Hi all,
on a linux device using yaffs we had discovered a second prototype
(over 650) showing a NAND corruption issue. To be clear, the NAND
hosts the linux file system and after programming all was OK. The
product boots correctly and was functionnally tested.
The next boot sequence fails at the end of the boot when the kernel
tries to access the file system on the NAND. We have started to
investigate:
1 - We changed the boot parameter and use a NFS image to get access to
our product ... success
2 - We try to mount the yaffs file system “/bin/mount –t yaffs2
/dev/mtdblock9 /mnt9”. The prompt returns: "mounting /dev/mtdblock9 on
/mnt9 failed: Invalid argument
3 - echo $ returns as per last command exit code: 255
4 - Our kernel message serial debug port returns: "yaffs: dev is
32505865 name is "mtdblock9". yaffs: passed flags "". yaffs:
Attempting MTD mount on 31.9, "mtdblock9" block 1841 is bad.
Any advice to help us further analyse this issue?
How to dump the content of the NAND to find where is the corruption
and potentitaly the root cause?
In addition we would like to insist on one point: the NAND was not
erased before the first copy of the image; although the first boot was
OK, do you think that this could be the root cause. Just imagine that
a bad block was not identified because of the missing erase ... just a
guess. We don't clearly know how does yaffs manage these bad blocks.
Thanks,
Marc
(text/plain)