Data CRC failed on node...

Data CRC failed on node...

[Massimo Calò]

Hi All,

    I have a 5272 Motorola Coldfire based board, running "uClinux dist. :
uClinux-20030305.tar.gz" +

"Tool Chain : m68k-elf-tools-20030314.sh.".

I have a 2Mb flash (AMD AM29LV160DB) using MTD + JFFS2.

It has worked fine for a long time, then I received the following message
during the bootstarp :

        Command: mkdir /var/run
        Command: mkdir /var/lock
        Command: mount -t jffs2 /dev/mtdblock2 /mnt
        jffs2_scan_inode_node(): Data CRC failed on node at 0x0005ca64: Read
0xb399e587c
        jffs2_scan_inode_node(): Data CRC failed on node at 0x000a1424: Read
0xee61d2b26
        Command: ifconfig lo 127.0.0.1
        Command: route add -net 127.0.0.0 netmask 255.0.0.0 lo

To fix the problem I do the command "eraseall /dev/mtd2" and the error was
cleared.

I have posted the problem to the uClinux List and I received the following
answer :

>>        Yes. Pls umount when u done using the flash. Because if u don't
unmount
>>        properly it will just scan for the nodes once again when the
filesystems
>>        comes up next time. It's just scans all the nodes,if u properly
unmount
>>        the filesystem before turning it off. So pls ensure u umount the
>>        filesystem before turning off.


Do I have to mount/unmount the flash every time I have to read/write from it
?

Which is the correct way to read/write from flash to avoid the "Data CRC"
problem ?

Thanks!

Massimo.

Re: Data CRC failed on node...

[David Woodhouse]

On Mon, 2003-06-16 at 12:34, Massimo Calò wrote:
> Hi All,
> 
>     I have a 5272 Motorola Coldfire based board, running "uClinux dist. :
> uClinux-20030305.tar.gz" +
> 
> "Tool Chain : m68k-elf-tools-20030314.sh.".
> 
> I have a 2Mb flash (AMD AM29LV160DB) using MTD + JFFS2.
> 
> It has worked fine for a long time, then I received the following message
> during the bootstarp :

>         jffs2_scan_inode_node(): Data CRC failed on node at 0x0005ca64: Read
> 0xb399e587c

That's normal and harmless. It's just a warning, not an error. You
powered down while a node was being written -- perhaps by the garbage
collection thread. In that case, the previous node containing the same
data will still be valid and will continue to be used. No data are lost.

> To fix the problem I do the command "eraseall /dev/mtd2" and the error was
> cleared.

Or you could have waited until the block containing the offending node
was garbage-collected, and it would have gone away on its own.

> Do I have to mount/unmount the flash every time I have to read/write from it
> ?

No, but if you want to avoid the cosmetic warnings, you should mount the
file system readonly or unmount it before each _reboot_.

-- 
dwmw2