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.

TDA4VM: Lost connection and UDMA error

Part Number: TDA4VM

Hi engineers of TI,

We got a problem when using TDA4 cpsw9g.

SDK verson: ti-processor-sdk-rtos-j721e-evm-07_02_00_06

UDMA ERROR.

Some times when TDA4 initial, we got this udma error below:

Ethnet is connecting for 20 minutes, after that, it lost connection, when try “ifconfig eth0 up/down”, the command couldn't quit, and the serial port got stuck.

How to fix this problem?

Thanks!

Liyuan Wang

  • Hello Tanmay,

    I have unlocked this thread. Please feel to to respond on this when you are ready.

    - Keerthy

  • Dear Tanmay,

        Here are some details for this issue:

        1. This error happens while booting MCU2_0 and the probability of this is about 10%.

        2. Customer runs ETHFW, Camera APPs and encoder on MCU2_0.

        3. Firstly when they use our RTOS MCU1_0, this error will occur. Then they changed it to AUTOSAR, then it will not reproduce again. And then they add some new functions to AUTOSAR in MCU1_0 , it reproduces. So they assume it should result from configurations on MCU1_0, however they do not know how to config it.

    BR

    Sikai  

  • Hi Sikai,

    Sorry for the delay in this response.

    What is running on MCU1_0 in the RTOS case? Is it just Sciserver or another application is running with sciserver.

    Is it possible for you to share the structure "rmOptFlowReq" and "rmOptFlowResp" after return from "Sciclient_rmUdmapFlowSizeThreshCfg()" function at [PSDK-RTOS]/pdk/packages/ti/drv/udma/src/udma_flow.c:522.

    Also can you share the complete remote log?

    Ethnet is connecting for 20 minutes, after that, it lost connection,

    The connection you are talking about, is it CPSW2G (MCU CPSW) or CPSW9G (MAIN CPSW).

    Regards,
    Tanmay

  • Dear Tanmay,

         Thanks for your support.

         I have talked with customer and they will upload related logs later.

         Here are something I can update:

         1. There are other applications run on MCU1_0 rather than Sciserver. Customer will update details.

         2. We are focusing on MCU2_0 so it will be CPSW9G

         3. "Firstly when they use our RTOS MCU1_0, this error will occur. Then they changed it to AUTOSAR, then it will not reproduce again." I have checked with them again today, and they told me even change to AUTOSAR version the error still reproduce. It just has less possibility to reproduce it, which means this error may not resulted from MCU1_0.

    BR

    Sikai

         

  • Dear Tanmay,

         After comparison, they find this is caused by the function appLogCpuSyncInit from app_init.c

        

        If they comment this function this error will occur while if they did not do anything to this function it will work fine.

        They used to comment it because the boot process will stuck here. However now they uncomment it the boot will not stuck any more.

        Could you help analyze what did this function do?

    BR

    Sikai

  • Hi Sikai,

    Sorry for the late reply.

    Could you help analyze what did this function do?

    This function is basically syncing all the logs from ethfw with other cores. This should not be responsible for this issue.

    Is it possible for you to share the structure "rmOptFlowReq" and "rmOptFlowResp" after return from "Sciclient_rmUdmapFlowSizeThreshCfg()" function at [PSDK-RTOS]/pdk/packages/ti/drv/udma/src/udma_flow.c:522.

    Also can you share the complete remote log?

    Can you provide me with these logs.

    Regards,
    Tanmay