I need to generate the expected ECC
values so our vendor can use this binary image program the NAND FLASH prior to installation onto the PCB.
The focus of the post was the observation that the ECC generation of the genecc tool provided by TI appeared to be different than the ECC generated by the dedicated hardware within the DaVinci DM355. I performed several tests using different patterns to verify the correctness of the calculated ECC. The results were inconsistent. With an incrementing pattern, the ECC's did not compare. With a sparse pattern, the ECC's did match.
I was wondering whether other developers had experienced any issues which would indicate the ECC correction within the DM355, core version 1.1, was pattern dependent.
On Fri, 24 Apr 2009 09:50:17 -0700
"Jerry Johns" <[email protected]> wrote:
> I'm not quite understanding the focus of your post - are
>you trying to
> flash u-boot, ubl, kernel, filesystem through the
>nandwriter program?
> The CCS nandwriter program does not calculate ECC
>values, since it
> leaves this up to the HW 512 byte 4-bit ECC engine to
>generate this. The
> program simply accepts these ECC values and places them
>in the correct
> area that can be read by the RBL when it loads in UBL,
>and consequently,
> the UBL when it reads in U-boot. Is the genecc utility
>the one provided
> by YAFFS? If so, that does not generate proper ECC
>values - it just
> generates column parities for a chunk of data that is
>not compatible
> with the HW ECC.
>
>
>
> You should not have to generate your own ECC values,
>with the exception
> being for YAFFS/YAFFS2 Filesystem images - YAFFS
>generates its own ECC
> (row, column parities) and lays it out in the free bytes
>in the spare
> area. The kernel would need to be configured for SW_ECC
>for this to work
> properly.
>
>
>
> Hope it helps,
>
>
>
> Jerry Johns
>
> Design Engineer
>
> Nuvation Research Corp - Canada
>
> Tel: (519) 746-2304 ext. 221
>
> www.nuvation.com <http://www.nuvation.com>
>
>
>
>
>
The focus of the post was the observation that the ECC generation of the genecc tool provided by TI appeared to be different than the ECC generated by the dedicated hardware within the DaVinci DM355. I performed several tests using different patterns to verify the correctness of the calculated ECC. The results were inconsistent. With an incrementing pattern, the ECC's did not compare. With a sparse pattern, the ECC's did match.
I was wondering whether other developers had experienced any issues which would indicate the ECC correction within the DM355, core version 1.1, was pattern dependent.
On Fri, 24 Apr 2009 09:50:17 -0700
"Jerry Johns" <[email protected]> wrote:
> I'm not quite understanding the focus of your post - are
>you trying to
> flash u-boot, ubl, kernel, filesystem through the
>nandwriter program?
> The CCS nandwriter program does not calculate ECC
>values, since it
> leaves this up to the HW 512 byte 4-bit ECC engine to
>generate this. The
> program simply accepts these ECC values and places them
>in the correct
> area that can be read by the RBL when it loads in UBL,
>and consequently,
> the UBL when it reads in U-boot. Is the genecc utility
>the one provided
> by YAFFS? If so, that does not generate proper ECC
>values - it just
> generates column parities for a chunk of data that is
>not compatible
> with the HW ECC.
>
>
>
> You should not have to generate your own ECC values,
>with the exception
> being for YAFFS/YAFFS2 Filesystem images - YAFFS
>generates its own ECC
> (row, column parities) and lays it out in the free bytes
>in the spare
> area. The kernel would need to be configured for SW_ECC
>for this to work
> properly.
>
>
>
> Hope it helps,
>
>
>
> Jerry Johns
>
> Design Engineer
>
> Nuvation Research Corp - Canada
>
> Tel: (519) 746-2304 ext. 221
>
> www.nuvation.com <http://www.nuvation.com>
>
>
>
>
>
_______________________________________________ Davinci-linux-open-source mailing list [email protected] http://linux.davincidsp.com/mailman/listinfo/davinci-linux-open-source
