[PATCH] usb: gadget: Check for NULL pointer in disconnect

[PATCH] usb: gadget: Check for NULL pointer in disconnect

[Badhri Jagan Sridharan]

Added a safety net to make sure that
composite_disconnect does not end up disconneting
a NULL device. Prevents NULL pointer crash.

Signed-off-by: Badhri Jagan Sridharan <Badhri@google.com>
---
 drivers/usb/gadget/composite.c | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/drivers/usb/gadget/composite.c b/drivers/usb/gadget/composite.c
index 13adfd1..90b37bd 100644
--- a/drivers/usb/gadget/composite.c
+++ b/drivers/usb/gadget/composite.c
@@ -1823,6 +1823,11 @@ void composite_disconnect(struct usb_gadget *gadget)
 	struct usb_composite_dev	*cdev = get_gadget_data(gadget);
 	unsigned long			flags;
 
+	if (!cdev) {
+		WARN(1, "Trying to disconnect a NULL composite device\n");
+		return;
+	}
+
 	/* REVISIT:  should we have config and device level
 	 * disconnect callbacks?
 	 */
-- 
2.2.0.rc0.207.ga3a616c

Re: [PATCH] usb: gadget: Check for NULL pointer in disconnect

[Peter Chen]

On Fri, Mar 20, 2015 at 04:40:52PM -0700, Badhri Jagan Sridharan wrote:
> Added a safety net to make sure that
> composite_disconnect does not end up disconneting
> a NULL device. Prevents NULL pointer crash.
> 
> Signed-off-by: Badhri Jagan Sridharan <Badhri@google.com>
> ---
>  drivers/usb/gadget/composite.c | 5 +++++
>  1 file changed, 5 insertions(+)
> 
> diff --git a/drivers/usb/gadget/composite.c b/drivers/usb/gadget/composite.c
> index 13adfd1..90b37bd 100644
> --- a/drivers/usb/gadget/composite.c
> +++ b/drivers/usb/gadget/composite.c
> @@ -1823,6 +1823,11 @@ void composite_disconnect(struct usb_gadget *gadget)
>  	struct usb_composite_dev	*cdev = get_gadget_data(gadget);
>  	unsigned long			flags;
>  
> +	if (!cdev) {
> +		WARN(1, "Trying to disconnect a NULL composite device\n");
> +		return;
> +	}
> +

Do you really see some udc drivers call it after composite_unbind is
called? If it is, you may add dump_stack() to track that error.

Besides this, function suspended_show is needed to add cdev NULL pointer
checking.

>  	/* REVISIT:  should we have config and device level
>  	 * disconnect callbacks?
>  	 */
> -- 
> 2.2.0.rc0.207.ga3a616c
> 
> --
> To unsubscribe from this list: send the line "unsubscribe linux-usb" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html

-- 

Best Regards,
Peter Chen

Re: [PATCH] usb: gadget: Check for NULL pointer in disconnect

[Badhri Jagan Sridharan]

 > Do you really see some udc drivers call it after composite_unbind is
 > called? If it is, you may add dump_stack() to track that error.
 >
 > Besides this, function suspended_show is needed to add cdev NULL pointer
 > checking.

We see this happening occasionally in *not yet* upstreamed UDC
code of some vendors (Yes, disconnect being called after unbind)
After reviewing the entire composite.c file, I did notice that
none of the functions check for NULL pointer when cdev is
obtained from get_gadget_data.
Is crashing/bringing down the whole kernel intentionally
left to happen ? Isn't printing a WARN/ERROR msg and
returning not the preferable approach ?

On 03/22/2015 12:43 AM, Peter Chen wrote:
> On Fri, Mar 20, 2015 at 04:40:52PM -0700, Badhri Jagan Sridharan wrote:
>> Added a safety net to make sure that
>> composite_disconnect does not end up disconneting
>> a NULL device. Prevents NULL pointer crash.
>>
>> Signed-off-by: Badhri Jagan Sridharan <Badhri@google.com>
>> ---
>>   drivers/usb/gadget/composite.c | 5 +++++
>>   1 file changed, 5 insertions(+)
>>
>> diff --git a/drivers/usb/gadget/composite.c b/drivers/usb/gadget/composite.c
>> index 13adfd1..90b37bd 100644
>> --- a/drivers/usb/gadget/composite.c
>> +++ b/drivers/usb/gadget/composite.c
>> @@ -1823,6 +1823,11 @@ void composite_disconnect(struct usb_gadget *gadget)
>>   	struct usb_composite_dev	*cdev = get_gadget_data(gadget);
>>   	unsigned long			flags;
>>
>> +	if (!cdev) {
>> +		WARN(1, "Trying to disconnect a NULL composite device\n");
>> +		return;
>> +	}
>> +
>
> Do you really see some udc drivers call it after composite_unbind is
> called? If it is, you may add dump_stack() to track that error.
>
> Besides this, function suspended_show is needed to add cdev NULL pointer
> checking.
>
>>   	/* REVISIT:  should we have config and device level
>>   	 * disconnect callbacks?
>>   	 */
>> --
>> 2.2.0.rc0.207.ga3a616c
>>
>> --
>> To unsubscribe from this list: send the line "unsubscribe linux-usb" in
>> the body of a message to majordomo@vger.kernel.org
>> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>

Re: [PATCH] usb: gadget: Check for NULL pointer in disconnect

[Peter Chen]

On Mon, Mar 23, 2015 at 04:00:37PM -0700, Badhri Jagan Sridharan wrote:
> > Do you really see some udc drivers call it after composite_unbind is
> > called? If it is, you may add dump_stack() to track that error.
> >
> > Besides this, function suspended_show is needed to add cdev NULL pointer
> > checking.
> 
> We see this happening occasionally in *not yet* upstreamed UDC
> code of some vendors (Yes, disconnect being called after unbind)

Does it a spurious interrupt?
I guess if you add struct usb_gadget_driver NULL pointer check check, it
will not occur:
	if (you_udc->driver)
		you_udc->driver->disconnect(&you_udc->gadget);
 
> After reviewing the entire composite.c file, I did notice that
> none of the functions check for NULL pointer when cdev is
> obtained from get_gadget_data.
> Is crashing/bringing down the whole kernel intentionally
> left to happen ?

Hmm, I don't think, the design may think it never occurs.

> Isn't printing a WARN/ERROR msg and
> returning not the preferable approach ?

I think BUG_ON is more suitable, let's see Felipe's comment.

> 
> On 03/22/2015 12:43 AM, Peter Chen wrote:
> >On Fri, Mar 20, 2015 at 04:40:52PM -0700, Badhri Jagan Sridharan wrote:
> >>Added a safety net to make sure that
> >>composite_disconnect does not end up disconneting
> >>a NULL device. Prevents NULL pointer crash.
> >>
> >>Signed-off-by: Badhri Jagan Sridharan <Badhri@google.com>
> >>---
> >>  drivers/usb/gadget/composite.c | 5 +++++
> >>  1 file changed, 5 insertions(+)
> >>
> >>diff --git a/drivers/usb/gadget/composite.c b/drivers/usb/gadget/composite.c
> >>index 13adfd1..90b37bd 100644
> >>--- a/drivers/usb/gadget/composite.c
> >>+++ b/drivers/usb/gadget/composite.c
> >>@@ -1823,6 +1823,11 @@ void composite_disconnect(struct usb_gadget *gadget)
> >>  	struct usb_composite_dev	*cdev = get_gadget_data(gadget);
> >>  	unsigned long			flags;
> >>
> >>+	if (!cdev) {
> >>+		WARN(1, "Trying to disconnect a NULL composite device\n");
> >>+		return;
> >>+	}
> >>+
> >
> >Do you really see some udc drivers call it after composite_unbind is
> >called? If it is, you may add dump_stack() to track that error.
> >
> >Besides this, function suspended_show is needed to add cdev NULL pointer
> >checking.
> >
> >>  	/* REVISIT:  should we have config and device level
> >>  	 * disconnect callbacks?
> >>  	 */
> >>--
> >>2.2.0.rc0.207.ga3a616c
> >>
> >>--
> >>To unsubscribe from this list: send the line "unsubscribe linux-usb" in
> >>the body of a message to majordomo@vger.kernel.org
> >>More majordomo info at  http://vger.kernel.org/majordomo-info.html
> >

-- 

Best Regards,
Peter Chen

Re: [PATCH] usb: gadget: Check for NULL pointer in disconnect

[Felipe Balbi]

[-- Attachment #1: Type: text/plain, Size: 568 bytes --]

On Mon, Mar 23, 2015 at 04:00:37PM -0700, Badhri Jagan Sridharan wrote:
> > Do you really see some udc drivers call it after composite_unbind is
> > called? If it is, you may add dump_stack() to track that error.
> >
> > Besides this, function suspended_show is needed to add cdev NULL pointer
> > checking.
> 
> We see this happening occasionally in *not yet* upstreamed UDC
> code of some vendors (Yes, disconnect being called after unbind)

fix your UDC. We leave these sort of thing open so people realise they
have to fix their drivers.

-- 
balbi

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 819 bytes --]