logfs unmount bug

srimugunthan dhandapani srimugunthan.dhandapani at gmail.com
Mon Aug 8 10:44:15 EDT 2011


Hi,

> Not good.  So far I managed to see a (the?) tiobench bug once, but
> couldn't make it reproduceable.  Nor could I reproduce the bonnie bug.
> Once the bugs are reproduceable, even my feeble eyes can usually spot
> the cause sooner or later.  With a backtrace alone and no idea what
> internal state it depends on, life is much harder.
>

I am sorry , i dont know what i am missing. I tried it 5 times. 4/5
trials, the bonnie test didnt complete. .Even when it completed once,
the logs show kernel bug. I have enabled logfs DEBUG and in all trials
the pattern of the crash is as follows:
-
Aug  8 19:31:25 mll kernel: [  516.247877] move_inode_to_page(1, 12, 0)
Aug  8 19:31:25 mll kernel: [  516.247895] logfs_segment_write(1, 12,
6) 340d74->340e58 0->e4
Aug  8 19:31:25 mll kernel: [  516.247900] logfs_segment_delete(1, 12,
6) 340d74->340e58 0->e4
Aug  8 19:31:25 mll kernel: [  516.248060] ------------[ cut here ]------------
Aug  8 19:31:25 mll kernel: [  516.248706] kernel BUG at
/home/srimugunthan/Documents/logfs/file.c:172!
-
I have attached the logfs DEBUG log for  the command
$ bonnie -d /mnt/flash_drive/ -s 10 -r 5

Writing intelligently...done
Rewriting...done
Reading a byte at a time...done
Reading intelligently...done
start 'em...done...done...done...done...
(hang)

> Just in case this depends on device size or erase size, can you send
> me the mklogfs output?


the mklogfs output is as follows:
$ sudo logfsprogs/mklogfs /dev/mtd0

Will create filesystem with the following details:
              hex:   decimal:
fssize=    8000000  134217728
segsize=      4000      16384
blocksize=    1000       4096
writesize=     200        512

Do you wish to continue (yes/no)
yes

Finished generating LogFS



Please let me know if you know any more info.
Thanks,
mugunthan
-------------- next part --------------
A non-text attachment was scrubbed...
Name: logfs_DEBUG.log
Type: text/x-log
Size: 2225773 bytes
Desc: not available
URL: <http://lists.infradead.org/pipermail/linux-mtd/attachments/20110808/97281b28/attachment-0001.bin>


More information about the linux-mtd mailing list