Hi Scott,
On Tue, Dec 7, 2010 at 8:23 AM, Scott Wood <scottw...@freescale.com> wrote:
> On Thu, Dec 02, 2010 at 04:53:30AM -0000, Lei Wen wrote:
>> Seems original implementation forget to set the pointer to point
>> to the oobbuf, so when we want to see oob buf, we see nothing...
>> Fix it by get pointer as the oobbuf set.
>>
>> Signed-off-by: Lei Wen <lei...@marvell.com>
>>
>> ---
>> common/cmd_onenand.c | 2 ++
>> 1 files changed, 2 insertions(+), 0 deletions(-)
>>
>> diff --git a/common/cmd_onenand.c b/common/cmd_onenand.c
>> index 33108f1..6d77495 100644
>> --- a/common/cmd_onenand.c
>> +++ b/common/cmd_onenand.c
>> @@ -319,6 +319,8 @@ static int onenand_dump(struct mtd_info *mtd, ulong off,
>> int only_oob)
>> }
>> puts("OOB:\n");
>> i = mtd->oobsize >> 3;
>> + p = oobbuf;
>> +
>> while (i--) {
>> printf("\t%02x %02x %02x %02x %02x %02x %02x %02x\n",
>> p[0], p[1], p[2], p[3], p[4], p[5], p[6], p[7]);
>
> Earlier in the function there's a comment saying:
> ops.oobbuf = oobbuf;
> /* must exist, but oob data will be appended to ops.databuf */
>
> Is this no longer accurate (I hope so, since that's awful)? If so, let's
> remove the comment.
>
Seem this comment should be removed...
For in onenand_read_ops_nolock, we transfer the oob data to the oobbuf
pass from the onenand dump function call in cmd_onenand.
I would send the updated patch later.
Best regards,
Lei
_______________________________________________
U-Boot mailing list
U-Boot@lists.denx.de
http://lists.denx.de/mailman/listinfo/u-boot
