[PATCH/RFC] MTD: Striping layer core

[Josh Boyer]

On 4/3/06, Jörn Engel <joern at wohnheim.fh-wedel.de> wrote:
> On Mon, 3 April 2006 15:06:19 +0200, Jörn Engel wrote:
> >
> > As a step in that direction, please take a look at this patchset:
> > http://wh.fh-wedel.de/~joern/mtd_type.tgzwh.fh-wedel.de/~joern/mtd_type.tgz
>
> Taking a closer look at the code, reader should notice three
> interesting things, btw:
>
> 1. Nico is a lazy moron,
> 2. Artem is a lazy moron and
> 3. Jörn is a lazy moron.

There should also be a Josh is a lazy moron.  I merged the NOR ECC
stuff in JFFS2.

> Nico, you didn't notice that the Sibley flashes are just another
> iteration of ST's ECC NOR flashes.  Special code for both flash types
> is nearly identical, except for the field used to store the writesize
> in struct mtd_info.

I've been thinking about that for a while now.  Real life has
intervened for the most part.

>
> Jörn, I didn't notice that ECC NOR has the exact same cleanup routine
> as NAND flash.  And for the initialization code, that could be shared
> as well.

And a large majority of the ECC NOR stuff could have just been merged
into cfi_cmdset_0001.c instead of writing a whole new file.  Indeed,
it really has to be because the larger 16MiB chips _report_ a command
set of 0001 even though they aren't.  I have code that does all this,
but it's based on old MTD and wants sanitizing.

> Nico, can you take a close look at the Sibley part of the patches and
> checker whether I broke anything?  Combining ECC NOR and Sibley was
> the low hanging fruit, so I already picked it.  Combining the three
> remaining sets is harder.  Maybe someone else is feeling the urge to
> combine those.

See above :).  I'm off traveling for work at the moment, but I'll try
to look at the patchset at some point next week.

josh