• State
  • Locked Locked
  • Replies 5 replies
  • Subscribers 82 subscribers
  • Views 662 views
  • Users 0 members are here
Support feedback
Options
Options
Related

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

F28332 Data verification failed during loading code in RAM

Michael Ivabov
Intellectual 300 points

Hello,

I'm using a custom board with F28332. On previous device i used F28335 with same program. Now I am getting some hardware mistakes (I hope that they are hardware).

Sometimes DSP reboots without any reason during work and it happens in random time. I've disabled WD, so it cant be it's work.

Sometimes during loading code CCS says - data verification failed during loading at address 0x8xxx (I saw it at two different addresses). I switched boards (so DSP's too), but errors dont stop. It's RAM, so this error is really serious. When I load my code in flash and start from it, it will copy some functions to flash, and what if it copies them with errors?

Yesterday my program wrote dummy data in one of my variables, went to ILLEGAL ISR and rebooted. Now it does not, but I'm not sure what I've changed.

So, I really don't know what to think. I'll try board with 28335 on next Monday. Can such problems be caused by bad contact in supply connectors? Maybe i should try changing emulator freq? I'm using SAU510 USB ISO. The lot trace code of DSP's is CA-8AAFJXW.

 

  • 0
    Expert 1070 points

    did you tailor the linker command file for the new chip? Is your code size too large to outfit the available memory on this 28332?

  • 0
    Expert 1720 points

    Hi Michael,

    I have few debugging questions for you, which is similar to the previous user's question:

    1. Is the command file of F28332 used and compiled and linked with no errors.

    2. Are you loading the compiled build used for F28335 as is to load in the F28332.

     

    The questions i asked because, there is quite a few difference in the amount of memory alloted to each of them.

     

     |DSP  | F28335| F28332|

    |Flash: | 256K    | 64K |

    |RAM    | 32K      | 24K |

  • 0 in reply to B M Mohan
    Intellectual 300 points

    Hi all,

    Sorry for the delay in my answer.

    Asking on your questions - it compiles with some warnings like unused vars and "file ends without of new line", but it's ok. I've added to the project 28332 RAM/FLASH files from examples. Now it's not important, because I think I've found a source of the error.

    I use all timers and their interrupts and sometimes I have to make requests to VFD. I get answers via interrupt, and after some time - about 30 seconds - my program crashes. SCI-a to VFD is pretty slow - 9200. I also found that if I forbid the SCI RX isr, program sends requests and works well. My stack is 0x500. What I am trying to say is that my program works fine for some time, gets correct answers from VFD and after a while it starts to rewrite even constant vars and goes to some unknown ISRs like RTOS and XINT3. I can choose this "living time" by changing the frequency of sending requests to the VFD, but I can not be sure that it will not crash in a lower freq.

     

  • 0 in reply to Michael Ivabov
    Intellectual 300 points

    eventually I'found the mistake. It was in the part of code, which is working with array and sometimes the counter falls out from borders and rewrites  some memory. Miracles doesn't happen..

    But I'm not sure about Data Verification at all.. Since I've lowered the emulator freq I've never seen it, but from the other hand I've never seen it before current project. Earlier I use the 28335 cmd file for both 335 and 332 DSP's and all was ok.

  • 0 in reply to Michael Ivabov
    Expert 1720 points

    Hi,

     

    Is the custom board contain Dual DSP's.

    In that case it can be because of other DSP.

     

    Regards,

    Mohan