What would cause large block of NAND to be marked as bad?

[Artem Bityutskiy]

On Mon, 2012-05-07 at 11:05 -0700, Subodh Nijsure wrote:
> Hello,
> 
> I am working with some prototype hardware and I have seen some weird 
> behaviour when testing power cut.
> 
> On my board I have Micron MT29F2G08ABAEAH4 part and now have three 
> boards on which close to 20-30 blocks have been marked as bad.
> 
> We mainly run UBIFS on these boards, if there was flash data corruption 
> we expected to see errors at UBIFS level i.e. not able to mount UBIFS 
> file system but we didn't expect NAND blocks themselves to be marked as 
> bad, due to power cut.
> 
> What would cause large number of NAND blocks to be marked as bad, due to 
> power cut?

Did you look at Mike Dunn's bitflip_threshold stuff? It is in my
l2-mtd.git tree. Look at this patch and the next ones from Mike:

http://git.infradead.org/users/dedekind/l2-mtd.git/commit/be1bb4ef1f98059f9d54623f3ed01423095fa967

Commit messages contain descriptions and references. May be for your
flash you need to set higher threshold?

-- 
Best Regards,
Artem Bityutskiy
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 836 bytes
Desc: This is a digitally signed message part
URL: <http://lists.infradead.org/pipermail/linux-mtd/attachments/20120507/8e135de2/attachment.sig>