On Mon, 1 Sep 2025 15:43:08 +0300 Andy Shevchenko <andriy.shevche...@intel.com> wrote:
> On Sun, Aug 31, 2025 at 10:19:38PM +0200, Duje Mihanović wrote: > > On Sunday, 31 August 2025 21:24:54 Central European Summer Time Karel Balej > > wrote: > > > Duje Mihanović, 2025-08-31T12:33:05+02:00: > > ... > > > > > + iio->dev.of_node = dev->parent->of_node; > > > > > > Didn't you mean to drop this since Jonathan pointed out that it's done > > > by the core? > > > > Actually I have found that device tree consumers fail to get their IO > > channels without this line, so I left it. > > because the passed device is not parent? Good point. I missed it seems to be going up another level. Absolutely fine to override in that case. > > In any case this line is problematic. Use device_set_node() instead with the > proper dev_fwnode() parameter. >
