| Message ID | 20230403200530.2103099-3-abel.vesa@linaro.org (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | Add dedicated Qcom ICE driver | expand |
On 03/04/2023 22:05, Abel Vesa wrote: > Starting with SM8550, the ICE will have its own devicetree node > so add the qcom,ice property to reference it. > > Signed-off-by: Abel Vesa <abel.vesa@linaro.org> > --- > > The v4 is here: > https://lore.kernel.org/all/20230327134734.3256974-4-abel.vesa@linaro.org/ > > Changes since v4: > * Added check for sm8550 compatible w.r.t. qcom,ice in order to enforce > it while making sure none of the other platforms are allowed to use it Why? Also, this does not solve my previous question still. Best regards, Krzysztof
On 23-04-04 07:41:55, Krzysztof Kozlowski wrote: > On 03/04/2023 22:05, Abel Vesa wrote: > > Starting with SM8550, the ICE will have its own devicetree node > > so add the qcom,ice property to reference it. > > > > Signed-off-by: Abel Vesa <abel.vesa@linaro.org> > > --- > > > > The v4 is here: > > https://lore.kernel.org/all/20230327134734.3256974-4-abel.vesa@linaro.org/ > > > > Changes since v4: > > * Added check for sm8550 compatible w.r.t. qcom,ice in order to enforce > > it while making sure none of the other platforms are allowed to use it > > Why? SM8550 will be the first platform to use the new DT bindings w.r.t ICE. > > Also, this does not solve my previous question still. Well, the clocks are not added for the a few platforms (which include SM8550). Same for 'ice' reg range.. So the only thing left is to enforce the qcom,ice property availability only for SM8550. I believe it solves the mutual exclusiveness of the "ice" reg range along with the clocks versus the qcom,ice property, by enforcing at compatible level. Is this not enough? > > Best regards, > Krzysztof >
On 04/04/2023 10:59, Abel Vesa wrote: > On 23-04-04 07:41:55, Krzysztof Kozlowski wrote: >> On 03/04/2023 22:05, Abel Vesa wrote: >>> Starting with SM8550, the ICE will have its own devicetree node >>> so add the qcom,ice property to reference it. >>> >>> Signed-off-by: Abel Vesa <abel.vesa@linaro.org> >>> --- >>> >>> The v4 is here: >>> https://lore.kernel.org/all/20230327134734.3256974-4-abel.vesa@linaro.org/ >>> >>> Changes since v4: >>> * Added check for sm8550 compatible w.r.t. qcom,ice in order to enforce >>> it while making sure none of the other platforms are allowed to use it >> >> Why? > > SM8550 will be the first platform to use the new DT bindings w.r.t ICE. This I understand, but why other platforms cannot use it? > >> >> Also, this does not solve my previous question still. > > Well, the clocks are not added for the a few platforms (which include > SM8550). Same for 'ice' reg range.. So the only thing left is to > enforce the qcom,ice property availability only for SM8550. I believe > it solves the mutual exclusiveness of the "ice" reg range along with the > clocks versus the qcom,ice property, by enforcing at compatible level. Ah, I think I understand. That would work except I don't understand why enforcing qcom,qce only for specific, new SoCs. Assuming it is a correct hardware representation, we want it for everyone, don't we? Best regards, Krzysztof
On 23-04-04 12:12:06, Krzysztof Kozlowski wrote: > On 04/04/2023 10:59, Abel Vesa wrote: > > On 23-04-04 07:41:55, Krzysztof Kozlowski wrote: > >> On 03/04/2023 22:05, Abel Vesa wrote: > >>> Starting with SM8550, the ICE will have its own devicetree node > >>> so add the qcom,ice property to reference it. > >>> > >>> Signed-off-by: Abel Vesa <abel.vesa@linaro.org> > >>> --- > >>> > >>> The v4 is here: > >>> https://lore.kernel.org/all/20230327134734.3256974-4-abel.vesa@linaro.org/ > >>> > >>> Changes since v4: > >>> * Added check for sm8550 compatible w.r.t. qcom,ice in order to enforce > >>> it while making sure none of the other platforms are allowed to use it > >> > >> Why? > > > > SM8550 will be the first platform to use the new DT bindings w.r.t ICE. > > This I understand, but why other platforms cannot use it? The platforms that do not have ICE support yet will be added in the same subschema along with SM8550 when the ICE DT node will be added in their dtsi. > > > > >> > >> Also, this does not solve my previous question still. > > > > Well, the clocks are not added for the a few platforms (which include > > SM8550). Same for 'ice' reg range.. So the only thing left is to > > enforce the qcom,ice property availability only for SM8550. I believe > > it solves the mutual exclusiveness of the "ice" reg range along with the > > clocks versus the qcom,ice property, by enforcing at compatible level. > > Ah, I think I understand. That would work except I don't understand why > enforcing qcom,qce only for specific, new SoCs. Assuming it is a correct > hardware representation, we want it for everyone, don't we? Yes, but they will be added to the subschema (qcom,ice one) when their their ICE support (ICE DT) will be added. This way, we keep the bindings check without failures (for now). > > Best regards, > Krzysztof >
On 04/04/2023 12:41, Abel Vesa wrote: >>>> >>>> Also, this does not solve my previous question still. >>> >>> Well, the clocks are not added for the a few platforms (which include >>> SM8550). Same for 'ice' reg range.. So the only thing left is to >>> enforce the qcom,ice property availability only for SM8550. I believe >>> it solves the mutual exclusiveness of the "ice" reg range along with the >>> clocks versus the qcom,ice property, by enforcing at compatible level. >> >> Ah, I think I understand. That would work except I don't understand why >> enforcing qcom,qce only for specific, new SoCs. Assuming it is a correct >> hardware representation, we want it for everyone, don't we? > > Yes, but they will be added to the subschema (qcom,ice one) when their > their ICE support (ICE DT) will be added. This way, we keep the bindings > check without failures (for now). I understand that then you will rework this if:then case, so I think it is just easier to make it correct from the first place. If there is qcom,qce, then reg is maxItems:1. Otherwise - maxItems can be 2. You achieve the same result, all DTS validate, without any need of further changes later. Best regards, Krzysztof
diff --git a/Documentation/devicetree/bindings/ufs/qcom,ufs.yaml b/Documentation/devicetree/bindings/ufs/qcom,ufs.yaml index c5a06c048389..874de31d2c41 100644 --- a/Documentation/devicetree/bindings/ufs/qcom,ufs.yaml +++ b/Documentation/devicetree/bindings/ufs/qcom,ufs.yaml @@ -70,6 +70,10 @@ properties: power-domains: maxItems: 1 + qcom,ice: + $ref: /schemas/types.yaml#/definitions/phandle + description: phandle to the Inline Crypto Engine node + reg: minItems: 1 maxItems: 2 @@ -187,6 +191,21 @@ allOf: # TODO: define clock bindings for qcom,msm8994-ufshc + - if: + properties: + compatible: + contains: + enum: + - qcom,sm8550-ufshc + then: + properties: + qcom,ice: + maxItems: 1 + else: + properties: + qcom,ice: false + + unevaluatedProperties: false examples:
Starting with SM8550, the ICE will have its own devicetree node so add the qcom,ice property to reference it. Signed-off-by: Abel Vesa <abel.vesa@linaro.org> --- The v4 is here: https://lore.kernel.org/all/20230327134734.3256974-4-abel.vesa@linaro.org/ Changes since v4: * Added check for sm8550 compatible w.r.t. qcom,ice in order to enforce it while making sure none of the other platforms are allowed to use it Changes since v3: * dropped the "and drop core clock" part from subject line Changes since v2: * dropped all changes except the qcom,ice property .../devicetree/bindings/ufs/qcom,ufs.yaml | 19 +++++++++++++++++++ 1 file changed, 19 insertions(+)