Conor Dooley wrote: > On Fri, Jul 26, 2024 at 08:29:39AM -0400, Emil Renner Berthing wrote: > > Conor Dooley wrote: > > > There is no rohm,dh2228fv on any of supported JH7110 boards - in fact > > > the dh2228fv almost certainly does not exist as it is not a valid Rohm > > > part number. Likely a typo by Maxime when adding the device originally, > > > and should have been bh2228fv, but these boards do not have a bh2228fv > > > either! Remove it from jh7110-common.dtsi - pretending to have a device > > > so that the spidev driver will be bound by Linux is not acceptable. > > > > This patch is correct, but as you mention the fake device was most likely added > > in order to use spidev from userspace with random devices added on the exposed > > pins. In case someone actually makes use of this wouldn't this be a regression? > > What is the right way to support this? > > Unfortunately, there's no "right way" that's supported for for this > particular case. If people want to use spidev for their device, they > should either document it in the bindings, add the compatible to the > spidev driver and use an overlay to add the device to the dts or they > can r bind the spidev driver to the device from userspace. > > The other thing, which doesn't exist yet, is a connector binding. The > folks are Beagle are currently working on creating a connector binding > for the Mikrobus connector - but that's rather far from complete at the > moment. > I see. Thanks for the explanation. At least now there is this thread any potential users might find. Reviewed-by: Emil Renner Berthing <emil.renner.berthing@xxxxxxxxxxxxx>
