On Tue, 2021-02-16 at 18:42 +0100, Christophe Leroy wrote: > I'd suggest also to find the good arguments to convince us that this > series has a real added value, not just "cisco use it in its kernels > so it is good".
Well, IIRC, this series was endorsed by the device tree maintainers as the preferred alternative to this: https://lore.kernel.org/linux-devicetree/1565020400-25679-1-git-send-email-dan...@gimpelevich.san-francisco.ca.us/T/#u The now-defunct patchwork.linux-mips.org link in that thread pointed to: https://lore.kernel.org/linux-mips/1510796793.16864.25.camel@chimera/T/#u When running modern kernels from ancient vendor bootloaders, it is sometimes necessary to pick and choose bits and pieces of the info they pass without taking it verbatim.
