Bug 59801 - kernel BUG at /kernel/fs/ext4/extents.c:1996!
Summary: kernel BUG at /kernel/fs/ext4/extents.c:1996!
Status: CLOSED PATCH_ALREADY_AVAILABLE
Alias: None
Product: File System
Classification: Unclassified
Component: ext4 (show other bugs)
Hardware: All Linux
: P1 normal
Assignee: fs_ext4@kernel-bugs.osdl.org
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2013-06-17 02:20 UTC by yyl810
Modified: 2013-06-18 01:14 UTC (History)
1 user (show)

See Also:
Kernel Version: 3.0.21
Subsystem:
Regression: No
Bisected commit-id:


Attachments
log (15.90 KB, text/plain)
2013-06-17 02:22 UTC, yyl810
Details

Description yyl810 2013-06-17 02:20:38 UTC
Is it data corrupted? this bug make the device reset again and again, if there is a tool which can repair this error? I tried e2fsck, it didn't work.

Please help, thanks!
Comment 1 yyl810 2013-06-17 02:22:55 UTC
Created attachment 104911 [details]
log
Comment 2 Theodore Tso 2013-06-17 03:22:31 UTC
What version of e2fsprogs are you using?   The BUG_ON looks like it was triggered by a zero-length extent, and older versions of e2fsprogs didn't check for this condition.

I would suggest upgrading to e2fsprogs 1.42.7 (or e2fsprogs 1.42.8, which will be released in a few days), especially if you are using something older than 1.42.2 or so.
Comment 3 yyl810 2013-06-17 05:32:42 UTC
Thank you so much.
I was using 1.41.11, do you have e2fsprogs 1.42.7 bin file for ARM? if yes, could you kindly share with me?
Comment 4 Theodore Tso 2013-06-17 12:35:29 UTC
Especially with ARM, where in general there is no standardization on C library shared library bindings, etc., you really need to build user packages for yourself.    There are more modern (1.42.5) packages for e2fsprogs compiled for ARM from Debian, but there's no guarantee that it will work on whatever system you are using.

Please note that support of old kernels, especially kernels as old as 3.0, is on "best effort" basis.  People who have to use old kernels on a commercial basis should get help from a commercial support organization (or build up ext4 expertise within their own organization, but the former is generally more economically effective).   There are a huge number of bug fixes since 3.0, and while there are a few people will try to backport fixes to the 3.0 kernel, the testing to make sure those fixes do not introduce new bugs, and the effort to backport patches if they do not apply cleanly to older kernels, is very much on a best effort basis.

If the problem is really obvious, as in this case, where I can find the problem after looking for a few seconds, the open source community can try to give you support.  But if it requires more work than that, we can't help you --- we just wouldn't have the time to improve ext4 for the upstream kernel if we were to supply free support for ancient kernels.  (Not to mention that it would not be fair for those engineers whose salaries come from getting paid to support ancient kernels.)
Comment 5 yyl810 2013-06-18 01:14:51 UTC
I totally understand you and appreciate your suggestion.

Actually yesterday I tried compiling 1.42.7 on my system, although it's succeeded, but unfortunately there is crash, anyway I found a way to try.

Thanks again.

Note You need to log in before you can comment on or make changes to this bug.