Hello Krzysztof, On Tue May 7, 2024 at 5:34 PM CEST, Krzysztof Kozlowski wrote: > On 07/05/2024 17:07, Théo Lebrun wrote: > > Proposal from Stephen Boyd of using auxiliary devices makes sense, that > > could be the future direction of this series. It won't change the > > dt-bindings aspect of it, only the driver implementations. > > > > [0]: https://lore.kernel.org/lkml/daa732cb31d947c308513b535930c729.sboyd@xxxxxxxxxx/ > > [1]: https://lore.kernel.org/lkml/20240124151405.GA930997-robh@xxxxxxxxxx/ > > So after Robs comment above, you still pushed the wrong approach and now > you revert it? Yes. The gist of it is that I had misunderstood the messages. Mostly, I did not understand how to implement separate Linux driver with the desired devicetree structure (no subnode on the syscon for each feature). I was missing knowledge about Linux infrastructure allowing for that. MFD and auxdevs are two approaches, with auxdevs being preferred. The latest revision finally takes those comments into account. Thanks Krzysztof, -- Théo Lebrun, Bootlin Embedded Linux and Kernel engineering https://bootlin.com
