Date: Wed, 8 Jul 2015 16:53:21 -0700
From: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
To: Jason Gunthorpe <jgunthorpe@obsidianresearch.com>
Cc: Dmitry Torokhov <dmitry.torokhov@gmail.com>,
        Jens Wiklander <jens.wiklander@linaro.org>,
        linux-kernel@vger.kernel.org, linux-arm-kernel@lists.infradead.org,
        devicetree@vger.kernel.org, Arnd Bergmann <arnd@arndb.de>,
        Rob Herring <robh+dt@kernel.org>,
        Herbert Xu <herbert@gondor.apana.org.au>, valentin.manea@huawei.com,
        jean-michel.delorme@st.com, emmanuel.michel@st.com, javier@javigon.com,
        Mark Rutland <mark.rutland@arm.com>,
        Michal Simek <michal.simek@xilinx.com>
Subject: Re: [PATCH v4 3/5] tee: generic TEE subsystem
Message-ID: <20150708235321.GB12393@kroah.com>
References: <1436350592-7732-1-git-send-email-jens.wiklander@linaro.org>
 <1436350592-7732-4-git-send-email-jens.wiklander@linaro.org>
 <20150708171026.GA11740@obsidianresearch.com>
 <20150708211129.GA29824@kroah.com>
 <20150708222649.GA20068@obsidianresearch.com>
 <20150708223325.GA5843@kroah.com>
 <20150708231612.GB20068@obsidianresearch.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20150708231612.GB20068@obsidianresearch.com>
User-Agent: Mutt/1.5.23+89 (0255b37be491) (2014-03-12)
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 2073
Lines: 56

On Wed, Jul 08, 2015 at 05:16:12PM -0600, Jason Gunthorpe wrote:
> On Wed, Jul 08, 2015 at 03:33:25PM -0700, Greg Kroah-Hartman wrote:
> > > The basic issue is that cdev_del doesn't seem to be synchronizing.
> > > 
> > > The use after free race is then something like:
> > > 
> > >    struct tpm_chip {
> > >  	struct device dev;
> > > 	struct cdev cdev;
> > 
> > Oops, right there's your problem.  You can't have two reference counted
> > objects trying to manage the memory of a single structure.  No matter
> > what you do, it's going to be a pain to deal with this, so don't :)
> 
> Sure, generally, yes, but that isn't done for no reason, it is to make
> open straightforward:
> 
> static int tpm_open(struct inode *inode, struct file *file)
> {
> 	struct tpm_chip *chip =
> 		container_of(inode->i_cdev, struct tpm_chip, cdev);
> 
> We need to recover the tpm_chip associated with the char device
> node, in a way that is holding a kref on it, without racing with
> cdev_del/etc
> 
> This scheme does mean that if we have a struct file we have a kref on
> the cdev, and if we have cdev then we have a kref on the tpm_chip,
> which is really easy to use properly.
> 
> > > Ie we need cdev to hold a ref on tpm_chip->dev until cdev_put is
> > > called.
> > 
> > No, separate them, make the cdev a pointer and all should be fine.
> 
> Okay, cdev_alloc takes care of the cdev lifetime.
> 
> Do you have a simple solution to replace container_of as well?
> 
> What would you think about something like:
> 
>  cdev_alloc(&chip->dev.kref)

Just pick either the cdev to handle the lifetime rules, or the struct
device, you'll still need a container_of().  Just don't do both as odds
are the lifetime rules is going to get really hard to debug and ensure
that everything is correct on the shutdown/release path.

thanks,

greg k-h
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/