[PATCH] [MTD] OneNAND: Do not allow oob write past end of page

Artem Bityutskiy dedekind at infradead.org
Mon Feb 5 03:30:39 EST 2007


On Mon, 2007-02-05 at 16:44 +0900, Kyungmin Park wrote:
> I mean that in the view point of software OnenNAND is NAND, but the one of
> hardware it is NOR.
> So we don't need to work it as NAND. I think we don't need to restrict the
> OneNAND

OneNAND stuff which we have now is _part of MTD_. MTD has its interface,
its model of the device. As long as we are part of this infrastructure,
we have to obey MTD's interface, rules and restrictions.

The whole idea of MTD is to make upper layers to access flashes in a
uniform layer. MTD achieves this at certain extent. There are flaws, but
anyway. So basically OneNAND implements more then NAND in this OOB
handling stuff. And it is bad. It encourages other people to write
OneNAND-only software, incompatible with NAND, which is totally
unreasonable in this case.

-- 
Best regards,
Artem Bityutskiy (Битюцкий Артём)





More information about the linux-mtd mailing list