[RFC] On-die ECC support

[Richard Weinberger]

Brian,

On Wed, Mar 25, 2015 at 3:02 PM, Richard Weinberger <richard at nod.at> wrote:
> Some Micron NAND chips offer an on-die ECC (AKA internal ECC) feature,
> in this mode all ECC is automatically performed by the NAND chips itself.
> Most user are happy with software ECC or hardware ECC performed by the
> NFC (NAND flash controller).
> But there are situations where multi-bit hardware ECC is not available,
> some AT91 based SoCs for example, and due to CPU limitations
> software ECC is not feasible. In such situations on-die ECC can help.
> People are using that feature in the wild for years.
> As this feature is not supported in mainline these guys are using out-of-tree
> patches of doubtful quality. That's why I did that patch series.
>
> This series is based on patches by David Mosberger and in-house patches
> by a customer.
>
> Currently I'm not happy with patch 2/3.
> To allow raw access (access without ECC) we have to disable/enable
> on-die ECC mode in various NAND core functions.
> I could also think of dropping support for raw access completely.
>
> I'd like to thank Ammonit Measurement GmbH for making this contribution to the MTD community happen!
>
> [PATCH 1/3] mtd: nand: Add on-die ECC support
> [PATCH 2/3] mtd: nand: Add support for raw access when using on-die
> [PATCH 3/3] mtd: nand: Wire up on-die ECC support

As the merge window is almost over I fear this patches won't make it into v4.1.
Do you want me to resent them for the v4.2 cycle or are they already
on your TODO list? :)

-- 
Thanks,
//richard