[Qi] bad magic when booting from NAND

Paul Fertser fercerpav at gmail.com
Sun Apr 5 16:50:45 CEST 2009


Dmitry Kurochkin <dmitry.kurochkin at gmail.com> writes:
> On Fri, Apr 3, 2009 at 8:44 PM, Werner Almesberger <werner at openmoko.org> wrote:
>> Dmitry Kurochkin wrote:
>>>  2: kernel              0x00820000
>>
>> Uh uh, a bad block. Maybe that's causing the trouble. You could
>> find out what's happening by first determining where the bad block
>> is located, e.g., with
>> http://svn.openmoko.org/developers/werner/badnand/
>>
>> and then checking if Qi finds it, e.g., by inserting a printf into
>> the bad block handling code in
>> src/cpu/s3c2442/nand_read.c:nand_read_ll
>
> I ran badnand /dev/mtd6 and got this:
>
> Factory 2, worn 0, good 2046
>
> Now I will look at Qi as you suggest.
>
> BTW How did you find out that there were badblocks?

U-boot command dynpart searches for bad blocks and creates nand
partition table accordingly.

Werner saw a unusual address for kernel partition; that means that the
previous partition was enlarged to compensate for a bad block.

Unfortunately, that badblock business is tricky as this information is
kept in 2 different places (OOB data and BBT at the end of NAND) and
is not always in sync. There was a discussion on the kernel list about
various ways of storing, searching and using bad blocks information
and the differences between NOR, NAND u-boots, Qi and the kernel.

-- 
Be free, use free (http://www.gnu.org/philosophy/free-sw.html) software!
mailto:fercerpav at gmail.com



More information about the openmoko-kernel mailing list