Re: [PATCH v2 4/9] dt-bindings: interrupt-controller: Add RISC-V incoming MSI controller

[Anup Patel]

On Sun, Feb 19, 2023 at 4:48 PM Vivian Wang <uwu@xxxxxxxxx> wrote:
>
> On 1/3/23 22:14, Anup Patel wrote:
> > We add DT bindings document for the RISC-V incoming MSI controller
> > (IMSIC) defined by the RISC-V advanced interrupt architecture (AIA)
> > specification.
> >
> > Signed-off-by: Anup Patel <apatel@xxxxxxxxxxxxxxxx>
> > ---
> >  .../interrupt-controller/riscv,imsics.yaml    | 168 ++++++++++++++++++
> >  1 file changed, 168 insertions(+)
> >  create mode 100644 Documentation/devicetree/bindings/interrupt-controller/riscv,imsics.yaml
> >
> > diff --git a/Documentation/devicetree/bindings/interrupt-controller/riscv,imsics.yaml b/Documentation/devicetree/bindings/interrupt-controller/riscv,imsics.yaml
> > new file mode 100644
> > index 000000000000..b9db03b6e95f
> > --- /dev/null
> > +++ b/Documentation/devicetree/bindings/interrupt-controller/riscv,imsics.yaml
> > <snip>
> > +
> > +  interrupts-extended:
> > +    minItems: 1
> > +    maxItems: 16384
> > +    description:
> > +      This property represents the set of CPUs (or HARTs) for which given
> > +      device tree node describes the IMSIC interrupt files. Each node pointed
> > +      to should be a riscv,cpu-intc node, which has a riscv node (i.e. RISC-V
> > +      HART) as parent.
> > +
>
> This property doesn't seem to describe guest external interrupts. Should
> we add a reference to e.g. <&cpuN_intc 12> to indicate that IMSIC can
> send a 'Supervisor guest external interrupt'? Or just an idea, maybe we
> can add an additional interrupt controller to the CPU nodes to handle
> SGEI: (Various properties omitted)
>
> cpu0: cpu@N {
>         compatible = "riscv";
>
>         cpu0_intc: interrupt-controller {
>                 compatible = "riscv,cpu-intc";
>
>                 cpu0_gei: interrupt-controller {
>                         /* intc for hart-local hgeie/hgeip */
>                         compatible = "riscv,..."; /* Something here */
>                         interrupt-parent = <&cpu0_intc>;
>                         interrupts = <12>; /* SGEI */
>                         interrupt-controller;
>                         #interrupt-cells = <1>;
>                 }
>         }
> }
>
> interrupt-controller@... {
>         compatible = "riscv,imsics";
>         interrupts-extended = <&cpu0_intc 11>, <&cpu0_gei 1>, <&cpu0_gei 2> /* ... */;
> }
>
> I feel that this would be more appropriate, since the guest external
> interrupts are defined in the privileged architecture specification and
> are not specific to AIA. Though please do suggest more appropriate ways
> to formulate it.

This is unnecessary because GEILEN can be detected by init time
writes to hgeie CSR. Please look at KVM RISC-V AIA implementation
for more details. We only need "riscv,guest-index-bits" DT property
for address space holes.

In fact, we have tested these DT bindings with a variety of NUMA
configurations containing different numbers of IMISC guest files
per-HART.

Regards,
Anup