xfs
[Top] [All Lists]

Re: [PATCH 3/4] fs: Remove i_size check from do_fallocate

To: Lukas Czerner <lczerner@xxxxxxxxxx>
Subject: Re: [PATCH 3/4] fs: Remove i_size check from do_fallocate
From: Theodore Ts'o <tytso@xxxxxxx>
Date: Sat, 12 Apr 2014 09:59:06 -0400
Cc: linux-fsdevel@xxxxxxxxxxxxxxx, linux-ext4@xxxxxxxxxxxxxxx, xfs@xxxxxxxxxxx, ceph-devel@xxxxxxxxxxxxxxx
Delivered-to: xfs@xxxxxxxxxxx
Dkim-signature: v=1; a=rsa-sha256; c=simple/simple; d=thunk.org; s=mail; t=1397311146; bh=hRKZGBAwOxVmiYyIyKp3rajCF72hwJ3Vl/+IcMgu/NY=; h=Date:From:To:Cc:Subject:References:In-Reply-To:From; b=iidJqMuAOq7dy5Lg8/Qez+N35PeRaOZ2hyV/SgQBM8aq3W/sXcLDU6Jz0HNwazKUW T+B/W18p4OhJAT8oGX/clLct3KYxpTMNZWLD8JqgLJlMrXCzJ5p93Ar5ZBK1ZJlWyh 42IZGpkIgLRJz4GaVEHw99N2PuJYuD5wIjdcrmao=
In-reply-to: <1397242665-2183-3-git-send-email-lczerner@xxxxxxxxxx>
References: <1397242665-2183-1-git-send-email-lczerner@xxxxxxxxxx> <1397242665-2183-3-git-send-email-lczerner@xxxxxxxxxx>
User-agent: Mutt/1.5.23 (2014-03-12)
On Fri, Apr 11, 2014 at 08:57:44PM +0200, Lukas Czerner wrote:
> Currently in do_fallocate in collapse range case we're checking whether
> offset + len is not bigger than i_size. However there is nothing which
> would prevent i_size from changing so the check is pointless. It should
> be done in the file system itself and the file system needs to make sure
> that i_size is not going to change.
> 
> As it is now we can easily crash kernel by having two processes doing
> truncate and fallocate collapse range at the same time. This can be
> reproduced on ext4 and it is theoretically possible on xfs even though I
> was not able to trigger it with this simple test.
> 
> This commit removes the check from do_fallocate and adds it to the file
> system.
> 
> Signed-off-by: Lukas Czerner <lczerner@xxxxxxxxxx>
> ---
>  fs/ext4/extents.c | 11 +++++++++--
>  fs/open.c         |  8 --------
>  fs/xfs/xfs_file.c | 10 +++++++++-
>  3 files changed, 18 insertions(+), 11 deletions(-)

Looks good to me.  Do the xfs folks mind if I carry this in the ext4
tree and push it to Linus shortly after -rc1?  If so, please send me
an ack'ed by.

If folks have a strong preference to handle this differently, let me
know.

Cheers,

                                                - Ted


<Prev in Thread] Current Thread [Next in Thread>