On Tue, Nov 03, 2020 at 01:42:47PM +0100, Benjamin Tissoires wrote: > > I also want to say that I like the general idea of Doug's patch. > Having a separate driver that handles the specific use case of goodix > is really nice, as it allows to just load this driver without touching > the core of i2c-hid. I believe this is in line with what Google tries > to do with their kernel that OEMs can not touch, but only add overlays > to it. The implementation is not polished (I don't think this new > driver belongs to the input subsystem), but I like the general idea of > having the "subclassing". Maybe we can make it prettier with Hans' > suggestion, given that this mainly means we are transforming > i2c-hid-core.c into a library. > > As for where this new goodix driver goes, it can stay in > drivers/hid/i2c-hid IMO.
Yep, I agree, it has nothing to do with input (except the device being physically a touchscreen ;) ), so driver/hid/i2c-hid makes most sense to me too. Thanks. -- Dmitry