[U-Boot] Q: Bad bits in Nand Write

[Ran Shalit]

On Thu, Feb 16, 2012 at 9:21 PM, Scott Wood <scottwood at freescale.com> wrote:
> On 02/14/2012 09:05 AM, Ran Shalit wrote:
>> Hello,
>>
>> I am using Micron Nand MT29F1G08ABC , with OMAP-L138 board, and I
>> write the Nand with the following sequence in U-Boot:
>> 1. erase block
>> 2, read complete page change the first byte and then write back the
>> complete page (2048 bytes)
>> 3, read complete page change the next byte and then write back the
>> complete page (2048 bytes)
>> 4. back to 3. until page is written completely
>>
>> What I see is that after about ~1200 bytes, bytes which are after the
>> last offset such as 1400 and 1500 offset are not 0xff but instead one
>> of the bits is zero (such as 0xbf)
>> What Do you think might be the problem here ...
>
> Read the manual for your NAND chip -- this is almost certainly violating
> the chip's specs.  There's a limit on the number of partial programmings
> that you can do to a page in between erases.  1, 4, and 8 are common
> limits.  2048 is not.
>
> -Scott
>
Hi Scott,
I checked and it is truely said in the manual, and no one was aware
about this limitation (4 page writes) in the design stage.
It is also interesting to add that the manual says that randsom pages
write is prohibeted (only sequential pages). I understand from this
that it is allowed to write the same page 4 times before moving to the
next (sequential) page.
Thank you! Ran