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.

dm365 and MMC interface performances

Hi all,

looking at the SDMMC application note I read that dm365 (TMS320DM365ZCE27, i.e. silicon rev 1.2) can communicate using 1 bit of data and 20Mhz of CLK and that it is compatible to MMC V3.31.

I have a custom board with inside an eMMC v4.3 and it works well with linux kernel arago r37 using 4 bit data and 50Mhz CLK.

Maybe the dm365-SDMMC application note is to be updated?

Could I have some bugs going on with 4 bit data and 50Mhz CLK?

Any info is appreciated.

  • Hi Raffaele,

    TI did validate the eMMC card with DM365 ROM however I think the validation was done without Linux/kernel. 

    My understanding is that MMC3.31 in DM365 will work with eMMC cards 4.x when these cards are used as mass storage devices (for read write). The MMC version 3.31 might not support some of the new features introduced in eMMC 4.x versions.

    yes, we can look into the DM365 - SDMMC app note to see it can be updated to add the above info soon.

    Hope this helps.

    Thanks,

    Tai Nguyen

  • Hi Tai,

    The purpose is to use MMC chipsets for reading and writing.
    I hope to share my experience with others in order to be more confident on this configuration.

    Thank you,
    Raffaele Recalcati

  • Hi Tai,
    it is possible to know when your validation for eMMC interface used as mass storage with 4 DATA pin at 50Mhz will be done.

    It is really important to know that, because the performance difference versus 1 DATA pin and 20Mhz is very big.

     

    Thx,
    Raffaele

  • Hi Raffaele,

    Sorry for the delay. I just got the reply from the ROM verification person and he indicates that the verification (50Mhz) has not been done. I will update you once I have the data.

    Thanks,

    Tai 

  • Tai Nguyen said:

    Hi Raffaele,

    Sorry for the delay. I just got the reply from the ROM verification person and he indicates that the verification (50Mhz) has not been done. I will update you once I have the data.

     

    Don't worry!

    What do you think about the possibility to do a validation test with your help and our hardware?

    Your help could consist in performance validation scripts and our hardware can offer at least 4 four different chipsets with inside v4.3,v4.4 eMMC.

    I'm offering to be a beta tester of this solution.

    We are working now on this issue and so if you point us to the right direction I can give you the results of endurance, stress, ... tests.

    The kernel is modified from dvsdk4.01, so 2.6.32 is a little bit old for v4.4, trying to fix it with backporting from 2.6.37, but u-boot can be used to do tests.

    I could also use blktrace or lauterbach or what  you want to validate dm365 with 4bit data and 50Mhz.

     

    Let me know what you think about it,

    Raffaele

  • Raffaele,

    Just want to let you know that I am waiting for the one who did the MMCSD testing to confirm if indeed he did test the eMMC in 4-data bit mode at 50Mhz. It was a little confusion before because there were two persons did the MMCSD testing. One reported that she did test 4-data bit mode at 50Mhz using a MMC card but not using the eMMC card and then the person who did test with an eMMC card claimed that he did use the same drivers and tested with an eMMC card. So I am not 100% sure but I think we had tested it in 4-data bit at 50Mhz with an eMMC card. Once I hear from the colleagues I will update the post if it is not the case.

    Thanks,

    Tai

  • Tai Nguyen said:
    Once I hear from the colleagues I will update the post if it is not the case.

     

    It should be nice to have a confirm about the eMMC dm365 interface working at 4bit/50Mhz.
    I'm sorry to be annoying, but you write that you post only if you haven't tested it, and so I don't know how much time I have to wait.

    Regards,
    Raffaele

  • Raffaele,

    Sorry for the delay. I finally get the response from the person who tested the MMCSD. He confirmed that he had tested the MMCSD with an eMMC card at 4bit/50Mhz using the DM355 silicon. He said it should work on the DM365 since the DM365 MMCSD module is identical IP design as the DM355.

    Regards,

    Tai Nguyen

  • Tai Nguyen said:

    Raffaele,

    Sorry for the delay. I finally get the response from the person who tested the MMCSD. He confirmed that he had tested the MMCSD with an eMMC card at 4bit/50Mhz using the DM355 silicon. He said it should work on the DM365 since the DM365 MMCSD module is identical IP design as the DM355.

    Regards,

    Tai Nguyen

    Hi Tai,
    that's quite nice, but I try to have something more...

    Are you maybe planning to validate also dm356 in an ufficial way modifying also the documentation?
    At the moment the dm365 mmc interface is compliant to v3.3 and now the normal eMMC are instead v4.4.
    Maybe you have a roadmap of this possible update?

    Regards,
    Raffaele

  • Hi Raffaele,

    Unfortunately we are not planning to validate the MMCSD on a DM36x silicon since the DM35x and DM36x MMCSD modules/hookups are identical.

    I am not aware of any upgrade on the module and I am also checking with the IP design team to see if there is any roadmap for this module. I will let you know if I hear anything.

    Thanks and regards,

    Tai