From mboxrd@z Thu Jan  1 00:00:00 1970
From: David Woodhouse <dwmw2@infradead.org>
Subject: Re: [PATCH v5] mtd: Do not corrupt backing device of device node
 inode
Date: Mon, 17 May 2010 15:10:41 +0100
Message-ID: <1274105441.6930.6920.camel@macbook.infradead.org>
References: <AANLkTimB1DsCd8HuMDc9jgVOnOAE_PVd2ktssp3TIwxc@mail.gmail.com>
Mime-Version: 1.0
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: 7bit
Cc: Jan Kara <jack@suse.cz>, Alexander Viro <viro@zeniv.linux.org.uk>,
	David Howells <dhowells@redhat.com>,
	Alexander Shishkin <virtuoso@slind.org>,
	Artem Bityutskiy <dedekind1@gmail.com>,
	linux-mtd@lists.infradead.org, linux-fsdevel@vger.kernel.org,
	stable@kernel.org, linux-kernel@vger.kernel.org
To: "Kirill A. Shutemov" <kirill@shutemov.name>
Return-path: <linux-fsdevel-owner@vger.kernel.org>
Received: from casper.infradead.org ([85.118.1.10]:47012 "EHLO
	casper.infradead.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org
	with ESMTP id S1750896Ab0EQOKw (ORCPT
	<rfc822;linux-fsdevel@vger.kernel.org>);
	Mon, 17 May 2010 10:10:52 -0400
In-Reply-To: <AANLkTimB1DsCd8HuMDc9jgVOnOAE_PVd2ktssp3TIwxc@mail.gmail.com>
Sender: linux-fsdevel-owner@vger.kernel.org
List-ID: <linux-fsdevel.vger.kernel.org>

On Mon, 2010-05-17 at 16:55 +0300, Kirill A. Shutemov wrote:
> +       if (!mtd->inode) {
> +               struct inode *mtd_inode;
> +               mtd_inode = new_inode(mtd_inode_mnt->mnt_sb);
> +               if (!mtd_inode) {
> +                       put_mtd_device(mtd);
> +                       ret = -ENOMEM;
> +                       goto out;
> +               }
> +               mtd_inode->i_mode = S_IFCHR;
> +               mtd_inode->i_rdev = inode->i_rdev;
> +               if (mtd->backing_dev_info) {
> +                       mtd_inode->i_data.backing_dev_info =
> +                               mtd->backing_dev_info;
> +               }
> +
> +               if (unlikely(cmpxchg(&mtd->inode, NULL, mtd_inode))) {
> +                       /* Somebody has already initialized mtd->inode
> */
> +                       iput(mtd_inode);
> +               }
> +       }
> +
> +       igrab(mtd->inode); 

Now you're just reimplementing iget().

Can we just use iget_locked() to get the appropriate inode (using the
mtd device number as the inode number)? Then we don't need to bother
storing it in mtd->inode at all, and we can iput() the last refcount on
it when the chardevice is closed -- we don't need it to stick around
until the MTD device is destroyed.

-- 
dwmw2