[PATCH 0/3] dt-bindings: Some schema fixes

[PATCH 0/3] dt-bindings: Some schema fixes

[Jean-Philippe Brucker]

The first patch fixes a syntax error that shows up on every use of
dt-validate.

The other two fix warnings that show up when trying to validate the QEMU
virt device-tree (with gic-version={2,3}). There are several other
warnings that need to be fixed in QEMU, but I believe these two are
schema issues.

Jean-Philippe Brucker (3):
  dt-bindings: regulator: Fix qcom,spmi-regulator schema
  dt-bindings: interrupt-controller: arm,gic-v3: Make 'interrupts'
    optional
  dt-bindings: interrupt-controller: arm,gic: Support two address and
    size cells

 .../devicetree/bindings/interrupt-controller/arm,gic-v3.yaml  | 1 -
 .../devicetree/bindings/interrupt-controller/arm,gic.yaml     | 4 ++--
 .../devicetree/bindings/regulator/qcom,spmi-regulator.yaml    | 3 +--
 3 files changed, 3 insertions(+), 5 deletions(-)

-- 
2.37.1

[PATCH 1/3] dt-bindings: regulator: Fix qcom,spmi-regulator schema

[Jean-Philippe Brucker]

The DT validator reports an error in the schema:

Documentation/devicetree/bindings/regulator/qcom,spmi-regulator.yaml: ignoring, error in schema: patternProperties: ^(5vs[1-2]|(l|s)[1-9][0-9]?|lvs[1-3])$: properties

Move the unevaluatedProperties statement out of the properties section
to fix it.

Fixes: 0b3bbd7646b0 ("regulator: qcom,spmi-regulator: Convert to dtschema")
Signed-off-by: Jean-Philippe Brucker <jean-philippe@linaro.org>
---
 .../devicetree/bindings/regulator/qcom,spmi-regulator.yaml     | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)

diff --git a/Documentation/devicetree/bindings/regulator/qcom,spmi-regulator.yaml b/Documentation/devicetree/bindings/regulator/qcom,spmi-regulator.yaml
index 8b7c4af4b551..faa4af9fd035 100644
--- a/Documentation/devicetree/bindings/regulator/qcom,spmi-regulator.yaml
+++ b/Documentation/devicetree/bindings/regulator/qcom,spmi-regulator.yaml
@@ -35,6 +35,7 @@ patternProperties:
     description: List of regulators and its properties
     type: object
     $ref: regulator.yaml#
+    unevaluatedProperties: false
 
     properties:
       qcom,ocp-max-retries:
@@ -100,8 +101,6 @@ patternProperties:
           SAW controlled gang leader. Will be configured as SAW regulator.
         type: boolean
 
-      unevaluatedProperties: false
-
 required:
   - compatible
 
-- 
2.37.1

Re: [PATCH 1/3] dt-bindings: regulator: Fix qcom,spmi-regulator schema

[Rob Herring]

On Mon, Aug 22, 2022 at 10:23 AM Jean-Philippe Brucker
<jean-philippe@linaro.org> wrote:
>
> The DT validator reports an error in the schema:
>
> Documentation/devicetree/bindings/regulator/qcom,spmi-regulator.yaml: ignoring, error in schema: patternProperties: ^(5vs[1-2]|(l|s)[1-9][0-9]?|lvs[1-3])$: properties
>
> Move the unevaluatedProperties statement out of the properties section
> to fix it.
>
> Fixes: 0b3bbd7646b0 ("regulator: qcom,spmi-regulator: Convert to dtschema")
> Signed-off-by: Jean-Philippe Brucker <jean-philippe@linaro.org>
> ---
>  .../devicetree/bindings/regulator/qcom,spmi-regulator.yaml     | 3 +--
>  1 file changed, 1 insertion(+), 2 deletions(-)

Reviewed-by: Rob Herring <robh@kernel.org>

Re: [PATCH 1/3] dt-bindings: regulator: Fix qcom,spmi-regulator schema

[Rob Herring]

On Mon, Aug 22, 2022 at 2:42 PM Rob Herring <robh+dt@kernel.org> wrote:
>
> On Mon, Aug 22, 2022 at 10:23 AM Jean-Philippe Brucker
> <jean-philippe@linaro.org> wrote:
> >
> > The DT validator reports an error in the schema:
> >
> > Documentation/devicetree/bindings/regulator/qcom,spmi-regulator.yaml: ignoring, error in schema: patternProperties: ^(5vs[1-2]|(l|s)[1-9][0-9]?|lvs[1-3])$: properties
> >
> > Move the unevaluatedProperties statement out of the properties section
> > to fix it.
> >
> > Fixes: 0b3bbd7646b0 ("regulator: qcom,spmi-regulator: Convert to dtschema")
> > Signed-off-by: Jean-Philippe Brucker <jean-philippe@linaro.org>
> > ---
> >  .../devicetree/bindings/regulator/qcom,spmi-regulator.yaml     | 3 +--
> >  1 file changed, 1 insertion(+), 2 deletions(-)
>
> Reviewed-by: Rob Herring <robh@kernel.org>

You sent this to the wrong Mar[kc]. Can you resend it and also Cc LKML.

Rob

[PATCH 2/3] dt-bindings: interrupt-controller: arm,gic-v3: Make 'interrupts' optional

[Jean-Philippe Brucker]

A GICv3 implementation without virtualization, such as the base QEMU
virt machine (without -M virtualization=on), does not issue maintenance
interrupts. Therefore its device-tree node does not need an 'interrupts'
property. Currently, validating the QEMU virt device-tree throws a
warning that 'interrupts' is missing. Make it optional.

Signed-off-by: Jean-Philippe Brucker <jean-philippe@linaro.org>
---
 .../devicetree/bindings/interrupt-controller/arm,gic-v3.yaml     | 1 -
 1 file changed, 1 deletion(-)

diff --git a/Documentation/devicetree/bindings/interrupt-controller/arm,gic-v3.yaml b/Documentation/devicetree/bindings/interrupt-controller/arm,gic-v3.yaml
index 3912a89162f0..9f7d3e11aacb 100644
--- a/Documentation/devicetree/bindings/interrupt-controller/arm,gic-v3.yaml
+++ b/Documentation/devicetree/bindings/interrupt-controller/arm,gic-v3.yaml
@@ -170,7 +170,6 @@ dependencies:
 
 required:
   - compatible
-  - interrupts
   - reg
 
 patternProperties:
-- 
2.37.1

Re: [PATCH 2/3] dt-bindings: interrupt-controller: arm,gic-v3: Make 'interrupts' optional

[Rob Herring]

On Mon, Aug 22, 2022 at 10:23 AM Jean-Philippe Brucker
<jean-philippe@linaro.org> wrote:
>
> A GICv3 implementation without virtualization, such as the base QEMU
> virt machine (without -M virtualization=on), does not issue maintenance
> interrupts. Therefore its device-tree node does not need an 'interrupts'
> property. Currently, validating the QEMU virt device-tree throws a
> warning that 'interrupts' is missing. Make it optional.
>
> Signed-off-by: Jean-Philippe Brucker <jean-philippe@linaro.org>
> ---
>  .../devicetree/bindings/interrupt-controller/arm,gic-v3.yaml     | 1 -
>  1 file changed, 1 deletion(-)

Reviewed-by: Rob Herring <robh@kernel.org>

Re: [PATCH 2/3] dt-bindings: interrupt-controller: arm,gic-v3: Make 'interrupts' optional

[Marc Zyngier]

On Mon, 22 Aug 2022 16:22:23 +0100,
Jean-Philippe Brucker <jean-philippe@linaro.org> wrote:
> 
> A GICv3 implementation without virtualization, such as the base QEMU
> virt machine (without -M virtualization=on), does not issue maintenance
> interrupts. Therefore its device-tree node does not need an 'interrupts'
> property. Currently, validating the QEMU virt device-tree throws a
> warning that 'interrupts' is missing. Make it optional.
> 
> Signed-off-by: Jean-Philippe Brucker <jean-philippe@linaro.org>

Acked-by: Marc Zyngier <maz@kernel.org>

	M.

-- 
Without deviation from the norm, progress is not possible.

Re: [PATCH 2/3] dt-bindings: interrupt-controller: arm,gic-v3: Make 'interrupts' optional

[Rob Herring]

On Mon, 22 Aug 2022 16:22:23 +0100, Jean-Philippe Brucker wrote:
> A GICv3 implementation without virtualization, such as the base QEMU
> virt machine (without -M virtualization=on), does not issue maintenance
> interrupts. Therefore its device-tree node does not need an 'interrupts'
> property. Currently, validating the QEMU virt device-tree throws a
> warning that 'interrupts' is missing. Make it optional.
> 
> Signed-off-by: Jean-Philippe Brucker <jean-philippe@linaro.org>
> ---
>  .../devicetree/bindings/interrupt-controller/arm,gic-v3.yaml     | 1 -
>  1 file changed, 1 deletion(-)
> 

Applied, thanks!

[PATCH 3/3] dt-bindings: interrupt-controller: arm,gic: Support two address and size cells

[Jean-Philippe Brucker]

It should be valid for a GICv2m node, child of a GICv2 node, to use two
cells per reg address and size. The QEMU virt device-tree currently
fails validation because the schema imposes a single address and size
cell. Amend the rule.

Signed-off-by: Jean-Philippe Brucker <jean-philippe@linaro.org>
---
 .../devicetree/bindings/interrupt-controller/arm,gic.yaml     | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/Documentation/devicetree/bindings/interrupt-controller/arm,gic.yaml b/Documentation/devicetree/bindings/interrupt-controller/arm,gic.yaml
index 62219a5c21c5..220256907461 100644
--- a/Documentation/devicetree/bindings/interrupt-controller/arm,gic.yaml
+++ b/Documentation/devicetree/bindings/interrupt-controller/arm,gic.yaml
@@ -64,9 +64,9 @@ properties:
   interrupt-controller: true
 
   "#address-cells":
-    enum: [ 0, 1 ]
+    enum: [ 0, 1, 2 ]
   "#size-cells":
-    const: 1
+    enum: [ 1, 2 ]
 
   "#interrupt-cells":
     const: 3
-- 
2.37.1

Re: [PATCH 3/3] dt-bindings: interrupt-controller: arm,gic: Support two address and size cells

[Marc Zyngier]

On Mon, 22 Aug 2022 16:22:24 +0100,
Jean-Philippe Brucker <jean-philippe@linaro.org> wrote:
> 
> It should be valid for a GICv2m node, child of a GICv2 node, to use two
> cells per reg address and size. The QEMU virt device-tree currently
> fails validation because the schema imposes a single address and size
> cell. Amend the rule.
> 
> Signed-off-by: Jean-Philippe Brucker <jean-philippe@linaro.org>

Acked-by: Marc Zyngier <maz@kernel.org>

	M.

-- 
Without deviation from the norm, progress is not possible.

Re: [PATCH 3/3] dt-bindings: interrupt-controller: arm,gic: Support two address and size cells

[Rob Herring]

On Mon, Aug 22, 2022 at 2:46 PM Marc Zyngier <maz@kernel.org> wrote:
>
> On Mon, 22 Aug 2022 16:22:24 +0100,
> Jean-Philippe Brucker <jean-philippe@linaro.org> wrote:
> >
> > It should be valid for a GICv2m node, child of a GICv2 node, to use two
> > cells per reg address and size. The QEMU virt device-tree currently
> > fails validation because the schema imposes a single address and size
> > cell. Amend the rule.
> >
> > Signed-off-by: Jean-Philippe Brucker <jean-philippe@linaro.org>
>
> Acked-by: Marc Zyngier <maz@kernel.org>

Did you want me to pick these up? If not:

Reviewed-by: Rob Herring <robh@kernel.org>

Re: [PATCH 3/3] dt-bindings: interrupt-controller: arm,gic: Support two address and size cells

[Marc Zyngier]

On Mon, 22 Aug 2022 20:50:16 +0100,
Rob Herring <robh+dt@kernel.org> wrote:
> 
> On Mon, Aug 22, 2022 at 2:46 PM Marc Zyngier <maz@kernel.org> wrote:
> >
> > On Mon, 22 Aug 2022 16:22:24 +0100,
> > Jean-Philippe Brucker <jean-philippe@linaro.org> wrote:
> > >
> > > It should be valid for a GICv2m node, child of a GICv2 node, to use two
> > > cells per reg address and size. The QEMU virt device-tree currently
> > > fails validation because the schema imposes a single address and size
> > > cell. Amend the rule.
> > >
> > > Signed-off-by: Jean-Philippe Brucker <jean-philippe@linaro.org>
> >
> > Acked-by: Marc Zyngier <maz@kernel.org>
> 
> Did you want me to pick these up? If not:
> 
> Reviewed-by: Rob Herring <robh@kernel.org>

Yes, please.

	M.

-- 
Without deviation from the norm, progress is not possible.

Re: [PATCH 3/3] dt-bindings: interrupt-controller: arm,gic: Support two address and size cells

[Rob Herring]

On Mon, 22 Aug 2022 16:22:24 +0100, Jean-Philippe Brucker wrote:
> It should be valid for a GICv2m node, child of a GICv2 node, to use two
> cells per reg address and size. The QEMU virt device-tree currently
> fails validation because the schema imposes a single address and size
> cell. Amend the rule.
> 
> Signed-off-by: Jean-Philippe Brucker <jean-philippe@linaro.org>
> ---
>  .../devicetree/bindings/interrupt-controller/arm,gic.yaml     | 4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
> 

Applied, thanks!