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.

Error logs when I try to connect to ti816x

Hi All,

I've been trying to connect to ti816x with XDS 100v2 USB emulator and have seen the following errors all the time when I ran "Launch Selected Configuration".

CortexM3_RTOS_0: This Target Does Not Support The GEL_DriverString() Function
CortexM3_ISS_0: This Target Does Not Support The GEL_DriverString() Function
CortexM3_secSS_0: This Target Does Not Support The GEL_DriverString() Function

These errors were printed out several times, then "XDS100v2 USB Emulator_0/CortexA8" came up in Debug view.
Does anyone know what these mean and how to solve ?

I'm using CCS 5.0.1 and Host PC is Win7. Of course, I chose the "ti816x" when I made a target configuration file.

  • Hello,

    Can you please try to reset your perspective (right click on the "CCS Debug" tab in the upper right corner and select "Reset") to ensure all non-default views are closed, and then you can start re-opening views to see which view triggers the problem?

    Regards,
    Marc

  • Hi Marc,

    I'm not sure what exactly did you ask to me but I got the following perspective when I did "Reset" and "Launch Selected Configuration" again. 

    I can connect to the target board and flash the u-boot and other binaries using XDS 100v2 even though the errors are still seen but it took almost 15 seconds.

    Thanks,
    Wonhong

     

  • Mark,

    Could you please capture the debugger diagostic logs and attach them to this thread?  The proceedure is described here: http://processors.wiki.ti.com/index.php/Troubleshooting_CCS#Debug_Server_Logging

    Thanks,

    Darian

  • Hi Darian,

    I've attached the log captured when those errors were seen.


    Thanks,
    Wonhong

    7838.DebugView.zip

  • Wonhong,

    It looks like you attached the debug view executable and not the data it captured.  Once you've captured the data in debug view, you need to save it to a file and attach the file you specified.

    Darian

  • Darian,

    Sorry for confusing. I had attached the wrong zip file. Here's the log.

    Wonhong.

    2620.dbgview.zip

  • Wonhong,

    The GEL_DriverString message you are seeing is part of the status check that is perform on all Cortex M3 devices. The error message is printed because not all Cortex M3 devices support checking status using GEL_DriverString.

    I will investigate an improvement on this process by only performing the check if the target supports it to prevent the error message from appearing; but please note that these error messages should not affect any functionality with CCS or the debugger.

    Please let me know if you have any additional questions.

    Thanks.

    Ricky

  • Hi Ricky,

    As you mentioned, I've been using CCS without any issues in spite of these errors. However, because of the errors it seemed that "Launch Selected Configuration" took more time than I expected. So I just wanted to know if there is a way to eliminate these errors.

    Thanks for your helps and please let me know when you get the improved version of CCS.

    Wonhong

  • Wonhong,

    I too ran into this same error, which was causing my debug launch time to stretch to about 1 minute.  My workaround was to edit the debug configuration (the ccxml file) by going to the advanced tab and marking all the M3 cores as bypass (I don't plan to connect to and debug any of these core, so I'm hoping this doesn't alter the behavior of CCS or connecting to the other cores).

    This may help until an official solution is available.

    Regards, Daniel

  • Wonhong, Daniel,

    These errors shouldn't be what's causing the slowdown.  Bypassing targets will increase the launch speed, and it's safe to do, but if you see a noticable change with bypassing certain targets then that's likely a bug.

    If you can re-capture the debugger logging data that I described earlier in the thread (perhaps one with the M3's bypassed and one without) I'll take a look as to where the time is being spent.  I'll need the full log from the moment you hit launch until the moment it appears to be complete - the logs attached to the thread right are only some of the launch, not the entire launch.

    Darian

  • Darian,

    The delay seems to be because the function is called 11 times (per M3 core, though for my attached logs I had only one M3 not in bypass) during the debug launch, with a ~5 second spacing between calls:

    CortexM3_RTOS_0: This Target Does Not Support The GEL_DriverString() Function
    CortexM3_RTOS_0: This Target Does Not Support The GEL_DriverString() Function
    CortexM3_RTOS_0: This Target Does Not Support The GEL_DriverString() Function
    CortexM3_RTOS_0: This Target Does Not Support The GEL_DriverString() Function
    CortexM3_RTOS_0: This Target Does Not Support The GEL_DriverString() Function
    CortexM3_RTOS_0: This Target Does Not Support The GEL_DriverString() Function
    CortexM3_RTOS_0: This Target Does Not Support The GEL_DriverString() Function
    CortexM3_RTOS_0: This Target Does Not Support The GEL_DriverString() Function
    CortexM3_RTOS_0: This Target Does Not Support The GEL_DriverString() Function
    CortexM3_RTOS_0: This Target Does Not Support The GEL_DriverString() Function
    CortexM3_RTOS_0: This Target Does Not Support The GEL_DriverString() Function

    Hopefully the attached logs show what you need.  The time to launch the debugger increased from about 15 seconds to about 60 seconds by having one M3 core not in bypass.

    Regards, Daniel

     

    Debug_Logs_TI861x_M3_Issues.zip
  • Daniel,

    Your workaround works for me too. Launching debug tab took just less than 1 sec.
    Thanks for your help :)

    Regards,
    Wonhong

  • Wonhong, Daniel,

    It does sound like there is a problem with the launch time. I will need to investigate on the cause and provide a fix for this for future CCS releases.

    I have filed a bug (SDSCM00039361) for tracking this issue. Please track the progress using our public bug tracking system if interested. But the current workaround for the bug is the solution provided by Daniel; ie; bypass the M3 cores if not debugging them.

    Let me know if you have any additional questions.

    Thanks.

    Ricky

  • Ricky,

    I saw in the bug report that you listed 4.2 as the CCS version.  I'm using 5.0.1 - not sure if that matters, but I have not tried 4.2 so I don't know if the issue is there.

    Regards, Daniel

  • Daniel,

    It is true that I have not verify the bug on CCS4.2 yet either, but the affected code would be the same between CCS4.2 and CCS5.0, which is why I have tagged it for CCS4.2. Once I have finish my investigation, I will commit the fix to the appropriate CCS versions.

    Regards,

    Ricky