[U-Boot] NAND flash question

[Peter Barada]

On 09/20/2013 11:27 AM, ANDY KENNEDY wrote:
>> -----Original Message-----
>> From: u-boot-bounces at lists.denx.de [mailto:u-boot-bounces at lists.denx.de] On Behalf Of Peter Barada
>> Sent: Thursday, September 19, 2013 3:26 PM
>> To: u-boot at lists.denx.de; Peter Barada
>> Subject: Re: [U-Boot] NAND flash question
>>
>> On 09/19/2013 04:04 PM, ANDY KENNEDY wrote:
>>> All,
>>>
>>> We have a design that has NAND as a secondary device (not the boot
>>> device).  The last four pages of the NAND flash are reported as bad.
>>> Should this be true for all NAND flash devices we have?
>>>
>>>
>> No, I wouldn't think so.  Manufacturers qualify their parts and mark bad
>> blocks found during qualification testing.  Most data sheets indicate
>> that the number of bad blocks marked bad during manufacturer is below a
>> set percentage(if above thent he part is rejected).  Some parts that are
>> meant to be used during boot (such as NAND meant for OMAP parts) will
>> have certain blocks that are guaranteed good (i.e. the boot blocks).
>> Past that bad blocks could be anywhere in a particular device.
>>
> Okay, well, next question:  So on EVERY unit we have designed with a
> NAND flash, when u-Boot reads the NAND flash it reports that it couldn't
> locate a bad block table and states that it cannot read the last four
> pages of the flash.  Next, when one does a 'nand bad' on these boards,
> these last four pages are reported by u-Boot as bad.  Looking at the
> code, we believe that this is by design.  Is that correct?  Has anyone
> else done a 'nand bad' on a board and seen the same information?
>
> Thanks again for any information you can provide,
>
> Andy
If you are using a NAND-based bad block table (i.e. set the
NAND_USE_FLASH_BBT bit in nand->options), then the MTD layer will
internally use the last four blocks to hold the bad block table _and_
report those blocks as bad to prevent the user from accidentally
modifying them...



-- 
Peter Barada
peter.barada at logicpd.com