TMS320F28379D: CLA debugging problem

Hi  Ralf, I am going to pass your question to the Code Composer Team.  I'm not sure what is causing the inability to read the variable. 

A couple of questions - is the C28x doing anything or is it halted?  Is it always the same addresses that can not be read?

Lori

Hi Ralf,

I see that you are using a modified version of the cla_sqrt example from C2000Ware. Would it be possible to share your project, or at least the modified bits? If you wish to share privately, please start a private E2E conversation with me.

Thanks

ki

Dear ki and Lori,

 first I want to answer the questions:

 - Yes, the C28x runs when the problem occurs

 - Yes, the read problem of the CLA occurs at the same adress (0xa06e).

 In order to fix the problem I have restartet my notebook and up to now the problem does not happen. I´m not sure if this is the final solution of the problem. Attached you will find the code based on cla_sqrt example.

Best regards

Ralf

cpu01.zip

Hello Ralf,

Thank you for sharing your test case. I tried this with CCS 9.3.0 and CCS 10.1.1. In both cases, I don't see the issue with the opcode in the Disassembly view. I do see the same errors regarding the reading of MR* registers however:

Ralf Bartling said:

 In order to fix the problem I have restartet my notebook and up to now the problem does not happen. I´m not sure if this is the final solution of the problem

The problem you are referring to is just for the opcode issue, correct?

Hello ki,
yes, that´s right. My problem is referred to the opcode issue (see attachment).
Unfortunately the local variables, which are realized in MRx registers, are not shown in the "variables" window, but it is possible to see the values of MRx in the “registers” window. During debugging is would be helpful to change the values of MRx in the "registers" window, but this is not possible.
For later versions of CCS I suggest to show the values of MRx in the "variables" windows with the possibility of modification.

Best regards
Ralf