From: Louis Rilling <Louis.Rilling@kerlabs.com>
To: Joel.Becker@oracle.com
Cc: linux-fsdevel@vger.kernel.org, cluster-devel@redhat.com,
ocfs2-devel@oss.oracle.com
Subject: Re: [Ocfs2-devel] [RFC] configfs: Pin configfs subsystems separately from new config_items.
Date: Tue, 24 Jun 2008 07:04:23 +0200 [thread overview]
Message-ID: <20080624050423.GA4193@localdomain> (raw)
In-Reply-To: <20080623191027.GF592@mail.oracle.com>
On Mon, Jun 23, 2008 at 12:10:28PM -0700, Joel Becker wrote:
> On Mon, Jun 23, 2008 at 05:44:57PM +0200, Louis Rilling wrote:
> > make_item()
> > new_item = kmalloc();
> > config_item_init_type_long_name();
> > return new_item;
> >
> > drop_item(item)
> > config_item_put(item);
> > kfree(item);
>
> This is never, ever safe. Consider that someone has an
> attribute file open - it has a reference to the item. You can still
> rmdir() the item - doing stuff to the attribute after drop_item() will
> just get ignored. But you can't free it in drop_item().
Yup, I realized it this night (prevented me from sleeping by the way).
The previous arguments remain valid though. I hope that you won't discard
them because of this buggy one :)
While looking at file.c, I found this in configfs_release():
struct config_item * item = to_item(filp->f_path.dentry->d_parent);
[...]
if (item)
config_item_put(item);
It looks strange:
1/ either item may be NULL, and there is a probable memory leak because of the
reference grabbed in configfs_open_file();
2/ or item may never be NULL and this check is just useless and (at least for
me) misleading.
If I understand correctly, option 2 is correct because
a/ even if .dentry gets unhashed and .dentry->d_parent gets unhashed as well,
VFS ensures that filp->f_path.dentry->d_parent is unchanged unless .dentry was
renamed, which is not permitted by configfs, and, I guess, will never be
permitted for a configfs attribute;
b/ once dentry->d_fsdata points to a configfs_dirent, it never changes
and remains valid for the rest of dentry's life (until dentry_iput());
c/ configfs_dirent->s_element never changes once it is set.
Am I wrong somewhere?
Thanks
Louis
--
Dr Louis Rilling Kerlabs - IRISA
Skype: louis.rilling Campus Universitaire de Beaulieu
Phone: (+33|0) 2 99 84 71 52 Avenue du General Leclerc
Fax: (+33|0) 2 99 84 71 71 35042 Rennes CEDEX - France
http://www.kerlabs.com/
next prev parent reply other threads:[~2008-06-24 5:04 UTC|newest]
Thread overview: 13+ messages / expand[flat|nested] mbox.gz Atom feed top
2008-06-17 22:41 [RFC] configfs: Pin configfs subsystems separately from new config_items Joel Becker
2008-06-18 12:31 ` [Ocfs2-devel] " Louis Rilling
2008-06-18 16:12 ` Joel Becker
2008-06-18 16:51 ` Louis Rilling
2008-06-18 20:07 ` Joel Becker
2008-06-19 11:13 ` [Ocfs2-devel] " Louis Rilling
2008-06-19 22:07 ` Joel Becker
2008-06-20 12:46 ` Louis Rilling
2008-06-20 22:36 ` Joel Becker
2008-06-23 15:44 ` Louis Rilling
2008-06-23 19:10 ` Joel Becker
2008-06-24 5:04 ` Louis Rilling [this message]
2008-06-24 17:03 ` Joel Becker
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20080624050423.GA4193@localdomain \
--to=louis.rilling@kerlabs.com \
--cc=Joel.Becker@oracle.com \
--cc=cluster-devel@redhat.com \
--cc=linux-fsdevel@vger.kernel.org \
--cc=ocfs2-devel@oss.oracle.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).