collector.

It was easy to reproduce for me with 2 sets of launchpads.

Regards,

Sameena Shaikh

0 Sid over 3 years ago in reply to Sameena Shaikh

TI__Mastermind 20412 points

Hi Sameena,

If I understand you correctly, you have flashed the same examples on two sets of launchpads. You sent me 1 code example. I wanted to confirm if you have changed the PAN ID, and the network name when flashing the firmware on the second set of devices.

Regards,

Siddanth

0 Sameena Shaikh over 3 years ago in reply to Sid

Prodigy 145 points

Hi Siddanth,

Yes that had been taken care of.

The 2 sets of launchpads are kept close to each other.

Regards,

Sameena Shaikh

0 Sid over 3 years ago in reply to Sameena Shaikh

TI__Mastermind 20412 points

Hi Sameena,

Please can you provide the changes that you made for the project when flashing the two sets of devices and the steps that you followed to produce the error, so that I can reproduce it here.

Regards,
Siddanth

0 Sameena Shaikh over 3 years ago in reply to Sid

Prodigy 145 points

Hi Siddanth,

Minimal changes made to sensor collector project

The changes made on the example code were as follows:

* Enabled FH mode

* Added dummy data of 192 bytes in a packet to be sent from sensor to collector

* Tx - Rx stats added to compare output of each set.

* Kept PanId for one set as 0xFFFF and changed for the other set 0x1FFF

* Kept Co-ord short address for one set as 0xAABB and changed for the other set 0xAABC

Tested with above changes for multiple attempts

This is the observation when the FH mode is running on both sets of launchpad and the launchpads are placed close to each other.

Regards,

Sameena Shaikh

0 Sameena Shaikh over 3 years ago in reply to Sameena Shaikh

Prodigy 145 points

Hi Siddanth,

Also would like to mention one thing, I have been trying to reproduce the issue on my side since yesterday since you said it was not reproduced on your end.

I have been able to reprodcue it.

The issue is reproduced at alternated times. At times the issue is not reproducible. Also I sometimes need to keep it for longer duration, overnight too to get the issue reproduced.

Regards,

Sameena Shaikh

0 Sid over 3 years ago in reply to Sameena Shaikh

TI__Mastermind 20412 points

Hi Sameena,

Sameena Shaikh said:
* Kept PanId for one set as 0xFFFF and changed for the other set 0x1FFF

Please can you try with distinct PAN IDs on both sets of devices. Do not keep the default value of 0xFFFF on any set. Change it for both sets to two different PAN IDs. Check if the error persists.

Regards,

Siddanth

0 Sameena Shaikh over 3 years ago in reply to Sid

Prodigy 145 points

Hi Siddanth,

I am on leave for a week.

I will get back to you with the results once I'm back.

Regards,

Sameena Shaikh

0 Sameena Shaikh over 3 years ago in reply to Sameena Shaikh

Prodigy 145 points

Hi Siddhant,

I did try with the changes as mentioned by you but the observations have still been the same.

With 2 sets running for about 5-6 hours, below are the observations. Following changes were done as per your suggestions

With 2 sets of devices kept at a distance of 2cm each, following are the stats.

Attempt 2 also yielded similar results

0 Sameena Shaikh over 3 years ago in reply to Sameena Shaikh

Prodigy 145 points

Following are the results with one set of launchpads. One sensor and one collector.

The test was run for 5-6 hours each for the issue to be reproduced.

No movement of the launchpads were done while the test was ongoing.

Regards,

Sameena Shaikh

0 Sid over 3 years ago in reply to Sameena Shaikh

TI__Mastermind 20412 points

Hi Sameena,

If it took 5-6 hours to produce the error, I need some time to reproduce it. I will try to reproduce today.

Regards,
Siddanth

0 Sid over 3 years ago in reply to Sid

TI__Mastermind 20412 points

Hi Sameena,

I looked into your code and see a possible issue with the increment of the failQueue counter.

Please log/observe the status code returned by ApiMac_mcpsDataReq(&dataReq) in the file sensor.c.

I see that you increment the failQueue stat whenever the status is not equal to ApiMac_status_success.

Please check what is the status returned with ApiMac_mcpsDataReq and increment the failQueue counter only when the status is 0xF1 or 241.

Regards,
Siddanth

0 Sameena Shaikh over 3 years ago in reply to Sid

Prodigy 145 points

Hi Siddanth,

I'll make the change as suggested by you and test and update the results.

Regards,

Sameena Shaikh

0 Sameena Shaikh over 3 years ago in reply to Sameena Shaikh

Prodigy 145 points

Hi Siddanth,

I did try with the changes as you suggested

Incrememnted the failQueue counter only when the status is F1

The results are still similar.

But each time I the issue is reproduced at different intervals.

Sometimes it took me 3-4 hours to reproduce. Sometimes an overnight with 14-15 hours.

But I have always been able to reproduce the issue.

I haven't been able to get a solution yet.

Regards,

Sameena Shaikh

0 Sid over 3 years ago in reply to Sameena Shaikh

TI__Mastermind 20412 points

Hi Sameena,

I do see from your screenshots that the number has improved. And in the API documentation linked here, it seems that it is the defined behavior of the MAC.

https://dev.ti.com/tirex/explore/content/simplelink_cc13x2_26x2_sdk_5_20_00_52/docs/ti154stack/mac_api/html/api__mac_8h.html#a8a4a0c53c8e995032157263b24f6d2e1

If you are seeing the buffer overflow, you might have to make the requests at a lower frequency or buffer the request and make the request again when the MAC is decongested.

Regards,

Siddanth

Sub-1 GHz

Sub-1 GHz forum

LAUNCHXL-CC1352P: Queue Full Issue in sensor collector example code in FH mode