2010-01-03 16:38:57

by Tvrtko Ursulin

[permalink] [raw]
Subject: [2.6.33-rc2] kernel BUG at fs/ext4/inode.c:1063!


Hi,

As much as it was visible, all writeout stopped after this happened so I could
only capture it with the camera. Please see the attachment.

Tvrtko


Attachments:
ext4bug.png (69.48 kB)

2010-01-04 03:10:24

by Theodore Ts'o

[permalink] [raw]
Subject: Re: [2.6.33-rc2] kernel BUG at fs/ext4/inode.c:1063!

On Sun, Jan 03, 2010 at 04:38:57PM +0000, Tvrtko Ursulin wrote:
>
> As much as it was visible, all writeout stopped after this happened so I could
> only capture it with the camera. Please see the attachment.

This is a known bug. The the latest mainline version of Linux (I
think 2.6.33-rc2-git4 includes the necessary patch) will no longer
BUG. There are two more patches which I'm about to push to Linus that
are necessary to fix up a typo and some potential undesirable side
effects of the initial fix of the problem. If you want to get all of
the fixes, pull in the changes from:

git://git.kernel.org/pub/scm/linux/kernel/git/tytso/ext4.git for_linus

Regards,

- Ted


2010-01-04 08:14:11

by Tvrtko Ursulin

[permalink] [raw]
Subject: Re: [2.6.33-rc2] kernel BUG at fs/ext4/inode.c:1063!

On Monday 04 Jan 2010 01:08:36 [email protected] wrote:
> On Sun, Jan 03, 2010 at 04:38:57PM +0000, Tvrtko Ursulin wrote:
> > As much as it was visible, all writeout stopped after this happened so I
> > could only capture it with the camera. Please see the attachment.
>
> This is a known bug. The the latest mainline version of Linux (I
> think 2.6.33-rc2-git4 includes the necessary patch) will no longer
> BUG. There are two more patches which I'm about to push to Linus that
> are necessary to fix up a typo and some potential undesirable side
> effects of the initial fix of the problem. If you want to get all of
> the fixes, pull in the changes from:
>
> git://git.kernel.org/pub/scm/linux/kernel/git/tytso/ext4.git
> for_linus

Was it triggered by no free space? Later I realised that may be what happened
when this triggered. I was writing out a large file which may filled the
filesystem. After reboot file in question had zero size, due to journal replay I
think, but given the amount of free space on the filesystem I think it might
have used all space while writing out.

If this is related I will wait for -rc3 and just make sure I don't run out of
space. :) (So far I avoided starting to build git trees etc)

Tvrtko