UBIFS failed to recover master node

[twebb]

On Tue, Jun 8, 2010 at 10:17 AM, Artem Bityutskiy <dedekind1 at gmail.com> wrote:
> On Tue, 2010-06-08 at 09:44 -0400, twebb wrote:
>> I've attached a log file that captures the problem I'm seeing.
>
> twebb, you did not enable debugging messages. But it looks like they are
> not actually needed.

I config'd kernel with...
CONFIG_MTD_UBI_DEBUG=y
CONFIG_UBIFS_FS_DEBUG=y

Did I miss something else?

>
>>   The
>> log includes failures on 2 of 4 UBIFS volumes that are part of a
>> single MTD device (and shows the other 2 volumes being
>> mounted/umounted without error).
>>
>> It appears from the log that the problem is associated with "corrupt
>> empty space" which can be seen in the log at lines 28, 313, 347, and
>> 359.  Could this be HW ECC failing to correct bit errors on reads?
>> Any suggestions on what this problem would be helpful.
>
> Yes, HW fails to correct bit errors. Why these errors are there - this
> it the question. May be MLC-related. You need to debug this. Or send me
> your board, I can try to take a look

I'll look into it, thanks.

>
> Did you enable UBI extra self checks?
>

No.