CC3220SF: SP outside stack!

Roman Jordan

Expert 2051 points

Part Number: CC3220SF

Hi there,

I am using the following environment:
- CCS Version: 10.2.0.00009
- SDK simplelink_cc32xx_sdk_4_40_00_07
- costumer board with CC3220MODASF
- TI-RTOS
- use of Posix threads in the application/ firmware

I have debugged my program. At one point the ROV shows me the error shown in the figure: SP outside stack!
ROV marked this error (?) in red.

This error is reproducible and occurs when I step over the function "getData" that loads some values from an SPI flash.
The breakpoint is just before the function call, step over this function and the ROV shows me the the error.

I attached two pictures, see below:
The first shows the ROV output just before calling the function "getData" ( 4 active threads ).
The second shows the ROV output after stepping over the function "getData" ( 6 active threads ).

The function "getData" itself is part of an initialization routine that is called from inside a Posix thread.
During the time the function is working two other threads are started in the background (see the second picture).

If I move the breakpoint to the end of the initialization routine or at a point somewhere later in the program, no error message appears.
I found that the error only occurs when the getData function uses SPI transfers. Guarding SPI access using mutexes does not change this behavior.

In addition, I cannot find any errors in the program. The requested data is loaded, checked and then processed.
The threads started at the time the function getData is working, are working as expected.
If I hadn't accidentally set the breakpoint exactly in front of the getData function, I would never have seen "SP outside stack!".

I can imaging that the ROV message occurs because of the combination of creating threads are started and the location of the breakpoint.

This points me: I can ignore the message.

But I have to sure about this.
Can you please help me about this ROV message?

Thanks for reading all of this.

Best regards,
Roman

over 4 years ago

0 Sarah P over 4 years ago

TI__Mastermind 38980 points

Hi Roman,

That is odd you are seeing that error and your application is working as expected, but this may just mean your application is not using the memory location that is being overwritten. Typically "SP outside stack" means you have a stack overflow or you are jumping to a bad pointer. Can you try breaking earlier in your application and stepping through?

Here are two other threads as references for debugging: https://e2e.ti.com/support/legacy_forums/embedded/tirtos/f/355/t/254793 and https://e2e.ti.com/support/microcontrollers/other/f/908/t/685353

Best regards,

Sarah

0 Roman Jordan over 4 years ago in reply to Sarah P

Expert 2051 points

Hello Sarah,

Thank you for your reply.

I've already stepped through the code in the function where the error occurred. It was a bit of a mystery because ROV showed me the "SP out of stack" error in this case in an another thread while in step mode. It is possible that both threads are using the same error function - i will check

OK. If you think there is a bug somewhere, I need to invest more time on this topic.

Many thanks,
Roman

0 Arnaud Moeschwitzer over 4 years ago in reply to Roman Jordan

Intellectual 785 points

Hi Roman,

are you running in an error handler ?

Which debug chain are you using ?

BR,

Arnaud

0 Roman Jordan over 4 years ago in reply to Arnaud Moeschwitzer

Expert 2051 points

Hello Arnaud,

I don't run into the error handler.

The stack check of the TI RTOS is activated.
halHwi.checkStackFlag = true;
Because of this, I expected that the system would encounter an exception handler when there are errors in the stack size. But that doesn't happen in my case.
If I force a stack size error for test then the exception handler is called.

At second, the data read are as expected. If I disable all breakpoints the program runs fine. At least it looks like this.

I use the Ti debugger XDS 110 of the LAUNCHCC3220MODASF that is connected to my board.

Best regards,
Roman

0 Roman Jordan over 4 years ago in reply to Roman Jordan

Expert 2051 points

Hello,

sorry, my first picture didn't show all details.

The priority of the "faulty" task is -1 at the time the "SP outside Stack!" message is shown.

Could it be possible that the "faulty" task is in the creation state and therefore the ROV message can be ignored?

Best regards,
Roman

0 Sarah P over 4 years ago in reply to Roman Jordan

TI__Mastermind 38980 points

Hi Roman,

I also suggest you enable Task.checkStackFlag, m3Hwi.enableException, and BIOS.assertsEnabled.

In your second image, all the text is red. Are you using a different view?

Best regards,

Sarah

0 Roman Jordan over 4 years ago in reply to Sarah P

Expert 2051 points

Hi Sarah,

for the second view I used the ROV classic.

Thanks fo the hint. I will set the checkStackFlag, m3Hwi.enableException, and BIOS.assertsEnabled and come back.

Best regards,
Roman

Wi-Fi

Wi-Fi forum

CC3220SF: SP outside stack!