On Mon, 2012-07-02 at 17:49 +1000, Iwo Mergler wrote:
> It's similar to a BeagleBone, but with NAND FLASH, I'm using the
> omap2 ECC driver.
> 
> So, if I set NAND_NO_SUBPAGE_WRITE in chip.options, UBI will
> leave the subpages alone?

I think so, but it is better to dig and find out why sub-pages to not
work - this sounds like a bug. With sub-pages support you'll waste less
flash space because UBI overhead will be smaller.

> So I had to make two debug runs, one each for ubi and ubifs debugging
> enabled.

I've reproduced this bug using nandsim and fixed it. I've sent you the
fix few minutes ago. Could you please test it? After I've got a
"Tested-by:", I'll send it to Linus and it will also go to -stable.
Thanks!

-- 
Best Regards,
Artem Bityutskiy