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.

TMS320F2812: Internal Flash acts partially locked (can view but not update) after failed image load using Code Composer.

Part Number: TMS320F2812

We are having an issue with a 2812 on a custom board we are developing.  We have been running the 2812 DSP chips from flash for some weeks.  Recently when load images into Flash memory of the F2812 using code composer 8.0.0 and 7.4 on Windows 7 the erase part of the load failed leaving memory with all zeros rather than the all ones it is supposed to have when erased.  When we have observed the issue Code composer stopped with an error stating it could not access the image.  This was in the middle of the erase rather than at the start of the erase like it usually does when updating memory. 

 

On the boards with the flash set to zero Code composer seems to be able to read data from the F2812 flash rather than reporting that the chip is locked.  Previously we had a flash problem updating flash from code on an eZdsp, it left the CSM locked and completely unaccusable.  With the previously locked chip Code Composer could not view memory at all.  Code composer does not seem to consider the chip as locked but the CSM region shows zeros as does the rest of memory and we get the generic con connected or locked message when trying to flash the chip.  Is there a way to recover from this?  The chip does not seem to be secured is it actually locked?  Since the libraries that update flash seem to write all of flash to various values when erasing a region of memory it seems like there needs to be a way for the hardware to distinguish between cleared memory in region A (where the keys are stored) and the chip being locked. 

  • Jerome,

    if you read from secure memory you see all zeros?

    what is the value of your CSMSCR.SECURE bit? You may have interrupted the flash while the device was erasing clearing or writing to the password locations. 

    If that is the case then there is nothing we can do, the device is simply locked.


    Regards,
    Cody 

  • Thanks Cody

    We do see all zeros.  I thought in the past CCS refused to read when locked rather than reading zeros. 

    So both code composer 8 and code composer 7 have stopped in the middle of the update locking the 2812s recently.  As a workaround to avoid loosing more chips I have been flashing on an old version of windows that is not networked using CCS 6.  I there a safe way to flash these chips in a more reasonable environment?

    Jerome

  • Jerome,

    i am looking into compatibility issues with CCS7/8 and F2812. Are you using the Flash API at all?

    Did you get a chance to check what the values of CSMSCR.SECURE of a 'damaged' unit is?

    Regards,
    Cody 

  • The area up near the end of region A with the keys were all zeros as was B. I am not able to look at the chip today.

    Our application uses the flash API. That has been working on the boards after they are flashed. I locked the 2812 on an eZdsp board months ago developing the update to flash but have not have issues with the application and flash otherwise.
  • One of the F2812 chips has already been replaced.  The on the other card the CSMSCR register reads 0x0047 after power up.  A working chip CSMSCR reads 0x0070.  So secure is set on the non working chip.

  • Jerome,

    OK, it does seem like the non-working chip is locked... We now know what is happening, I assume the important question is "Why?".

    When you are programming with CCS7.4 or CCS8.0.0 how often does the device become locked? Is it occasionally, does it fail every time? Can you provide the exact text from the error message?

    I am having someone within TI look into this issue, the answers to the above questions will be helpful.

    Regards,
    Cody

  • Jerome,

    additionally, when you updated to CCS7/8 did you change any hardware? Maybe you are using a new power supply or PC? For example, power instabilities or a bad USB port on a new PC could affect the flash programming.

    Regards,
    Cody

  • I assume your issue is solved. I am going to close this thread. If you issue is not resolved feel free to reply here or start a new thread.

    Regards,
    Cody