• State Resolved
  • Locked Locked
  • Replies 4 replies
  • Subscribers 98 subscribers
  • Views 853 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.

NDK Issues: Hello World for OMAP L138 EVM using C6748 SOM

DanB57495
Expert 1340 points

Hello,

I am using a 30-day evaluation of CCS 4.2. My evaluation kit came with CCS3.3. I am Using the latest NDK Release ndk_2_20_03_24 and am using the NSP release: nsp_1_00_00_09.

I was able to load the Hello World program on to the target and get it running. The console displays:

*******************************************

TCP/IP Stack 'Hello World!' Application

 

Using default MAC address

Using MAC Address: 00-08-ee-03-14-99

MAC Address = 00-08-ee-03-14-99

******************************************

But I am never able to get an IP to display regardless of whether I set the Local IP or leave it 0.0.0.0 (for DHCP). When a halt the debugger (after directing CCS to the \nsp_1_00_00_09\packages\ti\drv directory) the EMAC_open() function in csl_emac.c. The program gets stuck on line 986:

ECTL_REGS->SOFTRESET = 0x01;

 

while (ECTL_REGS->SOFTRESET != 0x00000000); /* Wait until reset has occured */

I tried this with 2 Experimenter board sets with the same result. Does anyone know why we get stuck here?

I will be away for 3-days, but would love any feedback that can be provided in the meantime.

Thanks in advance.

Dan.

  • 0
    TI__Mastermind 45025 points

    Hi Dan,

    Are you running on the ARM or DSP?

    I see that you're using the  MAC address that's hard-coded in the example for when BSL (board support library) isn't present (in the file emacHooks.c):

    static UINT8 bMacAddr[8]= { 0x00, 0x08, 0xee, 0x03, 0x14, 0x99 };

    If someone else is already running any of the NDK examples on the same network, this could cause a conflict.

    can you try changing the above line to a different address?  E.g.:

    static UINT8 bMacAddr[8]= { 0x00, 0x08, 0xee, 0x03, 0x19, 0x79 };

    Then rebuild, run.  Does it get you further?

    If you still are having a problem, please zip up your entire *built* project directory and attach to this post.

    Thanks,

    Steve

     

  • 0 in reply to Steven Connell
    Expert 1340 points

    Hi Steve,

    I found the answer last night from the forum. I was not able to find the same link today to post it, but I believe you had commented in it.

    It seems that in CCS 4.2 that you need to launch your debug session from the C/C++ view Target Configuration TAB, by right-clicking on the configuration and selecting "Launch Selected Configuration", then connecting to target,  loading out file etc. I had started the debug session using Target=>Debug Active Project (in C/C++ view) and while the file was loaded and running, I believe either the reset or some other portion of the debug interface was not running correctly.

    The Hello World issued it's namesake using the Windows app and I was able to ping the device.

    Thanks for looking into it, I imagine I will be back.

    It is good to know about the MAC address issue as well.

     

    Dan.

     

     

  • 0 in reply to DanB57495
    TI__Mastermind 45025 points

    Dan,

    I'm glad that you were able to get past that issue.  Yes, when the CCS debug session is launched that way, it loads the program without performing a GEL reset, and this results in the problematic behavior that you ran into.

    I think this it's too easy for the user to run into this, so I'll bring this problem to the CCS folks attention.

    Steve

  • 0 in reply to Steven Connell
    TI__Mastermind 45025 points

    I filed a bug for this:

     

    SDSCM00039929 CCSv4 "debug launch" button lack of GEL reset leading to example failures, customer confusion