| Message ID | 20230516184626.154892-2-afd@ti.com |
|---|---|
| State | Changes Requested, archived |
| Headers | show |
| Series | None | expand |
| Context | Check | Description |
|---|---|---|
| robh/checkpatch | success | |
| robh/patch-applied | success | |
| robh/dt-meta-schema | fail | build log |
On 17/05/23 00:16, Andrew Davis wrote: > This node's register space is not accessed by any other node, which > is the traditional use for the "syscon" hint. Unfortunately that's not the case across SoCs. Eg AM65x See TRM section Table 5-582. CTRLMMR_EPWM0_CTRL Register Field Descriptions TB_CLKEN is clubbed with SYNCIN_SEL and ePWM tripzone configuration signals which may require register to be shared with other drivers in future > It looks to have been > added here to make use of a Linux kernel helper syscon_node_to_regmap(). > The Linux driver now uses a more appropriate helper that does not > require the hint, so let's remove it from the binding. > > Signed-off-by: Andrew Davis <afd@ti.com> > --- > .../devicetree/bindings/clock/ti,am654-ehrpwm-tbclk.yaml | 5 ++--- > 1 file changed, 2 insertions(+), 3 deletions(-) > > diff --git a/Documentation/devicetree/bindings/clock/ti,am654-ehrpwm-tbclk.yaml b/Documentation/devicetree/bindings/clock/ti,am654-ehrpwm-tbclk.yaml > index 66765116aff5..64b8bce5962c 100644 > --- a/Documentation/devicetree/bindings/clock/ti,am654-ehrpwm-tbclk.yaml > +++ b/Documentation/devicetree/bindings/clock/ti,am654-ehrpwm-tbclk.yaml > @@ -16,7 +16,6 @@ properties: > - ti,am654-ehrpwm-tbclk > - ti,am64-epwm-tbclk > - ti,am62-epwm-tbclk > - - const: syscon > > "#clock-cells": > const: 1 > @@ -33,8 +32,8 @@ additionalProperties: false > > examples: > - | > - ehrpwm_tbclk: syscon@4140 { > - compatible = "ti,am654-ehrpwm-tbclk", "syscon"; > + ehrpwm_tbclk: clock@4140 { > + compatible = "ti,am654-ehrpwm-tbclk"; > reg = <0x4140 0x18>; > #clock-cells = <1>; > };
On Tue, 16 May 2023 13:46:26 -0500, Andrew Davis wrote: > This node's register space is not accessed by any other node, which > is the traditional use for the "syscon" hint. It looks to have been > added here to make use of a Linux kernel helper syscon_node_to_regmap(). > The Linux driver now uses a more appropriate helper that does not > require the hint, so let's remove it from the binding. > > Signed-off-by: Andrew Davis <afd@ti.com> > --- > .../devicetree/bindings/clock/ti,am654-ehrpwm-tbclk.yaml | 5 ++--- > 1 file changed, 2 insertions(+), 3 deletions(-) > My bot found errors running 'make DT_CHECKER_FLAGS=-m dt_binding_check' on your patch (DT_CHECKER_FLAGS is new in v5.13): yamllint warnings/errors: dtschema/dtc warnings/errors: /builds/robherring/dt-review-ci/linux/Documentation/devicetree/bindings/mfd/ti,j721e-system-controller.example.dtb: scm-conf@100000: clock-controller@4140:compatible: ['ti,am654-ehrpwm-tbclk', 'syscon'] is too long From schema: /builds/robherring/dt-review-ci/linux/Documentation/devicetree/bindings/mfd/ti,j721e-system-controller.yaml /builds/robherring/dt-review-ci/linux/Documentation/devicetree/bindings/mfd/ti,j721e-system-controller.example.dtb: clock-controller@4140: compatible: ['ti,am654-ehrpwm-tbclk', 'syscon'] is too long From schema: /builds/robherring/dt-review-ci/linux/Documentation/devicetree/bindings/clock/ti,am654-ehrpwm-tbclk.yaml See https://patchwork.ozlabs.org/patch/1782200 This check can fail if there are any dependencies. The base for a patch series is generally the most recent rc1. If you already ran 'make dt_binding_check' and didn't see the above error(s), then make sure 'yamllint' is installed and dt-schema is up to date: pip3 install dtschema --upgrade Please check and re-submit.
On 16/05/2023 20:46, Andrew Davis wrote: > This node's register space is not accessed by any other node, which > is the traditional use for the "syscon" hint. It looks to have been > added here to make use of a Linux kernel helper syscon_node_to_regmap(). > The Linux driver now uses a more appropriate helper that does not > require the hint, so let's remove it from the binding. > > Signed-off-by: Andrew Davis <afd@ti.com> > --- Acked-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org> Best regards, Krzysztof
On 5/16/23 11:36 PM, Vignesh Raghavendra wrote: > > > On 17/05/23 00:16, Andrew Davis wrote: >> This node's register space is not accessed by any other node, which >> is the traditional use for the "syscon" hint. > > Unfortunately that's not the case across SoCs. Eg AM65x See TRM section > Table 5-582. CTRLMMR_EPWM0_CTRL Register Field Descriptions > Not sure what version of the TRM you have, latest (Rev. E) has this register as Table 5-636.. but I found it and see your point here. > TB_CLKEN is clubbed with SYNCIN_SEL and ePWM tripzone configuration > signals which may require register to be shared with other drivers in future > This looks to only be a problem in AM65x, all later devices we have fixed the issue and now group the clock enable bits all together. Do we actually expect this to be an issue and have a user of these other bits? If so then we modeled this region wrong in AM65x DT, these registers are not "tbclk gate registers" any more then they are to the other functions they provide. These registers should be a syscon node and then each function within should be a child node. syscon@4140 { compatible = "ti,am654-epwm-crtl", "syscon"; reg = <0x4140 0x18>; ehrpwm_tbclk: clock { compatible = "ti,am654-ehrpwm-tbclk"; #clock-cells = <1>; }; pwm_mux: mux-controller { compatible = "mmio-mux"; #mux-control-cells = <1>; }; }; Something like that. That way we do not give preference to one device and have to have it give out shared registers. Either that or split the binding compatible, one for AM65x with syscon and one for all later device compatibles that do not share the register: compatible: oneOf: - items: - const: ti,am654-ehrpwm-tbclk - const: syscon - items: - enum: - ti,am64-epwm-tbclk - ti,am62-epwm-tbclk Would rather the first option. Andrew > >> It looks to have been >> added here to make use of a Linux kernel helper syscon_node_to_regmap(). >> The Linux driver now uses a more appropriate helper that does not >> require the hint, so let's remove it from the binding. >> >> Signed-off-by: Andrew Davis <afd@ti.com> >> --- >> .../devicetree/bindings/clock/ti,am654-ehrpwm-tbclk.yaml | 5 ++--- >> 1 file changed, 2 insertions(+), 3 deletions(-) >> >> diff --git a/Documentation/devicetree/bindings/clock/ti,am654-ehrpwm-tbclk.yaml b/Documentation/devicetree/bindings/clock/ti,am654-ehrpwm-tbclk.yaml >> index 66765116aff5..64b8bce5962c 100644 >> --- a/Documentation/devicetree/bindings/clock/ti,am654-ehrpwm-tbclk.yaml >> +++ b/Documentation/devicetree/bindings/clock/ti,am654-ehrpwm-tbclk.yaml >> @@ -16,7 +16,6 @@ properties: >> - ti,am654-ehrpwm-tbclk >> - ti,am64-epwm-tbclk >> - ti,am62-epwm-tbclk >> - - const: syscon >> >> "#clock-cells": >> const: 1 >> @@ -33,8 +32,8 @@ additionalProperties: false >> >> examples: >> - | >> - ehrpwm_tbclk: syscon@4140 { >> - compatible = "ti,am654-ehrpwm-tbclk", "syscon"; >> + ehrpwm_tbclk: clock@4140 { >> + compatible = "ti,am654-ehrpwm-tbclk"; >> reg = <0x4140 0x18>; >> #clock-cells = <1>; >> }; >
Quoting Andrew Davis (2023-05-16 11:46:26) > This node's register space is not accessed by any other node, which > is the traditional use for the "syscon" hint. It looks to have been > added here to make use of a Linux kernel helper syscon_node_to_regmap(). > The Linux driver now uses a more appropriate helper that does not > require the hint, so let's remove it from the binding. > > Signed-off-by: Andrew Davis <afd@ti.com> > --- Applied to clk-next
On Wed, May 17, 2023 at 1:53 AM Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org> wrote: > > On Tue, 16 May 2023 13:46:26 -0500, Andrew Davis wrote: > > This node's register space is not accessed by any other node, which > > is the traditional use for the "syscon" hint. It looks to have been > > added here to make use of a Linux kernel helper syscon_node_to_regmap(). > > The Linux driver now uses a more appropriate helper that does not > > require the hint, so let's remove it from the binding. > > > > Signed-off-by: Andrew Davis <afd@ti.com> > > --- > > .../devicetree/bindings/clock/ti,am654-ehrpwm-tbclk.yaml | 5 ++--- > > 1 file changed, 2 insertions(+), 3 deletions(-) > > > > My bot found errors running 'make DT_CHECKER_FLAGS=-m dt_binding_check' > on your patch (DT_CHECKER_FLAGS is new in v5.13): > > yamllint warnings/errors: > > dtschema/dtc warnings/errors: > /builds/robherring/dt-review-ci/linux/Documentation/devicetree/bindings/mfd/ti,j721e-system-controller.example.dtb: scm-conf@100000: clock-controller@4140:compatible: ['ti,am654-ehrpwm-tbclk', 'syscon'] is too long > From schema: /builds/robherring/dt-review-ci/linux/Documentation/devicetree/bindings/mfd/ti,j721e-system-controller.yaml > /builds/robherring/dt-review-ci/linux/Documentation/devicetree/bindings/mfd/ti,j721e-system-controller.example.dtb: clock-controller@4140: compatible: ['ti,am654-ehrpwm-tbclk', 'syscon'] is too long > From schema: /builds/robherring/dt-review-ci/linux/Documentation/devicetree/bindings/clock/ti,am654-ehrpwm-tbclk.yaml Now failing in linux-next. Rob
On 6/23/23 2:59 PM, Rob Herring wrote: > On Wed, May 17, 2023 at 1:53 AM Krzysztof Kozlowski > <krzysztof.kozlowski@linaro.org> wrote: >> >> On Tue, 16 May 2023 13:46:26 -0500, Andrew Davis wrote: >>> This node's register space is not accessed by any other node, which >>> is the traditional use for the "syscon" hint. It looks to have been >>> added here to make use of a Linux kernel helper syscon_node_to_regmap(). >>> The Linux driver now uses a more appropriate helper that does not >>> require the hint, so let's remove it from the binding. >>> >>> Signed-off-by: Andrew Davis <afd@ti.com> >>> --- >>> .../devicetree/bindings/clock/ti,am654-ehrpwm-tbclk.yaml | 5 ++--- >>> 1 file changed, 2 insertions(+), 3 deletions(-) >>> >> >> My bot found errors running 'make DT_CHECKER_FLAGS=-m dt_binding_check' >> on your patch (DT_CHECKER_FLAGS is new in v5.13): >> >> yamllint warnings/errors: >> >> dtschema/dtc warnings/errors: >> /builds/robherring/dt-review-ci/linux/Documentation/devicetree/bindings/mfd/ti,j721e-system-controller.example.dtb: scm-conf@100000: clock-controller@4140:compatible: ['ti,am654-ehrpwm-tbclk', 'syscon'] is too long >> From schema: /builds/robherring/dt-review-ci/linux/Documentation/devicetree/bindings/mfd/ti,j721e-system-controller.yaml >> /builds/robherring/dt-review-ci/linux/Documentation/devicetree/bindings/mfd/ti,j721e-system-controller.example.dtb: clock-controller@4140: compatible: ['ti,am654-ehrpwm-tbclk', 'syscon'] is too long >> From schema: /builds/robherring/dt-review-ci/linux/Documentation/devicetree/bindings/clock/ti,am654-ehrpwm-tbclk.yaml > > Now failing in linux-next. > Sent fix: https://lore.kernel.org/lkml/20230623201519.194269-1-afd@ti.com/T/#u Andrew
diff --git a/Documentation/devicetree/bindings/clock/ti,am654-ehrpwm-tbclk.yaml b/Documentation/devicetree/bindings/clock/ti,am654-ehrpwm-tbclk.yaml index 66765116aff5..64b8bce5962c 100644 --- a/Documentation/devicetree/bindings/clock/ti,am654-ehrpwm-tbclk.yaml +++ b/Documentation/devicetree/bindings/clock/ti,am654-ehrpwm-tbclk.yaml @@ -16,7 +16,6 @@ properties: - ti,am654-ehrpwm-tbclk - ti,am64-epwm-tbclk - ti,am62-epwm-tbclk - - const: syscon "#clock-cells": const: 1 @@ -33,8 +32,8 @@ additionalProperties: false examples: - | - ehrpwm_tbclk: syscon@4140 { - compatible = "ti,am654-ehrpwm-tbclk", "syscon"; + ehrpwm_tbclk: clock@4140 { + compatible = "ti,am654-ehrpwm-tbclk"; reg = <0x4140 0x18>; #clock-cells = <1>; };
This node's register space is not accessed by any other node, which is the traditional use for the "syscon" hint. It looks to have been added here to make use of a Linux kernel helper syscon_node_to_regmap(). The Linux driver now uses a more appropriate helper that does not require the hint, so let's remove it from the binding. Signed-off-by: Andrew Davis <afd@ti.com> --- .../devicetree/bindings/clock/ti,am654-ehrpwm-tbclk.yaml | 5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-)