smatch stuff: serial: ifx6x60: oops on load

smatch stuff: serial: ifx6x60: oops on load

[Dan Carpenter]

Smatch complains that the ifx6x60 driver in linux-next has a NULL
dereference.

drivers/serial/ifx6x60.c +986 ifx_spi_spi_probe(27)
	error: potential null derefence 'pl_data'.

   984          ifx_dev->spi_timer.function = ifx_spi_timeout;
   985          ifx_dev->spi_timer.data = (unsigned long)ifx_dev;
   986          ifx_dev->is_6160 = pl_data->is_6160;
                                   ^^^^^^^
"pl_data" is always NULL at this point.  Presumably this crashes on load
every time.

regards,
dan carpenter

RE: smatch stuff: serial: ifx6x60: oops on load

[Gorby, Russ]

Yes, unfortunately there are a few issues with this driver that prevent it from loading and running properly.
I will be proving follow-on patches to address this and other issues that have been discovered that should address this. 

>>-----Original Message-----
>>From: Dan Carpenter [mailto:error27@gmail.com]
>>Sent: Monday, December 20, 2010 1:16 AM
>>To: Gorby, Russ
>>Cc: Alan Cox; linux-kernel@vger.kernel.org
>>Subject: smatch stuff: serial: ifx6x60: oops on load
>>
>>Smatch complains that the ifx6x60 driver in linux-next has a NULL
>>dereference.
>>
>>drivers/serial/ifx6x60.c +986 ifx_spi_spi_probe(27)
>>	error: potential null derefence 'pl_data'.
>>
>>   984          ifx_dev->spi_timer.function = ifx_spi_timeout;
>>   985          ifx_dev->spi_timer.data = (unsigned long)ifx_dev;
>>   986          ifx_dev->is_6160 = pl_data->is_6160;
>>                                   ^^^^^^^
>>"pl_data" is always NULL at this point.  Presumably this crashes on
>load
>>every time.
>>
>>regards,
>>dan carpenter

Re: smatch stuff: serial: ifx6x60: oops on load

[Alan Cox]

On Mon, 20 Dec 2010 12:15:55 +0300
Dan Carpenter <error27@gmail.com> wrote:

> Smatch complains that the ifx6x60 driver in linux-next has a NULL
> dereference.
> 
> drivers/serial/ifx6x60.c +986 ifx_spi_spi_probe(27)
> 	error: potential null derefence 'pl_data'.
> 
>    984          ifx_dev->spi_timer.function = ifx_spi_timeout;
>    985          ifx_dev->spi_timer.data = (unsigned long)ifx_dev;
>    986          ifx_dev->is_6160 = pl_data->is_6160;
>                                    ^^^^^^^
> "pl_data" is always NULL at this point.  Presumably this crashes on load
> every time.

Getting the platform enabling patch in is on the todo list but it depends
on some other bits which are also not yet in the tree.

is_6160 needs to go anyway as it's not what the platform data really
indicates.