On 04/01/2025 20:20, Andreas Kemnade wrote: > Am Sat, 4 Jan 2025 19:45:35 +0200 > schrieb Roger Quadros <rogerq@xxxxxxxxxx>: > >> Hi Andreas, >> >> On 30/12/2024 21:55, Andreas Kemnade wrote: >>> Usually interrupts are overwritten in the board file to specify a >>> mux-dependent dedicated wakeup irq, so there is interrupts and >>> interrupts-extended property which is not allowed. That has generated a >>> lot of noise during dts changes if just a phandle involved has randomly >>> changed. >>> >>> Avoid that mess by specifying interrupts-extended in the dtsi file. >>> >>> Signed-off-by: Andreas Kemnade <andreas@xxxxxxxxxxxx> >>> Reported-by: Rob Herring <robh@xxxxxxxxxx> >>> Closes: https://lore.kernel.org/linux-omap/173558214240.2262575.18233884215338168789.robh@xxxxxxxxxx/ >>> Closes: https://lore.kernel.org/linux-omap/172784021601.525825.18405282128990798038.robh@xxxxxxxxxx/ >>> --- >>> arch/arm/boot/dts/ti/omap/omap4-l4.dtsi | 8 ++++---- >>> 1 file changed, 4 insertions(+), 4 deletions(-) >>> >>> diff --git a/arch/arm/boot/dts/ti/omap/omap4-l4.dtsi b/arch/arm/boot/dts/ti/omap/omap4-l4.dtsi >>> index 3fcef3080eae..150dd84c9e0f 100644 >>> --- a/arch/arm/boot/dts/ti/omap/omap4-l4.dtsi >>> +++ b/arch/arm/boot/dts/ti/omap/omap4-l4.dtsi >>> @@ -1414,7 +1414,7 @@ SYSC_OMAP2_SOFTRESET | >>> uart3: serial@0 { >>> compatible = "ti,omap4-uart"; >>> reg = <0x0 0x100>; >>> - interrupts = <GIC_SPI 74 IRQ_TYPE_LEVEL_HIGH>; >>> + interrupts-extended = <&wakeupgen GIC_SPI 74 IRQ_TYPE_LEVEL_HIGH>; >> >> At this point interrupts-extended is not applicable. >> > we have it this way also in omap3. I do not understand what is the > problem with it. Do you have a pointer where it is forbidden? > At least > Documentation/devicetree/bindings/interrupt-controller/interrupts.txt > says nothing againt using it in such cases. Pasting the quoted text here for reference. "The "interrupts-extended" property is a special form; useful when a node needs to reference multiple interrupt parents or a different interrupt parent than the inherited one." I understood both were false so said that it is not applicable here. > >> We could use >> /delete-property/ interrupts >> in the board files that needs multiple interrupt parents? >> > What is the advantage of using that more complex solution? I would then > prefer to have the same with omap3 and omap4. If we do anything about > interrupts in board file here, they will have multiple parents. I was suggesting from the point to avoid churn in SoC dtsi file. But I see your point now. I don't have any objections. -- cheers, -roger
