From mboxrd@z Thu Jan  1 00:00:00 1970
From: Al Viro <viro@ZenIV.linux.org.uk>
Subject: Re: [PATCH 4/4] xfs: Protect xfs_file_aio_write() with
 sb_start_write - sb_end_write
Date: Thu, 12 Jan 2012 21:29:41 +0000
Message-ID: <20120112212941.GI23916@ZenIV.linux.org.uk>
References: <1326331253-6497-1-git-send-email-jack@suse.cz>
 <1326331253-6497-5-git-send-email-jack@suse.cz>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Cc: linux-fsdevel@vger.kernel.org, LKML <linux-kernel@vger.kernel.org>,
	linux-ext4@vger.kernel.org, xfs@oss.sgi.com,
	Eric Sandeen <sandeen@sandeen.net>,
	Dave Chinner <dchinner@redhat.com>,
	Surbhi Palande <csurbhi@gmail.com>,
	Kamal Mostafa <kamal@canonical.com>,
	Christoph Hellwig <hch@infradead.org>, Ben Myers <bpm@sgi.com>,
	Alex Elder <elder@kernel.org>
To: Jan Kara <jack@suse.cz>
Return-path: <linux-ext4-owner@vger.kernel.org>
Content-Disposition: inline
In-Reply-To: <1326331253-6497-5-git-send-email-jack@suse.cz>
Sender: linux-ext4-owner@vger.kernel.org
List-Id: linux-fsdevel.vger.kernel.org

On Thu, Jan 12, 2012 at 02:20:53AM +0100, Jan Kara wrote:
> Replace racy xfs_wait_for_freeze() check with reliable sb_start_write()
> - sb_end_write() locking. Due to lock ranking dictated by the page fault
> code we have to call sb_start_write() after we acquire ilock.
> 
> CC: Ben Myers <bpm@sgi.com>
> CC: Alex Elder <elder@kernel.org>
> Signed-off-by: Jan Kara <jack@suse.cz>
> ---
>  fs/xfs/xfs_file.c |    6 ++++--
>  1 files changed, 4 insertions(+), 2 deletions(-)
> 
> diff --git a/fs/xfs/xfs_file.c b/fs/xfs/xfs_file.c
> index 753ed9b..c5f879b 100644
> --- a/fs/xfs/xfs_file.c
> +++ b/fs/xfs/xfs_file.c
> @@ -862,9 +862,11 @@ xfs_file_dio_aio_write(
>  		*iolock = XFS_IOLOCK_SHARED;
>  	}
>  
> +	sb_start_write(inode->i_sb);
>  	trace_xfs_file_direct_write(ip, count, iocb->ki_pos, 0);
>  	ret = generic_file_direct_write(iocb, iovp,
>  			&nr_segs, pos, &iocb->ki_pos, count, ocount);
> +	sb_end_write(inode->i_sb);

What lock ordering do you have in mind?  Explicit description in fs/super.c,
please...