Re: KASAN: use-after-free Read in adu_disconnect

[Johan Hovold]

On Thu, Sep 19, 2019 at 12:35:04PM +0200, Johan Hovold wrote:
> On Fri, Aug 09, 2019 at 01:24:04PM -0700, syzbot wrote:
> > syzbot has found a reproducer for the following crash on:
> > 
> > HEAD commit:    e96407b4 usb-fuzzer: main usb gadget fuzzer driver
> > git tree:       https://github.com/google/kasan.git usb-fuzzer
> > console output: https://syzkaller.appspot.com/x/log.txt?x=13871a4a600000
> > kernel config:  https://syzkaller.appspot.com/x/.config?x=cfa2c18fb6a8068e
> > dashboard link: https://syzkaller.appspot.com/bug?extid=0243cb250a51eeefb8cc
> > compiler:       gcc (GCC) 9.0.0 20181231 (experimental)
> > syz repro:      https://syzkaller.appspot.com/x/repro.syz?x=11c4c8e2600000
> > C reproducer:   https://syzkaller.appspot.com/x/repro.c?x=11d80d2c600000
> > 
> > IMPORTANT: if you fix the bug, please add the following tag to the commit:
> > Reported-by: syzbot+0243cb250a51eeefb8cc@xxxxxxxxxxxxxxxxxxxxxxxxx
> > 
> > usb 1-1: USB disconnect, device number 4
> > ==================================================================
> > BUG: KASAN: use-after-free in atomic64_read  
> > include/asm-generic/atomic-instrumented.h:836 [inline]
> > BUG: KASAN: use-after-free in atomic_long_read  
> > include/asm-generic/atomic-long.h:28 [inline]
> > BUG: KASAN: use-after-free in __mutex_unlock_slowpath+0x96/0x670  
> > kernel/locking/mutex.c:1211
> > Read of size 8 at addr ffff8881d1d0aa00 by task kworker/0:1/12
> 
> Should be fixed by the below patch.
> 
> #syz test: https://github.com/google/kasan.git f0df5c1b

Hmm. Let's try that again with the commit id from the report...

#syz test: https://github.com/google/kasan.git e96407b4

Johan

> From 6f09430ae18085a1552fc641e53d3a3e678db6f3 Mon Sep 17 00:00:00 2001
> From: Johan Hovold <johan@xxxxxxxxxx>
> Date: Thu, 19 Sep 2019 11:48:38 +0200
> Subject: [PATCH] USB: adutux: fix use-after-free on disconnect
> 
> The driver was clearing its struct usb_device pointer, which it uses as
> an inverted disconnected flag, before deregistering the character device
> and without serialising against racing release().
> 
> This could lead to a use-after-free if a racing release() callback
> observes the cleared pointer and frees the driver data before
> disconnect() is finished with it.
> 
> This could also lead to NULL-pointer dereferences in a racing open().
> 
> Fixes: f08812d5eb8f ("USB: FIx locks and urb->status in adutux (updated)")
> Reported-by: syzbot+0243cb250a51eeefb8cc@xxxxxxxxxxxxxxxxxxxxxxxxx
> Cc: stable <stable@xxxxxxxxxxxxxxx>     # 2.6.24
> Signed-off-by: Johan Hovold <johan@xxxxxxxxxx>
> ---
> 
>  drivers/usb/misc/adutux.c | 7 ++++---
>  1 file changed, 4 insertions(+), 3 deletions(-)
> 
> diff --git a/drivers/usb/misc/adutux.c b/drivers/usb/misc/adutux.c
> index 344d523b0502..bcc138990e2f 100644
> --- a/drivers/usb/misc/adutux.c
> +++ b/drivers/usb/misc/adutux.c
> @@ -762,14 +762,15 @@ static void adu_disconnect(struct usb_interface *interface)
>  
>  	dev = usb_get_intfdata(interface);
>  
> -	mutex_lock(&dev->mtx);	/* not interruptible */
> -	dev->udev = NULL;	/* poison */
>  	usb_deregister_dev(interface, &adu_class);
> -	mutex_unlock(&dev->mtx);
>  
>  	mutex_lock(&adutux_mutex);
>  	usb_set_intfdata(interface, NULL);
>  
> +	mutex_lock(&dev->mtx);	/* not interruptible */
> +	dev->udev = NULL;	/* poison */
> +	mutex_unlock(&dev->mtx);
> +
>  	/* if the device is not opened, then we clean up right now */
>  	if (!dev->open_count)
>  		adu_delete(dev);