[PATCH 1/1] UBI

Josh Boyer jwboyer at gmail.com
Wed Jul 5 08:43:05 EDT 2006


On 7/5/06, Artem B. Bityutskiy <dedekind at infradead.org> wrote:
> Hello Josh,
>
> On Wed, 2006-07-05 at 06:36 -0500, Josh Boyer wrote:
> > Does the range for bad eraseblock reserve really need to be up to 1/2
> > the device?
> Probably no, which digit would you conceive?

Well, most ship from the manufacturer with up to 10% bad.  So taking
that as a starting point, I would claim that 25% bad is enough of a
problem to seriously think of replacing the chip.  However, it's all
arbitrary so I suppose it doesn't matter.

> > Why all the forward declarations of structures in the various header
> > files?  Couldn't this be done by properly including the header files
> > that are needed?
> Well, I deliberately used forward declaration as it is faster then
> including headers just for the sake of getting some structure defined.
>
> Why is it bad? Please, refer a particular example.

It's not particularly bad... it just looks unclean in my opinion.
Certainly not a show-stopper.

josh




More information about the linux-mtd mailing list