Re: [PULL] mtd: nand: changes for 4.7

[Boris Brezillon <boris.brezillon@free-electrons.com>]

Hi Brian,

On Thu, 5 May 2016 11:38:15 -0700
Brian Norris <computersforpeace@gmail.com> wrote:

> Hi Boris,
> 
> On Fri, Apr 29, 2016 at 10:13:48AM +0200, Boris Brezillon wrote:
> > Hi Brian,
> > 
> > This is my first PR for the NAND subsystem, so don't hesitate to tell me
> > if I did something wrong ;).  
> 
> Overall things aren't too bad, but I do have some comments.
> 
> > This pull request contains the following infrastructure changes:
> > * introduction of the ECC algo concept to extend the ECC mode one
> > * replacement of the nand_ecclayout infrastructure by something more
> >   future-proof.
> > * addition of an mtd-activity led trigger to replace the nand-activity
> >   one
> > 
> > And a bunch of specific NAND driver improvements/fixes. Here are the
> > changes that are worth mentioning:
> > * rework of the OMAP GPMC and NAND drivers
> > * prepare the sunxi NAND driver to receive DMA support
> > * handle bitflips in erased pages on GPMI revisions that do not support
> >   this in hardware.  
> 
> Thanks for the summary; it'd be nice if you can include this in the tag
> description next time. Then it'll get auto-populated below with
> git-request-pull, and it'll get pulled in automatically for me.
> 
> I also see this when I try to build a reasonably simple kernel:
> 
> depmod: ERROR: Found 5 modules in dependency cycles!
> depmod: ERROR: Cycle detected: nand -> nand_bch -> nand
> 
> I think the new reference to &nand_ooblayout_lp_ops in nand_bch.c is
> causing this (I can at least build/boot if I hack that out).

Oops, indeed. I should really setup an environment to test all the
cases on a build server (+ some static code checker like sparse).

Anyway, I fixed the bug as proposed on IRC (moved the
mtd_set_ooblayout() call into nand_base.c), and resent the PR.

Thanks,

Boris

-- 
Boris Brezillon, Free Electrons
Embedded Linux and Kernel engineering
http://free-electrons.com