On Fri, Jul 17, 2026 at 10:31:26AM +0100, Rodrigo Alencar wrote: > On 17/07/26 11:15, Andy Shevchenko wrote: > > On Fri, Jul 17, 2026 at 07:51:35AM +0100, Rodrigo Alencar wrote: > > > On 16/07/26 21:42, Andy Shevchenko wrote: > > > > On Thu, Jul 16, 2026 at 01:14:18PM +0100, Rodrigo Alencar via B4 Relay > > > > wrote:
... > > > > > struct iio_dev *indio_dev; > > > > > int ret, i; > > > > > > > > Move an assignment here as well. > > > > > > > > chip_info = ... > > > > > > assignment? this is an input param validation > > > > Can we make sure we always get a correct one to begin with? > > With that an assignment and check can be coupled together. > > Also consider use -ENODATA as it's most likely comes from > > driver_data. > > I suppose you are suggesting to move the check to the bus code with: > > info = i2c_get_match_data(i2c); > if (!info) > return -ENODATA; > > rather than here in the core/common code. Yes! This avoids layering violation and makes the API contract cleaner. > > > > > + if (!chip_info) > > > > > + return -ENODEV; -- With Best Regards, Andy Shevchenko
