Kernel removing directories

Murtada Shah mshah at yottayotta.com
Mon Apr 29 18:38:07 EDT 2002


My mistake David. That was Jffs2 debug set to 2, that is why it took hours.

Nonetheless, I attached the serial output of jffs2debug set to 1... so 
you can take a look at it. I copied/pasted the relevant parts at the end.

The inode #7 seems to be causing problems. You can find the relevant 
sections from the attached file by finding the following quotes. As you 
see inode #7 has important children :) (i.e. "mv", "ps", 
/etc/init.d/rc&). I can boot from nfs & mount the /dev/mtdblock/0 no 
problem after another reboot.

--------------------------------------------
jffs2_scan_inode_node(): Node at 0x00ed3768
Node is ino #7, version 44. Range 0x0-0x0
....
....
....

Pass 2: ino #7, nlink 0, ic c0f0e7e8, nodes c0ec8550
JFFS@: Removing ino #7 with nlink == zero.
obsoleting node at 0x00fdca58.
...
...
...
obliterating obsoleted node at 0x008b9e70
Inode #7 was a directory with children - removing those too...
Removing child "mv", ino #14
Removing child "ps", ino #15
Removing child....
...
...
...
/etc/init.d/rcY
...
...
INIT: cannot execute "/etc/init.d/rc"
....
...


-------------------------------------------------
David Woodhouse wrote:

>
>mshah at yottayotta.com said:
>
>>When I set jffs2 debug level to 1... and default_console_loglevel to
>>9...  I get output for an extremely long time. Is this normal? I've
>>been  waiting for about 3 hours now and it's still not done. Normally
>>it takes  several seconds to load the filesystems. 
>>
>
>A long time is normal. Hours is a little excessive. How big is your file 
>system? Are you running at 115200 baud or faster?
>
>--
>dwmw2
>


-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: serialB
Url: http://lists.infradead.org/pipermail/linux-mtd/attachments/20020429/811991dd/attachment.pl 


More information about the linux-mtd mailing list