[PATCH] mtd: nand: fix byte addressing on 16-bit wide devices

[Ezequiel Garcia]

(Ccing Brian and Uwe)

On Mar 21, David Mosberger-Tang wrote:
> On Fri, Mar 21, 2014 at 12:05 PM, Ezequiel Garcia <
> ezequiel.garcia at free-electrons.com> wrote:
> >
> > You can take a look at these commits:
> >
> > 3dad2344e92c6e1aeae42df1c4824f307c51bcc7
> >
> > AFAIK, the READID command data is now issued as 8-bit, disregarding of
> > NAND_BUSWIDTH_16.
> >
> > Maybe you can re-test this building a kernel from linux-next.git and
> > report if your issues are now fixed?
> >
> 
> That's not sufficient.  GET_FEATURES and SET_FEATURES  also need
> byte-addressing.
> 

AFAIK, the issue for GET_FEATURES and SET_FEATURES was also fixed.
IIRC, it was done here:

commit 05f7835975dad6b3b517f9e23415985e648fb875
Author: Uwe Kleine-König <u.kleine-koenig at pengutronix.de>
Date:   Thu Dec 5 22:22:04 2013 +0100

    mtd: nand: don't use {read,write}_buf for 8-bit transfers

> No offense, but my patch seems both simpler and more complete?
> 

Could be.

Have you actually tested linux-next? It would be good to check if we still
have issues with 16-bit devices on the latest kernel, or if this is
already fixed. Last time I checked, my 16-bit device was ONFI-detected.

Brian: Related to this, the current code for fixing READID on 16-bit devices
seem to apply *only* for drivers that don't override cmdfunc(), right?
So each of the drivers that have their cmdfunc() currently need some
tweak to support 16-bit devices.

Or am I missing anything?
-- 
Ezequiel García, Free Electrons
Embedded Linux, Kernel and Android Engineering
http://free-electrons.com