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.

AM2732: Reading some addresses of MSS_VIM_R5A causes interrupts to stop working

Part Number: AM2732

Hi, I am using timer that frequently generate interrupts on AM2732. Everything works just fine, until I open Memory at 0x02080000 (MSS_VIM_R5A ) or perform any other type of read on that address. It looks like the addresses 0x02080000 - 0x02080078 are somehow protected. Can anyone provide any useful information, what could be the cause. After that it is not possible to enable interrupts back, until the reset.

Thank you and best regards 

  • Hi, We are looking in to it. Will update here on the progress by 09/29. Could you please provide more info on SDK version/application that you are using ?

    Best Regards, Shiv

  • Hi, I am using my own bare-metal application (initialization of timer for interrupt generation and GPIO initialization for toggling when interrupt occur), debugging with iSystem tool. Everything works as it should until the read in that area, after that code is executed normally, just interrupts stop working.

    Best regards, Jaka

  • Hi Jaka,

    I tried recreating the issue that you're facing with a simple project triggering an interrupt whenever the RTI reaches a certain count. It works fine even when I open a memory browser at the addresses 0x02080000 - 0x02080078. I tried reading from individual addresses in this range, and also the whole range that you've mentioned.
    Could you please clarify the following:
    1. Is this issue seen when you read from any of the individual addresses in the range that you've mentioned, or only when you read from a certain range?
    2. Have you used any debug statements like Debug Log or printf for reading those addresses? I see that the interrupts get disabled while using printf sometimes, which I'm still debugging.
    3. Could you also please verify if you have placed the Timer module in "STALL MODE OFF" and the timer is running when you've paused in the breakpoint for reading the memory, and hence missing the interrupt event?

    I will keep you posted with any further progress.

    Thanks,
    Sahana

  • Hi Sahana,

    This issue occurs in both scenarios, reading from individual address or reading a certain range. If I stop the core, read something in that address range and then run the core again, the interrupt is not generated anymore.

    For reading I use memory read window (also tried with forced align), no debug log or printf is used in application.

    The timer is in stop mode, when the CPU is stopped in debug mode. RTIGCTRL -> COS is 0 (Counters are stopped while in halting debug mode.). So that means that interrupt is not missed.

    What debugger you are using? Can you provide your test application.

    Thanks,
    Jaka

  • Hi Jaka,

    I've shared the project with you, please let me know if you can reproduce the issue on this project as well. I'm using XDS110 with Code Composer Debug tool for debugging. Could you also confirm what happens when you read into any other address (say 0x02F7A000)?

    Thanks,

    Sahana

  • Hi Jaka,

    Have you made any progress on this? 

    I'm marking this as resolved, as there has been no communication for a long time. We are waiting on your response, please let us know if you need anything else from us.

    Thanks,

    Sahana

  • Hi, thank you for sharing the project. Tried with other timer for generating the interrupt, but the same is happening. Reading any other address (for example 0x02F7A000) is not causing this issue, interrupts still work.

  • Hi Jaka,

    Thanks for your response. Could you also reproduce this using the Code Composer Studio IDE and Debug tool to rule out that it is not a Debugger specific issue, as I'm unable to see the issue that you are facing when I try it on my side.

    Regards,

    Sahana

  • Hi Jaka,

    Could you please let us know if you've had the chance to try out the suggestion above?

    Thanks,

    Sahana