On Tue, Sep 17, 2024 at 10:21:16AM +0200, Krzysztof Kozlowski wrote: > On 17/09/2024 09:45, Francesco Dolcini wrote: > > On Mon, Sep 16, 2024 at 06:10:55PM +0200, Krzysztof Kozlowski wrote: > >> On 10/09/2024 17:22, Emanuele Ghidoli wrote: > >>> From: Emanuele Ghidoli <emanuele.ghidoli@xxxxxxxxxxx> > >>> > >>> The device tree defines the touchscreen controller, but it cannot be > >>> enabled because it lacks a reference label. > >> > >> It can be. Just enable it... > > > > colibri-imx8x is a SoM, it's not a complete self-contained device. The > > display touch controller is part of the SoM, however it is kept disabled > > since you need an actual touchscreen to make any use of it. > > > > This label would be used where an actual touchscreen is defined, this can > > happen with a DT overlay, for example. > > DT overlay should be in the tree. > > > > >>> This commit adds a label to allow it to be referenced and enabled. > >> > >> You changed here nothing. For me this patch is churn and pointless. > >> You add the label when you need to use it. > > > > DT files from the Linux kernel GIT are used also outside the Linux kernel > > sources, see for example U-Boot OF_UPSTREAM [1], to me it's fair to add > > a label for an out-of-tree user, am I wrong? > > > For U-Boot or any other upstream project: yes. It's enough to point to > lore patch or lore link. For downstream projects: I don't care, just > churn. Downstream should upstream their stuff to be considered. In the specific case this is a SOM dtsi that can be included in an actual whole self-contained product DTS file. It's not just about DT overlay. In the Linux kernel DT files we do have labels all the time into SOC dtsi files to allow the actual board user to augment/edit the node, even when not used (yet). As a user of SOC dtsi file this is great to me, and the same reasoning just apply to a SOM dtsi. While I understand you do not care, IMO it has some valid use case and should be considered for merging. Francesco
