> On 14.07.2015, at 14:56, Stephen Warren <swarren@xxxxxxxxxxxxx> wrote: > > I don't care so much about the internal code details; anything there can > be trivially changed. But please do make sure the DT correctly > represents the HW by: > > * Having a separate DT node for each HW block (SPI controller, UART, and > any shared interrupt mux/demux/controller/...) That is what we have with the v3 patch: * spi1 * spi2 * syscon controlling the aux-enable register (only used to enable/disable the hw block) > * If e.g. the SPI controller driver is going to need to manipulate the > "shared interrupt mux/demux/controller/...", there should be a phandle > from the SPI controller node to the "shared interrupt > mux/demux/controller/..." node, rather than listing the shared registers > in each "client"'s reg property. That is not what is in the v3 release of the patch any longer. Each of the above dt-nodes has their own (non-overlapping) register ranges and interrupts are enabled/disabled in the respective ranges for spi1/spi2. Also the status register of spi1/2 contain the corresponding flags for the reason of an interrupt - so no need to access any of the shared aux registers for that - hence the use of IRQF_SHARED. -- To unsubscribe from this list: send the line "unsubscribe devicetree" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html