Re: [PATCH next] iio: adc: ad4851: Fix signedness bug in ad4851_calibrate()

[Jonathan Cameron]

On Fri, 28 Feb 2025 13:24:20 +0000
Nuno Sá <noname.nuno@xxxxxxxxx> wrote:

> On Fri, 2025-02-28 at 12:36 +0300, Dan Carpenter wrote:
> > The "c" variable is used to store error codes from ad4851_find_opt()
> > so it has to be signed for the error handling to work.  Change it
> > to type int.
> > 
> > Fixes: 6250803fe2ec ("iio: adc: ad4851: add ad485x driver")
> > Signed-off-by: Dan Carpenter <dan.carpenter@xxxxxxxxxx>
> > ---  
> 
> Reviewed-by: Nuno Sá <nuno.sa@xxxxxxxxxx>
Julia + 0-day / coccinelle caught this one as well but seeing as you
were both first to report and you sent me a patch I'll just apply this one.

Applied to the togreg branch of iio.git.

Thanks,

Jonathan

> 
> > Btw, in ad4851_write_raw() there are some checks for negative but
> > the multiplication could still wrap and turn negative so we might want to
> > add an upper bounds as well:
> > 
> >     drivers/iio/adc/ad4851.c:826 ad4851_write_raw()
> >     warn: potential integer overflow from user 'val * st->osr'
> > 
> >  drivers/iio/adc/ad4851.c | 4 ++--
> >  1 file changed, 2 insertions(+), 2 deletions(-)
> > 
> > diff --git a/drivers/iio/adc/ad4851.c b/drivers/iio/adc/ad4851.c
> > index 1ad37084355e..98ebc853db79 100644
> > --- a/drivers/iio/adc/ad4851.c
> > +++ b/drivers/iio/adc/ad4851.c
> > @@ -492,11 +492,11 @@ static int ad4851_find_opt(const unsigned long *field,
> > unsigned int start,
> >  static int ad4851_calibrate(struct iio_dev *indio_dev)
> >  {
> >  	struct ad4851_state *st = iio_priv(indio_dev);
> > -	unsigned int opt_delay, num_lanes, delay, i, s, c;
> > +	unsigned int opt_delay, num_lanes, delay, i, s;
> >  	enum iio_backend_interface_type interface_type;
> >  	DECLARE_BITMAP(pn_status, AD4851_MAX_LANES * AD4851_MAX_IODELAY);
> >  	bool status;
> > -	int ret;
> > +	int c, ret;
> >  
> >  	ret = iio_backend_interface_type_get(st->back, &interface_type);
> >  	if (ret)  
> 
>