Re: [PATCH] SELinux: Fix possible NULL pointer dereference in selinux_inode_permission()

From: Eric Paris <eparis@redhat.com>
To: Steven Rostedt <rostedt@goodmis.org>
Cc: Alexander Viro <viro@zeniv.linux.org.uk>,
	LKML <linux-kernel@vger.kernel.org>,
	Stephen Smalley <sds@tycho.nsa.gov>,
	James Morris <james.l.morris@oracle.com>,
	Paul Moore <paul@paul-moore.com>,
	Andrew Morton <akpm@linux-foundation.org>,
	"Paul E. McKenney" <paulmck@linux.vnet.ibm.com>,
	stable <stable@vger.kernel.org>,
	selinux@tycho.nsa.gov, linux-security-module@vger.kernel.org
Subject: Re: [PATCH] SELinux: Fix possible NULL pointer dereference in selinux_inode_permission()
Date: Thu, 09 Jan 2014 11:05:45 -0500	[thread overview]
Message-ID: <1389283545.15209.59.camel@localhost> (raw)
In-Reply-To: <1389283030.15209.56.camel@localhost>

[adding lsm and selinux]

Am I just crazy, or was this bug discussed (and obviously not fixed)
some time ago?

VFS can still use inodes after security_inode_free_security() was
called...

On Thu, 2014-01-09 at 10:57 -0500, Eric Paris wrote:
> On Thu, 2014-01-09 at 10:51 -0500, Steven Rostedt wrote:
> > On Thu, 9 Jan 2014 10:31:55 -0500
> > Eric Paris <eparis@parisplace.org> wrote:
> > 
> > > Didn't Al find this/something very similar.  I really hate this
> > 
> > I'm not involved with the vfs, so I'm unaware of other solutions
> > presented. I just hit this now and solving bugs is where I get a chance
> > to learn about other aspects of the kernel. ;-)
> > 
> > > solution.  Why should every LSM try to understand the intimate
> > > lifetime rules of the parent subsystems?  The real problem is that
> > > inode_free_security() is being called while the inode is still in use.
> > >  While I agree with the assessment, I disagree with the solution.  Let
> > > me try to find where Al and Christoph talked about this....
> > > 
> > 
> > The other obvious solution (but not as trivial to implement) is to call
> > the security_inode_free() and friends (probably __destroy_inode()
> > itself) after a synchronize_rcu().
> > 
> > Perhaps something like this?
> 
> I can't for the life of me find that conversation!  Maybe I'm just
> making it all up...  Usually I forget conversations, not remember ones
> that didn't happen...
> 
> Assuming the VFS guys say that delaying __destroy_inode() is safe like
> that, I like it better.  It also means that this is fixed for all LSMs,
> not just SELinux...
> 
> -Eric
> 
> > 
> > -- Steve
> > 
> > diff --git a/fs/inode.c b/fs/inode.c
> > index 4bcdad3..a8f3b88 100644
> > --- a/fs/inode.c
> > +++ b/fs/inode.c
> > @@ -252,16 +252,17 @@ EXPORT_SYMBOL(__destroy_inode);
> >  static void i_callback(struct rcu_head *head)
> >  {
> >  	struct inode *inode = container_of(head, struct inode, i_rcu);
> > +	__destroy_inode(inode);
> >  	kmem_cache_free(inode_cachep, inode);
> >  }
> >  
> >  static void destroy_inode(struct inode *inode)
> >  {
> >  	BUG_ON(!list_empty(&inode->i_lru));
> > -	__destroy_inode(inode);
> > -	if (inode->i_sb->s_op->destroy_inode)
> > +	if (inode->i_sb->s_op->destroy_inode) {
> > +		__destroy_inode(inode);
> >  		inode->i_sb->s_op->destroy_inode(inode);
> > -	else
> > +	} else
> >  		call_rcu(&inode->i_rcu, i_callback);
> >  }
> >  
> 