Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Date:   Wed, 20 Apr 2022 12:06:43 +0200
From:   Johan Hovold <johan@kernel.org>
To:     Oliver Neukum <oneukum@suse.com>
Cc:     Andy Shevchenko <andy.shevchenko@gmail.com>,
        Oleksij Rempel <linux@rempel-privat.de>,
        Dongliang Mu <dzm91@hust.edu.cn>,
        Oliver Neukum <oliver@neukum.org>,
        "David S. Miller" <davem@davemloft.net>,
        Jakub Kicinski <kuba@kernel.org>,
        Paolo Abeni <pabeni@redhat.com>,
        Dongliang Mu <mudongliangabcd@gmail.com>,
        syzbot+eabbf2aaa999cc507108@syzkaller.appspotmail.com,
        USB <linux-usb@vger.kernel.org>, netdev <netdev@vger.kernel.org>,
        Linux Kernel Mailing List <linux-kernel@vger.kernel.org>
Subject: Re: [PATCH] driver: usb: nullify dangling pointer in cdc_ncm_free
Message-ID: <Yl/bMxXKeNdLI87G@hovoldconsulting.com>
References: <20220409120901.267526-1-dzm91@hust.edu.cn>
 <YlQbqnYP/jcYinvz@hovoldconsulting.com>
 <CAHp75VeTqmdLhavZ+VbBYSFMDHr0FG4iKFGdbzE-wo5MCNikAA@mail.gmail.com>
 <d851497f-7960-b606-2f87-eb9bff89c8ac@suse.com>
 <Yl+utFmKEgILDFr5@hovoldconsulting.com>
 <aef0c568-e088-b897-f8ec-f22cfef124f6@suse.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <aef0c568-e088-b897-f8ec-f22cfef124f6@suse.com>
Precedence: bulk

On Wed, Apr 20, 2022 at 11:45:49AM +0200, Oliver Neukum wrote:

> >> -	if (dev->driver_info->unbind)
> >> -		dev->driver_info->unbind(dev, intf);
> >> +	if (dev->driver_info->disable)
> >> +		dev->driver_info->disable(dev, intf);
> >>  
> >>  	net = dev->net;
> >>  	unregister_netdev (net);
> >> @@ -1651,6 +1651,9 @@ void usbnet_disconnect (struct usb_interface *intf)
> >>  
> >>  	usb_scuttle_anchored_urbs(&dev->deferred);
> >>  
> >> +	if (dev->driver_info->unbind)
> >> +		dev->driver_info->unbind (dev, intf);
> >> +
> >>  	usb_kill_urb(dev->interrupt);

> > Don't you need to quiesce all I/O, including stopping the interrupt URB,
> > before unbind?

> If I do that, how do I prevent people from relaunching the URB between
> kill and unbind? Do I need to poison it?

You could, but it would seem you have bigger problems if something can
submit the URB after having deregistered the netdev.

Looks like the URB should already have been stopped by
usbnet_status_stop() so that the usb_kill_urb() above is (or should be)
a noop.

Johan