DM355 boot issue

Onoda,

Can you please provide more information:

1. Silicon Revision fof DM355.
2. Are all the DM355 same Rev
3. NAND part no
4. Version and source of UBL and flasher

Thanks,
Gaurav

_{We also have the problem like this. New UBL can solve this problem greatly. However we got three returned failed DM355 product from customers.After check, all of them had this kind of error}

_{messages and the Jtag can't be connected to boards. These boards were tested ok many times in factory before they were shipped to customer. It really was really hard to find what caused this.It looks like the board or the chip is very sensitive to certain variables.I am frustrated by this.} 

Awen,

I suppose this issue is now being discussed at another forum post:

https://e2e.ti.com/support/dsp/davinci_digital_media_processors/f/99/p/30619/216595.aspx#216595

Thanks,
Gaurav

Hello Chris,

I have tried the fix you suggested and am delighted to say it works great in dealing with the small percentage of failures we see in our product.  However, I'd like to follow up on this by asking for the rationale behind this fix.  If I interpret the code literally, I understand that this code sequence will pull the DM355 memory controller into reset and then out, which I guess will put the controller into a known "clean" state.   However, your datasheet also suggest that this is supposed to be the default state of this module anyway, when the DM355 comes out of reset.  So why now does this reset have to be done explicitly by the UBL code?  Is this a bug in the chip?

Regards,

WM