On Mon, Oct 26, 2020 at 4:30 PM Ahmad Fatoum <a.fatoum@xxxxxxxxxxxxxx> wrote: > > Hello Rob, > > On 10/26/20 3:36 PM, Rob Herring wrote: > > On Wed, Oct 21, 2020 at 12:28:55PM +0200, Ahmad Fatoum wrote: > >> The stm32mp1 TAMP (Tamper and backup registers) does tamper detection > >> and features 32 backup registers that, being in the RTC domain, may > >> survive even with Vdd switched off. > >> > >> This makes it suitable for use to communicate a reboot mode from OS > >> to bootloader via the syscon-reboot-mode binding. Add a "simple-mfd" > >> to support probing such a child node. The actual reboot mode > >> node could then be defined in a board.dts or fixed up by the bootloader. > > > > 'simple-mfd' implies there is no dependency on the parent node for the > > child (such as the regmap perhaps). Is that the case here? > > No, there's a dependency and the Linux driver does syscon_node_to_regmap > on the device tree node's parent but that's how the syscon-reboot-mode binding > is documented: > > The SYSCON mapped register is retrieved from the > parental dt-node plus the offset. So the SYSCON reboot-mode node > should be represented as a sub-node of a "syscon", "simple-mfd" node. > > How would you prefer this being done instead? Well, probably the syscon driver could just probe any children, but I'm not sure if that would break anyone. So I guess fine as-is. Reviewed-by: Rob Herring <robh@xxxxxxxxxx> Rob
