System Error in PXI-8106 (exception 0x80000003)

[kbsave]

Hi,

I have a PXI-8106 (PharLap OS version 13.0) mounted in PXI-1031 chasis.

I have built an rt application which I deploy on the above target.

The application is set as Startup aplication to start whenever PXI boots.

Some times when I run this application the PXI crashes and gets disconnected. (MAX shows it disconnected.)

Then I need to hard reset the PXI to make it running again.

This behaviour is not consistent and happens at occasions.

When I connected a monitor to the PXI it showed following exception on the screen when it crashed:

"

System Error kernel.448380. Assert Failed : tcp-sqxmit : tried to send sq header

Rebooting system due to exception 0x80000003

".

The setup consists of following components:

NI PXI-1031 Chasis

NI PXI-8106 RT Controller

NI PXI-7813R FPGA module

Hi speed CAN 9853

LabVIEW version 8.6

I would like to know the reson for the crash and whether it is happening due to the code which I am running or is it an environment issue.

It would be really helpfule if someone can help me solving this problem as this issue has started becoming a workstopper for me due to frequent reboots.

Thanks in advance,

-Kaustubh.

[kbsave]

Hi,

No one has a solution to this problem?

Atleast a hint would be very helpful so that I can start digging the problem.

- Kaustubh

[Herok]

Dear Kaustub,

Try formatting your RT controller through MAX and reinstall the software including labVIEW RT through remote computer.

Are you using RT FIFO in your program? If yes can you try deploying a simple RT program and test the same.

Also I want to know if you are using shared variables in your project. Kaustub, when does the exception occur? During initialisation, during main execution or shutdown?

With regards,

Pavan 

[kbsave]

Hi Pavan,

Thanks for your reply,

We have not reformatted the PXI but this week we got a new PXI and we are facing the same issue with it. Seems like issue with the code.

1. We are not using RT FIFOs.

2. We are using shared variables. These shared variables are deployed on host and not on RT.

3. We have an startup executable on RT target. This VI calls many VIs when a Start trigger receives through a boolean shared variable.

    All these VIs are stopped (all their while loops) when another Stop trigger is received through another boolean shared variable.

   As per the observation when the Stop trigger is send then this problem of crash occurs.

4.We are using FPGA FIFOs.

Do you think its and issue with the shared variables?

The Shared variables are deployed by TestStand and then they are triggered from the VIs called from TestStand.

We are undeploying the shared variables in TestStand once the sequence gets over. But the RT startup exe might be using it.

Can this be the source of problem?

I would appreciate your help in resolving this issue. 

Thanks & Regards,

Kaustubh.

[Herok]

Hi Kaustub,

Can you try to deploy a very simple program and check if you are facing similar problems. Do keep the stop trigger in your program. Let's try to zero in on the source of the issue first. Please try that and reply with the results.

Regards,

Pavan

[Amit Ahuja]

Hi

Suppose i access a VI (remote panel) that is not currently in the memory and when this

VI gets instantiated on the RT, Can this call that is pending for the vi that got instantiated

later cause this exception.

Waiting for the response

Regards

Amit Ahuja

[cs_BOT]

Hi,

one thing you can try is deploying/undeploying the shared variables through LV and not through TS. 

also, If it is happening on RT-PC, could you look at the Error Log. In MAX , under remote systems, right click on the RT-PC and select View Error Log. See if any of the logs have something written under the Crash log text box. If there are multiple see if they have the same call stacks.

Is multicore support enable on your controller? are you using it? some thread might not be thread safe....

[Amit Ahuja]

Hi

Thanks For the reply.

I have a error log but i am not able to analyze it.

Please find attached the same and can you please look into it and get something out of it.

Yes Multicore support is enabled on the controller.

I am not using the multicore support explicitly but i suppose labview itself is handling various threads

that are created out of the code.

Is there any way of knowing which thread is not thread safe.

Regrads

Amit Ahuja