Hi, On 11/24/22 12:35, Andy Shevchenko wrote: > On Thu, Nov 24, 2022 at 1:07 PM Stefan Binding > <sbinding@xxxxxxxxxxxxxxxxxxxxx> wrote: >> >> This allows the i2c driver to obtain the ACPI_COMPANION. > > As far as I get how it's done in the SPI case the real fix should lie > among i2c_acpi_new_device_by_fwnode(), right? Eventually maybe, but not for the initial change. It is complicated, making this change has side-effects and we want to limit those side-effects to only i2c-clients instantiated from serial-multi-instantiate for now, see my other reply to this patch. I do believe that we eventually want to make this change, to easily give drivers access to all sorts of info (e.g. _DSM methods) from the matching ACPI fw-node, but as I said it is complicated... Regards, Hans
