On 17/11/2026 14:26, Luca Leonardo Scorcia wrote: >> Turns out you bring here undocumented ABI. NAK. > > Hi, I'm sorry but I don't see the relation between this commit you > just NAKed and the bigger issue about mediatek-drm alias usage. It's > not like I'm introducing anything new in these DPI specific patches > and the individual display block drivers do not depend on the aliases > (they're used by mediatek-drm for building the pipeline).
To clarify: you do introduce a new compatible here and in the DTS. Here it is documented. This is correct. However your DTS introduces also new usage of the ABI - aliases - and that part is not documented. New DTS cannot use undocumented ABI. This means that the binding was not complete. There is a mismatch or missing pieces between the binding and DTS. > > I understand why the commit to the dtsi that includes the aliases is > NAKed, and the eventual fix is going to require a change in all of > Mediatek SoCs dtsi. I am fine with the dtsi changes being blocked > until an appropriate solution is found. I just don't understand the > NAKs on the DPI components. Because this binding in relation to your DTS is incomplete. Of course if you drop the aliases from your DTS and this will be your final and working solution (bindings and DTS should be complete) then this NAK is not relevant and can be ignored. But in current form your bindings are incomplete and DTS after correcting won't work... Best regards, Krzysztof
