Hi,
>>>> I did a "eraseall" of the device, but the moment I try to
>>>> write, it's all marked bad... Any new hints?
>>> Make sure that bad block checking in the MTD layer is commented
>>> out (it does wipe all the factory marked bad blocks but in my
>>> experience this is not a crisis for development work) or the
>>> blocks are not erased at all and remain "bad".
>> Well, the device does report 99% free according to "df" after
>> eraseall. Further, a "mkyaffs" on the device does not state that
>> there is thousands of bad blocks - just the factory ones, so I
>> think eraseall does it correct? Moreover, I tried it on a new NAND
>> device never before being used for YAFFS, and the same occured
>> (with my new kernel without YAFFS_ECC)... You still think I should
>> try without bad block checking?
> No. It sounds like the mtd layer is content. I am surprised that you
> are having this grief. You will probably need to add a liberal
> smattering of printks to tease ot the issue. Is it the mtd layer
> reporting write errors?
Now, new strange things are happening!
I enabled "CONFIG_MTD_NAND_VERIFY_WRITE=y" in the kernel, and suddenly
it begun working!!
However, I get these kernel messages (alot) whenever I try to read/write
anything:
=== 8< 8< 8< ===
Writing data without ECC to NAND-FLASH is not recommended
Writing data without ECC to NAND-FLASH is not recommended
Writing data without ECC to NAND-FLASH is not recommended
Writing data without ECC to NAND-FLASH is not recommended
Reading data from NAND FLASH without ECC is not recommended
I guess these must be from YAFFS...
Is there really no ECC running now? I thought that config-option for MTD
did the job?
> Is it read/write oob reporting rubbish?
Before, when it didn't work, I saw the "Cannot allocate memory"
messages, as I've seen others report on the list...
BR,
Martin
(text/plain)