• State
  • Locked Locked
  • Replies 5 replies
  • Subscribers 99 subscribers
  • Views 500 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.

SD Card is not detected by DM8107 when it is reset by software

Faiyaz Pisuwala
Prodigy 60 points
Other Parts Discussed in Thread: DM8107

Hi,

We have designed a hardware in which SD card is connected on SD1 interface of DM8107.  We have configured DM8107 for Boot mode sequence "NAND, NANDI2C, MMC, UART".  Now, by mistake, we have connected ball # AA29 of DM8107 to SD_CMD signal ; while ball # Y29 should be connected to this signal to support MMC boot mode.  So, MMC boot mode does not work in our hardware but SD card does get detected and works fine after DM8107 is booted. 

Now, the problem is when the SD card is already inserted in the system and working and when we reset the DM8107 through software kernel, SD Card does not get detected in the next boot of the system; neither at u-boot level nor at kernel level.  But when we remove the card and insert it again, it starts working.

We have observed an interesting thing that when we used a different boot mode which does not have MMC boot mode option then this problem got disappeared. 

Can somebody please throw some light on this strange behavior of the system?

Thanks & regards

Faiyaz Pisuwala

  • 0
    TI__Expert 7225 points

    Hi Faiyaz,

     

    Is this a major issue? Since the board is anyway not booted via MMC card, so the issue should not arise isn't it? All you have to do is to ensure no MMC boot option. Can you not release with this understanding ? I hope the next revision should have this hardware fixed.

     

    It seems like when the when MMC boot is used, the MMC/SD controller port is not reset correctly/left in an undefined state (due to the hardware bug). So it fails to get detected in due course as no action is taken by u-boot and kernel subsequently; However when reinserted the interrupt enables it to be detected.  

     

    This is expected as it seems.

     

    Is this on RDK 4.0 - which version?  Also does it happen even if the hardware is power cycled? Presume this is Safe Vision project(?)

     

    Please post responses of this also on e2e......

     

    Thanks.

    Best Regards,

    Feroz

  • 0 in reply to K Md Feroz Irfan
    Prodigy 60 points

    We cannot go ahead in releasing our hardware with no MMC boot option as we require it in our final product. 

    Not able to understand what you are trying to say in this sentence : "It seems like when the when MMC boot is used, the MMC/SD controller port is not reset correctly/left in an undefined state (due to the hardware bug). So it fails to get detected in due course as no action is taken by u-boot and kernel subsequently; However when reinserted the interrupt enables it to be detected."    As per our understanding, MMC/SD controller port inside DM8107 should reset itself once any reset (whether software or Hard) is provided to DM8107.

    RDK we are using is 03.50.00.05

    And yes, when we power cycle the hardware, SD card does get detected. 

    Could this be a SOC bug? If yes, Please advise us SOC revision with fixing of this issue.

    --Faiyaz

  • 0 in reply to Faiyaz Pisuwala
    TI__Expert 7225 points

    Hi Faiyaz,

     

    Is there not a hardware issue on the board as in the “Now, by mistake, we have connected ball # AA29 of DM8107 to SD_CMD signal ; while ball # Y29 should be connected to this signal to support MMC boot mode.” Since anyway you will be fixing this in hardware as MMC boot its needed for final product, this side effect related to the soft reset should be irrelvant. Isn’t it?

     

    Or do you suspect this issue is totally different and indpendent of this board bug? Please let me know…

     

    My suggestion/hypothesis springs from the fact that you will anyway fix the board bug and enable MMC boot anyway. That should resolve the soft reset side effect too in all possibility. So, surely I am not suggesting to release w/ the bug !

     

    If you fix the board hardware issue and still the issue is seen then we can suspect the SoC and dig further. As of today there is no known issue like this in errata. PFA the latest one. You may want to cross check on ROM code/ boot related Advisory 1.1.88 Control Module, Pin Configuration (PINCNTLx): ROM Modifies Bit 19  if that applies here and if the workaround is indeed implmented in the code to be on safer side

     

    Please let me know if I misunderstand anywhere…

     

    Thanks.

    Best Regards,

    Feroz

  • 0 in reply to K Md Feroz Irfan
    Prodigy 60 points

    Hi Feroz,

    Yes, there is a hardware issue but that is related only to the MMC boot option and that is we are not able to boot from MMC, but once our board gets booted; SD card interface does work fine.

    So, i think the current issue may be totally different than the hardware bug. 

    And what is the guarantee that if we resolve the bug in hardware, the current issue will also get resolved?

    Also, we cannot release our hardware without solving this issue as our next hardware release will be our production release.

    Thanks
    Faiyaz

  • 0 in reply to Faiyaz Pisuwala
    TI__Expert 7765 points

    Hi,

    I tried this out on 8107 EVM. I tried with RDK 3.5 prebuilt SD boot images. I’m able to boot using SD card with power reset on board (not switch). I verified 5-6 times, it boots up fine. Note following setup

     Images – RDK3.5 prebuilt

    SD card – SanDisk SDHC 4 GB class 4

    Formatting utility – HP USB disk format too version 2.0.6 (this is important)

    I tried mounting SD card everytime after I did power reset. I also checked this with reboot command and SD card is getting detected everytime I reboot.