linux-kernel - Re: [lkp] [direct] 187372a3b9: WARNING: CPU: 2 PID: 3999 at fs/ext4/inode.c:3315 ext4_direct

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: <20160218003929.GT14668@dastard>
Date:	Thu, 18 Feb 2016 11:39:29 +1100
From:	Dave Chinner <david@...morbit.com>
To:	kernel test robot <ying.huang@...ux.intel.com>
Cc:	Christoph Hellwig <hch@....de>, lkp@...org,
	LKML <linux-kernel@...r.kernel.org>,
	Dave Chinner <dchinner@...hat.com>, linux-ext4@...r.kernel.org,
	tytso@....edu
Subject: Re: [lkp] [direct] 187372a3b9: WARNING: CPU: 2 PID: 3999 at
 fs/ext4/inode.c:3315 ext4_direct_IO+0x56d/0x5b0()

On Tue, Feb 16, 2016 at 08:38:10AM +0800, kernel test robot wrote:
> FYI, we noticed the below changes on
> 
> https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git master
> commit 187372a3b9faff68ed61c291d0135e6739e0dbdf ("direct-io: always call ->end_io if non-NULL")
> 
> 
> run fstests generic/299 at 2016-02-16 06:10:03
> EXT4-fs (sda8): mounted filesystem with ordered data mode. Opts: acl,user_xattr
> ------------[ cut here ]------------
> WARNING: CPU: 2 PID: 3999 at fs/ext4/inode.c:3315 ext4_direct_IO+0x56d/0x5b0()
.....
>  [<ffffffff81422e52>] dump_stack+0x4b/0x69
>  [<ffffffff8107a476>] warn_slowpath_common+0x86/0xc0
>  [<ffffffff8107a56a>] warn_slowpath_null+0x1a/0x20
>  [<ffffffff81282a2d>] ext4_direct_IO+0x56d/0x5b0
>  [<ffffffff811719ea>] generic_file_direct_write+0xaa/0x170
>  [<ffffffff81171b70>] __generic_file_write_iter+0xc0/0x1f0
>  [<ffffffff8127d201>] ext4_file_write_iter+0x141/0x430
>  [<ffffffff81242436>] aio_run_iocb+0x226/0x290
>  [<ffffffff8124353d>] do_io_submit+0x29d/0x530
>  [<ffffffff812437e0>] SyS_io_submit+0x10/0x20
>  [<ffffffff818e012e>] entry_SYSCALL_64_fastpath+0x12/0x71

3309                 /*
3310                  * When no IO was submitted ext4_end_io_dio() was not
3311                  * called so we have to put iocb's reference.
3312                  */
3313                 if (ret <= 0 && ret != -EIOCBQUEUED && iocb->private) {
3314                         WARN_ON(iocb->private != io_end);
3315  >>>>>>>>               WARN_ON(io_end->flag & EXT4_IO_END_UNWRITTEN);
3316                         ext4_put_io_end(io_end);
3317                         iocb->private = NULL;
3318                 }

Pre-existing ext4 issue. It assumes that unwritten extent conversion
occurs even on IO submission error. i.e. when we avoid the stale
data exposure bug that the bisect-blamed commit fixed, ext4 warns
that it failed to expose stale data....

I'd just remove that warning, but I have no idea how that impacts on
the rest of ext4's end IO handling. Christoph, Ted, the ball is in
your court for this one.

Cheers,

Dave.
-- 
Dave Chinner
david@...morbit.com