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