lkml.org 
[lkml]   [2019]   [Aug]   [21]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
SubjectRe: [PATCH v2 11/11] arm64: dts: add dts nodes for MT6779
From
Date
Hi Marc

> >>> + soc {
> >>> + #address-cells = <2>;
> >>> + #size-cells = <2>;
> >>> + compatible = "simple-bus";
> >>> + ranges;
> >>> +
> >>> + gic: interrupt-controller@0c000000 {
> >>> + compatible = "arm,gic-v3";
> >>> + #interrupt-cells = <3>;
> >>
> >> You also haven't described the CPU PMUs. Depending on how they are wired
> >> (SPIs or PPIs), you may have to change the interrupt-cells property to
> >> include a cell for the PPI partitioning.
> >>
> >
> > pmu nodes would be:
> >
> > pmu {
> > compatible = "arm,armv8-pmuv3";
> > interrupt-parent = <&gic>;
> > interrupts = <GIC_PPI 7 IRQ_TYPE_LEVEL_LOW>;
> > };
> >
> > dsu-pmu-0 {
> > compatible = "arm,dsu-pmu";
> > interrupts = <GIC_SPI 18 IRQ_TYPE_LEVEL_HIGH>;
> > cpus = <&cpu0>, <&cpu1>, <&cpu2>, <&cpu3>,
> > <&cpu4>, <&cpu5>, <&cpu6>, <&cpu7>;
> > };
> >
> > so I think interrupt-cells could be <3>, will add pmu nodes in v3.
>
> No, that's wrong, at least for the CPU pmu node.
>
> First, you need two of them (one for the A55s, one for the A75s).
> Then you need to partition the corresponding PPI so that they can be
> described as separate affinity sets.
> Finally, this implies that #interrupt-cells goes up to 4, and all the
> interrupts directly routed to the GIC must be updated.
>
> You should have something like this:
>
> &gic {
> ppi-partitions {
> cluster0: interrupt-partition-0 {
> affinity = <&cpu0 &cpu1 &cpu2
> &cpu3 &cpu4 &cpu5>;
> };
>
> cluster1: interrupt-partition-1 {
> affinity = <&cpu6 &cpu7>;
> };
> };
>
> pmu_a55 {
> compatible = "arm,cortex-a55-pmu", "arm,armv8-pmuv3";
> interrupts = <GIC_PPI 7 IRQ_TYPE_LEVEL_LOW &cluster0>;
> };
>
> pmu_a75 {
> compatible = "arm,cortex-a75-pmu", "arm,armv8-pmuv3";
> interrupts = <GIC_PPI 7 IRQ_TYPE_LEVEL_LOW &cluster1>;
> };
>
> Please see the rk3399 usage of the binding, as it is the canonical example.
>
> >

Got the idea. Will check rk3399 and fix our part. Thanks for reviewing.

\
 
 \ /
  Last update: 2019-08-22 02:46    [W:0.126 / U:4.972 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site