• State Resolved
  • Locked Locked
  • Replies 7 replies
  • Subscribers 83 subscribers
  • Views 689 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.

TMS320F280041: Memory export error on Sector 8 (bank 0) and OTP issue

Wayne Huang
Genius 16395 points
Part Number: TMS320F280041
Other Parts Discussed in Thread: UNIFLASH

Dear Champs,

I am asking this for our customer.

The user could not use Uniflash or CCS to read the sector 8 starting from 0x88000.

See below figure.

It seems it has something to do with OTP because we found Z2_OTPSECLOCK is not same as normal.

They do not intend to use any CSM password, but they did intend to modify the boot mode pin from GPIO24/32 to Zero Boot Mode Select Pins and always boot to flash.

See below OTP registers

See Z2_OTPSECLOCK

Questions:

1) Their OTP setting is as below. Is there anything wrong?

In f28004x_codestartbranch.asm

In the .cmd file,

2) Is the weird Z2_OTPSECLOCK the reason that memory 0x88000 (sector 8) cannot read? Or is there any other cause?

Wayne Huang

  • 0
    TI__Guru** 115411 points

    Hi Wayne,

    No, OTPSECLOCK is correct and that should not have any impact on accessibility to Flash sectors. It is showing "????" in CCS memory watch window which means CCS is not able to read these location. What happens if you try to read these locations in code ? But it is strange that user is not programming the OTPSECLOCK field but still it is programmed.

    Can you check the flash ECC error status register to see if ECC error flag is set and if yes, what is the ERROR address ?

    Regards,

    Vivek Singh

  • 0 in reply to Vivek Singh
    TI__Genius 16395 points

    Dear Vivek,

    Please see the flash related registers below:

    FLASH_ECC_REGS

    Flash control register

    Also, note that the user originally used CCS 9.1 and found CCS 9.1 was not able to read out the flash sector 8 (0x88000). But today, they found they could use CCS 9.3 to read the same flash memory out.

    Please see the below figures.

    Questions:

    1) From the above flash ECC register, FLASH_ECC_REGS, it appears there was uncorrectable ECC errors, right?

    Do you have any comment? Was this the cause?

    2) Why could CCS 9.3 read it out, but CCS 9.1 couldn't? Is this a CCS 9.1 bug?

    3) We still could not relate these to the wrong OTPSECLOCK that I mentioned in the last post. Do you have any idea?

    4) What should the user do next to solve this? Or is there anything else you want the user to confirm?

    Wayne Huang

  • 0 in reply to Wayne Huang
    TI__Genius 16395 points

    Dear Vivek,

    Do you have any update?

    Wayne Huang

  • 0 in reply to Wayne Huang
    TI__Guru** 115411 points

    Hi Wayne,

    1) From the above flash ECC register, FLASH_ECC_REGS, it appears there was uncorrectable ECC errors, right?

    Do you have any comment? Was this the cause?

    Yes, there is ECC error for USER OTP setting for Zone2 and this is not good.  Not sure why that is happening if user is not programming that field. In general that should not cause issue with read of specific flash sector. That will cause issue while running any test on this device. Does this issue happens only after programming the settings in USER OTP ?

    2) Why could CCS 9.3 read it out, but CCS 9.1 couldn't? Is this a CCS 9.1 bug?

    I am not aware of any issue like this on CCS9.1 but I personally have not tried CCS9.1 version. Let me know if you see this issue on CCS9.1 on even a fresh device (not programming anything).

    3) We still could not relate these to the wrong OTPSECLOCK that I mentioned in the last post. Do you have any idea?

    Look into your map file if something is mapped in that space? If yes, then you may be missing DSECT for that section in cmd file.

    Regards,

    Vivek Singh

  • 0 in reply to Vivek Singh
    TI__Genius 16395 points

    Dear Vivek,

    Let's talk this offline with some more information.

    Wayne Huang

  • 0 in reply to Wayne Huang
    TI__Guru** 115411 points

    I found your offline email on this and replied to this. Sorry for late reply.

    Regards,

    Vivek Singh

  • 0 in reply to Vivek Singh
    TI__Genius 16395 points

    Dear Vivek,

    1) The original issue about reading memory address 0x88000 (????) is related to CCS and Uniflash. With newer CCS 9.2 and Uniflash 6, then this issue was gone.

    2) For the wrong FLASH_ECC_REGS and PSWDLOCK issue, it was no longer seen after they tested with newer samples. Now, it seems everything is good. The issue could not be reproduced.

    Therefore, this issue can be closed.

    Thank you for your support anyway.

    Wayne Huang