• State Resolved
  • Locked Locked
  • Replies 5 replies
  • Subscribers 98 subscribers
  • Views 594 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.

How to recover processor with CCS after very early crash

Dan Edmondson
Intellectual 860 points
Other Parts Discussed in Thread: AM1705

Hello,

Board is custom using AM1705 processor. It is close enough that I have been using the evmam1707 GEL file.

While adding some POST testing to the UBL code we introduced an exception very early in the boot process from NAND. Once the exception occurs I cannot access memory and cannot load nandwriter to recover this board. I believe the correct thing to do is to get the GEL file to correct the error condition. So far I have not been able to get the GEL loaded.

If I hit the reset button, the buggy code runs again before I can connect the debugger and so the exception gets re-introduced. I'm looking for a technique. :-)


Dan

  • 0
    TI__Guru**** 465851 points

    Hi Dan,

    Dan Edmondson said:
    Once the exception occurs I cannot access memory and cannot load nandwriter to recover this board.

    Do you lose your connection with CCS also?

    Dan Edmondson said:
    I believe the correct thing to do is to get the GEL file to correct the error condition. So far I have not been able to get the GEL loaded.

    What issue are you seeing when your try to load the GEL file?

    Thanks

    ki

  • 0 in reply to Ki
    Intellectual 860 points

    Hi Ki,

    I was loosing CCS connection but I found the button to keep debug going even if the program load fails. Here is a screen shot of where I get stuck. I have tried Abort and Force several times but you can see from the console output it hasn't been successful. This exception happens so early I haven't been able to get between the reset and the line of code causing this. I'm thinking I may need to change the boot target strapping so that it doesn't look to NAND and then would not crash. Make it look like board that hasn't been flashed at all.

  • 0 in reply to Dan Edmondson
    TI__Guru**** 465851 points
    Dan - sorry for the tardy reply.

    I'm not familiar with the error myself but when I search the forums I see some related issues in the AM1x forums. Some suggested turning down the JTAG TCLK settings if using an XDS100. Others have suggested potential configuration issues. I will move your thread to that forum where hopefully more people are experienced with your issue.

    Thanks
    ki
  • 0 in reply to Ki
    Intellectual 860 points

    Hi Ki,

    Thanks for re-posting this. I need to follow up and let you know that I did stumble upon recovering the board. I had stopped using the GEL file since our boards are all booting and being configured by boot code. Long story short, once I re-discovered the right place to configure the target configuration to use the GEL again the recovery went just like I expected. Maybe the experts in the other forum can spot something else I missing.

    Once the code had hit the exception, I could reset the processor with the emulator attached, then I could see the GEL played in the console. Once that happened the data bus hang was corrected and I could re-flash the board.

  • 0 in reply to Dan Edmondson
    TI__Guru**** 465851 points
    oh right! If you have target initialization done by some boot loader and then you launch the debugger with a target config that uses a startup GEL file that also does target initialization (holy run-on sentence batman!), then you would blow away the previous initialization and put the target in a bad state. Good catch!