U-Boot OMAP GPMC ECC change

[Colin Foster]

Hi Michael,

On Mon, Jun 26, 2023 at 07:03:19AM +0200, Michael Nazzareno Trimarchi wrote:
> Hi all
> 
> Il sab 20 mag 2023, 19:28 Colin Foster <colin.foster at in-advantage.com> ha
> scritto:
> 
> > On Fri, May 19, 2023 at 03:41:34PM +0300, Roger Quadros wrote:
> > > Hi Colin,
> > >
> > > On 19/05/2023 02:19, Colin Foster wrote:
> > > > Hi Roger,
> > > >
> > > >> Can you please share your spl/u-boot.cfg?
> > > >
> > > > Attached
> > >
> > > Couple of questions there
> > >
> > > 1) CONFIG_MTDPARTS_DEFAULT
> > "mtdparts=nandflash:0x20000(xload_raw),0x180000(u-boot),0x180000(u-boot-2),0x1fce0000(main)"
> > > Is this correct and matches with what kernel sees?
> > > I couldn't see the NAND partition table in the Kernel Device tree patch.
> >
> > Yes, this is correct. I intentionally left my MTD Partitions out of the
> > kernel patch, since I don't want any changes I might make to the flash
> > partitions to require further patches. I'm currently at this structure
> > (SPL, 2x U-Boot, and main UBI with A/B partitions and 2x U-Boot Envs)
> >
> > The SD Boot version of U-Boot doesn't use NAND, so it might have a stale
> > partition layout that I'll need to remove / modify.
> >
> 
> 
> Was any end up here?

I got pulled away from this before I could figure it out, but I have an
idea.

I suspect the ECC calculation is working as expected. It is just that
now it is pre-calculating the ECC for regions that my flash writing was
simply ignroing (just leaving erased). Previously the ECC would only be
calculated for pages that were correctly written, so wouldn't cause any
issues.

I do need to verify this hypothesis, but admittedly don't know how soon
that'll be. (I won't be near hardware for at least a week)

> 
> Michael