Peter B sent the following to help out Randy:
First thing:
losh> info mem
and send back the NAND bad block information so we can see if they are blocks YAFFS marked bad, or are manufacturer bad blocks. From the thread, I think they are YAFFS bad blocks, and something he did may have tweaked them into a bad state. He's trying to use Linux - is he using YAFFS in Linux? If so, then which kernel and where can I find his YAFFS code he's using. It may be that the Linux YAFFS is conflicting with the YAFFS we use in LoLo(in regards to what a bad block mark is).
If he wants to bring the blocks back, he can try(IIRC):
losh> erase B0 B178 /dev/nand0 force
to erase the NAND, and remark YAFFS-marked bad blocks as good. "info mem" at that point will tell what blocks are bad after the erase. If there are blocks marked bad by the manufacturer in that output, then he can try:
losh> erase B0 B178 /dev/nand0 force force
to *force* erasing the blocks, including manufacturer-marked bad blocks.
>From experience, the NAND chip on the IMX31 will complain that the
>erase
failed on a manufacturer-marked bad block and remark them as manufacturer-marked bad, but he should write down the manufacturer-marked bad blocks before the step (or save the LoLo output)...
Then try the "add-yaffs; mount" command and tell us if it still complains. It shouldn't since those blocks have been all erased.
Hope this helps.
|