[PATCH v3] mtd: gpmi: fix the bitflips for erased page

[Huang Shijie]

On Mon, Jan 13, 2014 at 08:52:48AM +0100, Elie De Brauwer wrote:
> >> A second note, you patch v2 is a lie, there in the v1->v2 you mention
> >> you return the number of bitflips to the upper layer, however your
> >> code was not modified accordingly, gpmi_erase_check still does not
> >> return the number of bitflips. Unless I'm missing something there too.
> > I added a new argument @max_bitflips for gpmi_erased_check which will
> > return the number of bitflips.
> 
> True, but if you look just  below your patch you see that
> mtd->ecc_stats.correct will not
> get properly updated. It normally gets updated with the number of
> bitflips reported
My bad now :(

I really missed the mtd->ecc_stats.corrected.
thanks a lot for pointing this.


> through the status. (Well to be honest the entire switching between chunks and
> pages confused me a lot on this monday morning, it could be me but when
> iterating over chunks, switching to pages and breaking out of the for
> loop isn't
> something I would consider very clean, I think some more elaborate comment
If we do not read the whole page, we have to issue the change column comand 0x5
and calculate the offset of the chunk.

> on the erased check might help.).
ok. i will add more comment in the next version.

thanks
Huang Shijie