[U-Boot] QorIQ P1020 NAND Flash 2k-Page-Size Problem

[Scott Wood]

On Thu, 2014-08-07 at 16:59 -0300, Allan Fislor wrote:
> On Thu, Aug 7, 2014 at 7:49 AM, Allan Fislor <allan.fislor at gmail.com> wrote:
> > On Wed, Aug 6, 2014 at 3:58 PM, Scott Wood <scottwood at freescale.com> wrote:
> >>
> >> On Tue, 2014-08-05 at 16:48 -0300, Allan Fislor wrote:
> >> > Hi all,
> >> >
> >> > I have a custom Freescale QorIQ P1020 board, with a serial NOR Flash (eSPI
> >> > bus) and a NAND Flash (eLBC/FCM bus). Im booting from the NOR flash.
> >> >
> >> > The old NAND (512 bytes page size) was working perfectly, so I swapped for
> >> > another with 2k bytes page size.
> >> >
> >> > I changed in the config file:
> >> >
> >> >    CONFIG_SYS_NAND_OR_PRELIM: added OR_FCM_PGS option
> >> >
> >> >    CONFIG_SYS_NAND_BLOCK_SIZE: set to (128 * 1024)
> >> >
> >> >
> >> > But this new NAND flash does not work. If I enter the command "nand
> >> > erase.chip", all sectors appears with bad block.
> >>
> >> Are the timings correct for your new NAND chip?
> >>
> >
> > My eLBC is at 25MHz, I also tried 50MHz and 100MHz.
> >
> > My current OR options are: OR_FCM_PGS | OR_FCM_CSCT | OR_FCM_CST |
> > OR_FCM_CHT  | OR_FCM_SCY_1  | OR_FCM_TRLX  | OR_FCM_EHTR
> >
> > I tried some combinations, without success.
> >
> >>
> >> What do you see if you dump the raw NAND contents?
> >>
> >
> > => nand dump 0
> > Page 00000000 dump:
> >         30 30 30 30 30 30 30 30  30 20 20 20 00 00 00 00
> >         00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00
> >         (all zeros)
> >         00 00 00 00 00 00 00 08  00 00 00 00 00 00 00 00
> >         (all zeros)
> >
> > OOB:
> >         00 00 00 00 00 00 00 00
> >         00 00 00 00 00 00 00 00
> >         00 00 00 00 00 00 00 00
> >         00 00 00 00 00 00 00 00
> >         00 00 00 00 00 00 00 00
> >         00 00 00 00 00 00 00 00
> >         00 00 00 00 00 00 00 00
> >         00 00 00 00 00 00 00 00
> >
> >>
> >> Have you seen this behavior with more than one instance of the new flash
> >> chip type?
> >>
> >
> > With 2k page-size I only tested this PN: MT29F2G08ABAEAWP
> >
> > And others boards with this same NAND, have the same behavior.
> >
> > Cheers,
> >
> > --
> > Allan
> 
> 
> I got a NAND from ST-Micro, with 2k page-size. It works perfectly! And
> it accepts a lot of OR timings combination.
> 
> I think is something related specifically to Micron....

I don't know what the cause of the problem you're seeing is, but
MT29F2G08ABAEAWP is not a good match for P1020 because it requires 4
bits of ECC per 528 bytes.  P1020 has eLBC which is only capable of one
bit correction per 512 bytes.  So you'd need to either use software ECC,
or enable support for the NAND chip's internal ECC.

-Scott