linux-ext4 - ext4_mb_generate_buddy block bitmap and bg descriptor inconsistent

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>] [day] [month] [year] [list]

Message-ID: <8e22ed00-f290-04de-15ba-eb6580e66ce8@luciad.com>
Date:   Tue, 22 Aug 2017 15:25:46 +0200
From:   Geert Lorang <geert.lorang@...iad.com>
To:     linux-ext4@...r.kernel.org
Subject: ext4_mb_generate_buddy block bitmap and bg descriptor inconsistent

Hi,

We have some QEMU virtual machines that go randomly into read-only state 
where the ext4 filesystem aborts with:
ext4_mb_generate_buddy:757: group X, block bitmap and bg descriptor 
inconsistent: Y vs 32768 free clusters

I know this has been discussed extensively already in the past e.g.

https://bugzilla.kernel.org/show_bug.cgi?id=102731
https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=818502
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1423672

Now the difference with those reports and our case is that the last 
number of the inconsistency (grp->bb_free) is *always* 32768 + We're 
sure the proposed patch is applied on both host & guest.

Is there anyone that can provide some insights in what /might/ be going 
on? E.g. what does this error actually mean and/or why it would always 
be exactly 32768?
It think it's unlikely something is wrong in the ext4 code - I'm just 
trying to reproduce the problem so we can figure out what's going on here.

My apologies if this does not belong on this list.

Thanks,
Geert