• Locked Locked
  • Replies 4 replies
  • Subscribers 53 subscribers
  • Views 2128 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

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.

GEL Errors evaluating DNUM on C6678

Ralf Goebel
Genius 3715 points

Hi,

sometimes I get errors when using my GEL script and loading symbols to multiple cores:

C66xx_2: GEL: Error while executing OnFileLoaded(0, 1): cannot left shift by void     at (1<<DNUM) [VIB_C66xx.gel:1341]     at OnFileLoaded(0, 1) .
C66xx_4: GEL: Error while executing OnFileLoaded(0, 1): cannot test equality of void     at (DNUM==0) [VIB_C66xx.gel:1343]     at OnFileLoaded(0, 1) .

Are there any known problems when using DNUM inside a GEL script?

Thanks,
Ralf

  • TI__Guru**** 169864 points

    Ralf,

    All memory mapped registers must be explicitly defined in GEL scripts

    The core registers like DNUM are accessible from the GEL environment (check the attached screen), but that is only valid if the device configured in the debugger has them properly defined (or visible in the registers view), which is usually the case for CCSv5.1.x (with the ti_emupack update) and newer.

    If that only happens sporadically, it is possible that a specific core configuration or a pre-loaded running software is somehow preventing reading this register, thus causing the GEL script to fail. In this case I would try to manually connect to each core and run the GEL to see if this register is properly resolved.

    Regards,

    Rafael

  • in reply to desouza
    Prodigy 60 points

    Hi Rafael,

    This problem occurs only sporadically. Loading the symbols a second time after I get this error seems to work without a problem.
    All cores are connected, not running and reset (GEL_Reset()) before I start loading the symbols.

    CCS also crashes very often with my system. This could also be related to my GEL script. I'm trying to figure it out.

    Sometimes I get this error:
    C66xx_5: GEL: Encountered a problem loading file: .....out DWARF information is corrupted.
    If I don't restart the debug session immediately, CCS crashes when trying to reload the symbols.

    Any ideas how to solve this? I have a lot of CCS error reports, but I'm not sure if they are useful.

    Ralf

  • in reply to elfuhr
    TI__Guru**** 169864 points

    Ralf,

    Considering that you have enough memory and CPU speed to run CCS (2GB of RAM, a fast CPU) and since you have been getting some inconsistent behaviour when using CCS with your target hardware, I would try the suggestions in the sections General IDE and Debugger of the CCSv5 troubleshooting page to rule out any problems that may be coming from a pure software perspective.

    If you still see some inconsistencies, I would start investigating the JTAG and hardware details of your debug system - a good strategy to do that is described in the page Debugging JTAG Connectivity Problems.

    All that said, you are not the first one to report the corrupt DWARF information error message (check here); if you are using the same target board as the OP on the other thread, this indicates a trend and we will keep an eye on this issue and try to create a reproducible test case that exposes this bug.

    Hope this helps,

    Rafael

  • in reply to desouza
    Genius 3715 points

    This problem seems to be solved in Version 5.3.0 and CCS is running a lot more stable in general with our GEL script.

    Ralf