linux-ext4 - Re: ext4 metadata corruption bug?

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20140506155159.GY5136@gradx.cs.jhu.edu>
Date:	Tue, 6 May 2014 11:51:59 -0400
From:	Nathaniel W Filardo <nwf@...jhu.edu>
To:	"Theodore Ts'o" <tytso@....edu>
Cc:	linux-ext4@...r.kernel.org, admins@....jhu.edu
Subject: Re: ext4 metadata corruption bug?

On Tue, May 06, 2014 at 11:42:39AM -0400, Theodore Ts'o wrote:
> On Thu, May 01, 2014 at 12:25:03PM -0400, Nathaniel W Filardo wrote:
> > Here's another kernel report, this time from /dev/sda1, which is a QEMU-IDE
> > view of a local LVM volume and is only 4060864 blocks big, so it falls into
> > neither the "Ceph's fault" nor "8TB is special" bins:

Ack, oops; my bad.  So I just checked the configuration and realized that,
while /dev/sda1 was in fact once upon a time a local view of LVM, it is now
in Ceph.  So it does eliminate the "8TB is special" bin but "Ceph's fault"
is still in play.

> > [922646.672586] EXT4-fs error (device sda1): ext4_mb_generate_buddy:756: group 17, 24652 clusters in bitmap, 24651 in gd; block bitmap corrupt.
> 
> So this is a different report from the ones where we see this error:
> 
> [817576.492468] EXT4-fs error (device vdd): ext4_mb_release_inode_pa:3729: group 59035, free 14, pa_free 12
> 
> Have you seen any more of these errors?

I think so, yes; I recall seeing bugs in both the allocation and the free
side of things, but I will keep an eye out.

> > [922646.712017] BUG: unable to handle kernel NULL pointer dereference at 0000000000000028
> > [922646.712017] IP: [<ffffffffa014a7cc>] __ext4_error_inode+0x2c/0x150 [ext4]
> 
> FYI, this BUG (which can happens after certain jbd2 errors, which in
> your case happened after the journal was aborted) is fixed with commit
> 66a4cb187b9 which will be in v3.15.

Excellent; I look forward to the new release and will stop nagging you with
these. :)

--nwf;

Content of type "application/pgp-signature" skipped