linux-kernel - Re: [syzbot] [ext4?] possible deadlock in jbd2_log_wait

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

Message-ID: <20230301124249.tj7oub3r35q5z3rx@quack3>
Date:   Wed, 1 Mar 2023 13:42:49 +0100
From:   Jan Kara <jack@...e.cz>
To:     syzbot <syzbot+9d16c39efb5fade84574@...kaller.appspotmail.com>
Cc:     hdanton@...a.com, jack@...e.com, jack@...e.cz,
        linux-ext4@...r.kernel.org, linux-fsdevel@...r.kernel.org,
        linux-kernel@...r.kernel.org, syzkaller-bugs@...glegroups.com,
        tytso@....edu
Subject: Re: [syzbot] [ext4?] possible deadlock in jbd2_log_wait_commit

On Wed 01-03-23 02:08:19, syzbot wrote:
> syzbot has bisected this issue to:
> 
> commit 0813299c586b175d7edb25f56412c54b812d0379
> Author: Jan Kara <jack@...e.cz>
> Date:   Thu Jan 26 11:22:21 2023 +0000
> 
>     ext4: Fix possible corruption when moving a directory
> 
> bisection log:  https://syzkaller.appspot.com/x/bisect.txt?x=14845d50c80000
> start commit:   e492250d5252 Merge tag 'pwm/for-6.3-rc1' of git://git.kern..
> git tree:       upstream
> final oops:     https://syzkaller.appspot.com/x/report.txt?x=16845d50c80000
> console output: https://syzkaller.appspot.com/x/log.txt?x=12845d50c80000
> kernel config:  https://syzkaller.appspot.com/x/.config?x=f763d89e26d3d4c4
> dashboard link: https://syzkaller.appspot.com/bug?extid=9d16c39efb5fade84574
> syz repro:      https://syzkaller.appspot.com/x/repro.syz?x=11d96208c80000
> C reproducer:   https://syzkaller.appspot.com/x/repro.c?x=176d917f480000
> 
> Reported-by: syzbot+9d16c39efb5fade84574@...kaller.appspotmail.com
> Fixes: 0813299c586b ("ext4: Fix possible corruption when moving a directory")
> 
> For information about bisection process see: https://goo.gl/tpsmEJ#bisection

Drat, yeah, lockdep is actually right. We should not be grabbing i_rwsem
while having transaction started in ext4_rename(). I'm somewhat surprised
we didn't hit this lockdep warning earlier during the testing. So we need
to move the locking earlier in ext4_rename(). I'll send a patch...

								Honza

-- 
Jan Kara <jack@...e.com>
SUSE Labs, CR