[bug report] iio: adc: adding support for MCP3564 ADC

[bug report] iio: adc: adding support for MCP3564 ADC

[Dan Carpenter]

Hello Marius Cristea,

The patch 33ec3e5fc1ea: "iio: adc: adding support for MCP3564 ADC"
from Aug 29, 2023 (linux-next), leads to the following Smatch static
checker warning:

	drivers/iio/adc/mcp3564.c:1426 mcp3564_probe()
	warn: address of NULL pointer 'indio_dev'

drivers/iio/adc/mcp3564.c
    1418 static int mcp3564_probe(struct spi_device *spi)
    1419 {
    1420         int ret;
    1421         struct iio_dev *indio_dev;
    1422         struct mcp3564_state *adc;
    1423 
    1424         indio_dev = devm_iio_device_alloc(&spi->dev, sizeof(*adc));
    1425         if (!indio_dev) {
--> 1426                 dev_err_probe(&indio_dev->dev, PTR_ERR(indio_dev),
                                       ^^^^^^^^^^^^^^^
This will crash and PTR_ERR(NULL) is useless.  Please could we just
delete this printk?  devm_iio_device_alloc() is going to print its own
error message.  (It doesn't print on every error path.  Like if we pass
something stupid instead of sizeof(*adc) then that looks like it might
not have a warning, but it's going to print a warning for everthing in
real life).

    1427                               "Can't allocate iio device\n");
    1428                 return -ENOMEM;
    1429         }
    1430 
    1431         adc = iio_priv(indio_dev);
    1432         adc->spi = spi;
    1433 

regards,
dan carpenter

Re: [bug report] iio: adc: adding support for MCP3564 ADC

[Dan Carpenter]

On Thu, Sep 14, 2023 at 01:01:02PM +0300, Dan Carpenter wrote:
> Hello Marius Cristea,
> 
> The patch 33ec3e5fc1ea: "iio: adc: adding support for MCP3564 ADC"
> from Aug 29, 2023 (linux-next), leads to the following Smatch static
> checker warning:
> 
> 	drivers/iio/adc/mcp3564.c:1426 mcp3564_probe()
> 	warn: address of NULL pointer 'indio_dev'
> 
> drivers/iio/adc/mcp3564.c
>     1418 static int mcp3564_probe(struct spi_device *spi)
>     1419 {
>     1420         int ret;
>     1421         struct iio_dev *indio_dev;
>     1422         struct mcp3564_state *adc;
>     1423 
>     1424         indio_dev = devm_iio_device_alloc(&spi->dev, sizeof(*adc));
>     1425         if (!indio_dev) {
> --> 1426                 dev_err_probe(&indio_dev->dev, PTR_ERR(indio_dev),
>                                        ^^^^^^^^^^^^^^^
> This will crash

Actually I think that dev_err_probe() has a check for this kind of bug
so it doesn't actually crash.  The check is burried deep into the call
tree where it prints the name and if the address is < PAGE_SIZE it just
prints the device name as "(efault)".  But it's still not ideal.

regads,
dan carpenter

Re: [bug report] iio: adc: adding support for MCP3564 ADC

[Marius.Cristea]

Hello Dan Carpenter,


  Thank you for your feedback.
Indeed, it will not add more information in this case so sure I can
remove that line.

Please, tell me how it is best to procced next: submitting another
PATCH (V6) in this case, or to create a Fixes to the PATCH_V5?

Best Regards,
Marius


On Thu, 2023-09-14 at 14:29 +0300, Dan Carpenter wrote:
> EXTERNAL EMAIL: Do not click links or open attachments unless you
> know the content is safe
> 
> On Thu, Sep 14, 2023 at 01:01:02PM +0300, Dan Carpenter wrote:
> > Hello Marius Cristea,
> > 
> > The patch 33ec3e5fc1ea: "iio: adc: adding support for MCP3564 ADC"
> > from Aug 29, 2023 (linux-next), leads to the following Smatch
> > static
> > checker warning:
> > 
> >       drivers/iio/adc/mcp3564.c:1426 mcp3564_probe()
> >       warn: address of NULL pointer 'indio_dev'
> > 
> > drivers/iio/adc/mcp3564.c
> >     1418 static int mcp3564_probe(struct spi_device *spi)
> >     1419 {
> >     1420         int ret;
> >     1421         struct iio_dev *indio_dev;
> >     1422         struct mcp3564_state *adc;
> >     1423
> >     1424         indio_dev = devm_iio_device_alloc(&spi->dev,
> > sizeof(*adc));
> >     1425         if (!indio_dev) {
> > --> 1426                 dev_err_probe(&indio_dev->dev,
> > PTR_ERR(indio_dev),
> >                                        ^^^^^^^^^^^^^^^
> > This will crash
> 
> Actually I think that dev_err_probe() has a check for this kind of
> bug
> so it doesn't actually crash.  The check is burried deep into the
> call
> tree where it prints the name and if the address is < PAGE_SIZE it
> just
> prints the device name as "(efault)".  But it's still not ideal.
> 
> regads,
> dan carpenter
> 

Re: [bug report] iio: adc: adding support for MCP3564 ADC

[Dan Carpenter]

On Thu, Sep 14, 2023 at 01:19:33PM +0000, Marius.Cristea@microchip.com wrote:
> Hello Dan Carpenter,
> 
> 
>   Thank you for your feedback.
> Indeed, it will not add more information in this case so sure I can
> remove that line.
> 
> Please, tell me how it is best to procced next: submitting another
> PATCH (V6) in this case, or to create a Fixes to the PATCH_V5?
> 

I believe that you should send a follow on patch instead of sending
another rev.

regards,
dan carpenter

Re: [bug report] iio: adc: adding support for MCP3564 ADC

[Jonathan Cameron]

On Thu, 14 Sep 2023 16:29:26 +0300
Dan Carpenter <dan.carpenter@linaro.org> wrote:

> On Thu, Sep 14, 2023 at 01:19:33PM +0000, Marius.Cristea@microchip.com wrote:
> > Hello Dan Carpenter,
> > 
> > 
> >   Thank you for your feedback.
> > Indeed, it will not add more information in this case so sure I can
> > remove that line.
> > 
> > Please, tell me how it is best to procced next: submitting another
> > PATCH (V6) in this case, or to create a Fixes to the PATCH_V5?
> >   
> 
> I believe that you should send a follow on patch instead of sending
> another rev.
> 
> regards,
> dan carpenter
> 
> 
Yes - at this stage follow on patch preferred.

Jonathan