Hello Uwe, On 07.03.22 14:23, Uwe Kleine-König wrote: > I would have gone with the mapping tables and I'd consider > /aliases/barebox,mmc0 = ... more ugly. But I agree this to be > a bit subjective. Mapping table will be more effort IMO. At least the `barebox,` prefix change would be mechanical and less error-prone. > If the soc-code continues to depend on having the mmc > aliases as defined in stm32mp151.dtsi, there should be a comment > describing that IMHO. It's generic code that depends on the aliases for fixup. >>>> Additionally having any alias at all ensures fixed naming that's >>>> not dependent on probe order. >>> >>> Fine for me. And if the board doesn't define the aliases, you get random >>> ordering. >> >> I prefer sane defaults. > > I prefer sane defaults iff they can be easily adapted by board code. If > you consider in board.dts: > > mmc0 = &sdmmc2; > mmc1 = &sdmmc3; > > (for whatever reasons), you end up with mmc1 and mmc2 both pointing to > sdmmc3. Ugly. /delete-property/ is easily adaptable IMO, but yes, not particularly nice to look at. >> I think most people involved have boards (often with trivial board support) >> that are not upstream. I do think we should avoid breaking them for no good >> reason. > > I think these people should mainline their trivial board support if they > want to be immune to such breaking. One could hope. :-) Cheers, Ahmad -- Pengutronix e.K. | | Steuerwalder Str. 21 | http://www.pengutronix.de/ | 31137 Hildesheim, Germany | Phone: +49-5121-206917-0 | Amtsgericht Hildesheim, HRA 2686 | Fax: +49-5121-206917-5555 | _______________________________________________ barebox mailing list barebox@xxxxxxxxxxxxxxxxxxx http://lists.infradead.org/mailman/listinfo/barebox
