On Wed Jun 17, 2026 at 12:05 PM CEST, Konrad Dybcio wrote: > On 5/22/26 4:46 PM, Luca Weiss wrote: >> Due to initial kodiak/sc7280 bringup being done for Chrome platforms, >> some Chrome-specific bits still remain in kodiak.dtsi, like the clocks >> and power-domains for the LPASS RX/TX/WSA/VA macros. >> >> Move them to sc7280-chrome-common.dtsi and put Elite (q6afecc) >> equivalents in its place. The qcs6490-audioreach.dtsi file can also drop >> deletion of power-domains properties then. >> >> This follows previous commits moving Chrome-specific configuration to >> the correct file, leaving kodiak.dtsi for Elite and >> qcs6490-audioreach.dtsi for AudioReach. >> >> No functional change intended. The clock-output-names property will now >> exist for both Chrome and AudioReach devices but this shouldn't have any >> relevant effect. And WSA macro clocks weren't added to Chrome because I >> don't believe this would've ever worked given it already referenced >> q6afecc and the nodes were originally added during AudioReach bringup. > > I think it's better to keep them, if only to make sure that the result > of dtx_diff isn't outside the expectation of a commit that claims to > only reshuffle data for these platforms
I documented this in the commit message for this purpose. But I'd argue not putting (likely incorrect) wsa_macro overrides there is the better thing to do. Probably somebody should've taken care of the Chrome platforms when adding the wsa nodes in the first place. > > It looks OK as-is for Elite and Audioreach platforms (for the record, I > checked rb3gen2 and FP5) Thanks! Regards Luca
