[PATCH] mtd: rawnand: omap: Use BCH private fields in the specific OOB layout

[Miquel Raynal]

On Tue, 2021-01-19 at 15:55:10 UTC, Miquel Raynal wrote:
> The OMAP driver may leverage software BCH logic to locate errors while
> using its own hardware to detect the presence of errors. This is
> achieved with a "mixed" mode which initializes manually the software
> BCH internal logic while providing its own OOB layout.
> 
> The issue here comes from the fact that the BCH driver has been
> updated to only use generic NAND objects, and no longer depend on raw
> NAND structures as it is usable from SPI-NAND as well. However, at the
> end of the BCH context initialization, the driver checks the validity
> of the OOB layout. At this stage, the raw NAND fields have not been
> populated yet while being used by the layout helpers, leading to an
> invalid layout.
> 
> The chosen solution here is to include the BCH structure definition
> and to refer to the BCH fields directly (de-referenced as a const
> pointer here) to know as early as possible the number of steps and ECC
> bytes which have been chosen.
> 
> Note: I don't know which commit exactly triggered the error, but the
> entire migration to a generic BCH driver got merged in one go, so this
> should not be a problem for stable backports.
> 
> Reported-by: Adam Ford <aford173 at gmail.com>
> Fixes: 80fe603160a4 ("mtd: nand: ecc-bch: Stop using raw NAND structures")
> Signed-off-by: Miquel Raynal <miquel.raynal at bootlin.com>
> Tested-by: Adam Ford <aford173 at gmail.com> #logicpd-torpedo-37xx-devkit-28.dts

Applied to https://git.kernel.org/pub/scm/linux/kernel/git/mtd/linux.git mtd/fixes.

Miquel