Re: [PATCH v9 2/6] iio: adc: ad4691: add initial driver for AD4691 family

[Andy Shevchenko <andriy.shevchenko@intel.com>]

On Fri, May 08, 2026 at 07:44:39AM +0300, Andy Shevchenko wrote:
> On Thu, May 07, 2026 at 03:15:49PM +0100, Jonathan Cameron wrote:
> > On Thu, 7 May 2026 09:26:00 +0000
> > "Sabau, Radu bogdan" <Radu.Sabau@analog.com> wrote:

...

> > > > +	st->info = spi_get_device_match_data(spi);  
> > > 
> > > "Is it possible for st->info to be NULL here?
> > > If the driver is manually bound to a device with a non-matching name using
> > > the sysfs bind interface in combination with driver_override,
> > > spi_get_device_match_data() could return NULL, which would cause a NULL
> > > pointer dereference later in the probe sequence."
> > > 
> > > Agreed, will add the NULL check with -ENODEV immediately after
> > > spi_get_device_match_data().
> > 
> > Andy, you seeing this one?  Looks like we are putting these checks back in again.
> > Whilst anyone forcing a bind like this is onto a looser anyway we shouldn't
> > crash due to a null dereference.
> 
> We should find a way how to disable that combination from the start. The driver
> makes no sense to be instantiated from user space. Actually most of the drivers
> nowadays should not be bound to the devices without driver data.

The 20260508095224.1275645-1-andriy.shevchenko@linux.intel.com has been just sent.

-- 
With Best Regards,
Andy Shevchenko