[PATCH 1/2] dt-bindings: remoteproc: qcom,sm8550-pas: Make SDX75, Hawi and Maili MPSS fallback to SM8650

[PATCH 1/2] dt-bindings: remoteproc: qcom,sm8550-pas: Make SDX75, Hawi and Maili MPSS fallback to SM8650

[Mukesh Ojha]

SDX75, Hawi and Maili MPSS are compatible with the SM8650 MPSS as
they share the same set of interrupts, memory regions and power domains.
Rather than describing them as standalone compatibles, model them as
fallbacks to qcom,sm8650-mpss-pas.

Signed-off-by: Mukesh Ojha <mukesh.ojha@oss.qualcomm.com>
---
 .../bindings/remoteproc/qcom,sm8550-pas.yaml  | 29 +++++++++----------
 1 file changed, 14 insertions(+), 15 deletions(-)

diff --git a/Documentation/devicetree/bindings/remoteproc/qcom,sm8550-pas.yaml b/Documentation/devicetree/bindings/remoteproc/qcom,sm8550-pas.yaml
index a062188861ff..f722c55aadf4 100644
--- a/Documentation/devicetree/bindings/remoteproc/qcom,sm8550-pas.yaml
+++ b/Documentation/devicetree/bindings/remoteproc/qcom,sm8550-pas.yaml
@@ -18,9 +18,6 @@ properties:
     oneOf:
       - enum:
           - qcom,eliza-cdsp-pas
-          - qcom,hawi-mpss-pas
-          - qcom,maili-mpss-pas
-          - qcom,sdx75-mpss-pas
           - qcom,sm8550-adsp-pas
           - qcom,sm8550-cdsp-pas
           - qcom,sm8550-mpss-pas
@@ -49,6 +46,12 @@ properties:
       - items:
           - const: qcom,sm8750-cdsp-pas
           - const: qcom,sm8650-cdsp-pas
+      - items:
+          - enum:
+              - qcom,hawi-mpss-pas
+              - qcom,maili-mpss-pas
+              - qcom,sdx75-mpss-pas
+          - const: qcom,sm8650-mpss-pas
 
   reg:
     maxItems: 1
@@ -205,11 +208,9 @@ allOf:
   - if:
       properties:
         compatible:
-          enum:
-            - qcom,hawi-mpss-pas
-            - qcom,maili-mpss-pas
-            - qcom,sdx75-mpss-pas
-            - qcom,sm8650-mpss-pas
+          contains:
+            enum:
+              - qcom,sm8650-mpss-pas
     then:
       properties:
         interrupts:
@@ -257,13 +258,11 @@ allOf:
   - if:
       properties:
         compatible:
-          enum:
-            - qcom,hawi-mpss-pas
-            - qcom,maili-mpss-pas
-            - qcom,sdx75-mpss-pas
-            - qcom,sm8550-mpss-pas
-            - qcom,sm8650-mpss-pas
-            - qcom,sm8750-mpss-pas
+          contains:
+            enum:
+              - qcom,sm8550-mpss-pas
+              - qcom,sm8650-mpss-pas
+              - qcom,sm8750-mpss-pas
     then:
       properties:
         power-domains:
-- 
2.53.0

[PATCH 2/2] arm64: dts: qcom: sdx75: Use SM8650 MPSS as fallback compatible

[Mukesh Ojha]

Following the dt-bindings change that expresses the SDX75 MPSS PAS as
a SM8650 MPSS fallback, list the SM8650 compatible alongside the
SDX75 one so the node validates against the updated schema.

Signed-off-by: Mukesh Ojha <mukesh.ojha@oss.qualcomm.com>
---
 arch/arm64/boot/dts/qcom/sdx75.dtsi | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/arch/arm64/boot/dts/qcom/sdx75.dtsi b/arch/arm64/boot/dts/qcom/sdx75.dtsi
index 85cc9c6e6f83..dc9c4ed00d4b 100644
--- a/arch/arm64/boot/dts/qcom/sdx75.dtsi
+++ b/arch/arm64/boot/dts/qcom/sdx75.dtsi
@@ -925,7 +925,8 @@ tcsr: syscon@1fc0000 {
 		};
 
 		remoteproc_mpss: remoteproc@4080000 {
-			compatible = "qcom,sdx75-mpss-pas";
+			compatible = "qcom,sdx75-mpss-pas",
+				     "qcom,sm8650-mpss-pas";
 			reg = <0 0x04080000 0 0x10000>;
 
 			interrupts-extended = <&intc GIC_SPI 250 IRQ_TYPE_EDGE_RISING>,
-- 
2.53.0

Re: [PATCH 2/2] arm64: dts: qcom: sdx75: Use SM8650 MPSS as fallback compatible

[Dmitry Baryshkov]

On Mon, Aug 03, 2026 at 11:30:21PM +0530, Mukesh Ojha wrote:
> Following the dt-bindings change that expresses the SDX75 MPSS PAS as
> a SM8650 MPSS fallback, list the SM8650 compatible alongside the
> SDX75 one so the node validates against the updated schema.
> 
> Signed-off-by: Mukesh Ojha <mukesh.ojha@oss.qualcomm.com>
> ---
>  arch/arm64/boot/dts/qcom/sdx75.dtsi | 3 ++-
>  1 file changed, 2 insertions(+), 1 deletion(-)
> 

Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@oss.qualcomm.com>


-- 
With best wishes
Dmitry

Re: [PATCH 1/2] dt-bindings: remoteproc: qcom,sm8550-pas: Make SDX75, Hawi and Maili MPSS fallback to SM8650

[sashiko-bot]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Medium] Removing `qcom,sdx75-mpss-pas` from the standalone enum and enforcing a fallback breaks schema validation for existing DTBs and the in-tree DTS.
--

--- Patch [1]: [PATCH 1/2] dt-bindings: remoteproc: qcom,sm8550-pas: Make SDX75, Hawi and Maili MPSS fallback to SM8650 ---
commit 47a9ace0c12a078f309c8a844f59ffd712d4d63a
Author: Mukesh Ojha <mukesh.ojha@oss.qualcomm.com>

dt-bindings: remoteproc: qcom,sm8550-pas: Make SDX75, Hawi and Maili MPSS fallback to SM8650

SDX75, Hawi and Maili MPSS are compatible with the SM8650 MPSS as
they share the same set of interrupts, memory regions and power domains.
Rather than describing them as standalone compatibles, model them as
fallbacks to qcom,sm8650-mpss-pas.

> diff --git a/Documentation/devicetree/bindings/remoteproc/qcom,sm8550-pas.yaml b/Documentation/devicetree/bindings/remoteproc/qcom,sm8550-pas.yaml
> --- a/Documentation/devicetree/bindings/remoteproc/qcom,sm8550-pas.yaml
> +++ b/Documentation/devicetree/bindings/remoteproc/qcom,sm8550-pas.yaml
> @@ -18,9 +18,6 @@ properties:
>      oneOf:
>        - enum:
>            - qcom,eliza-cdsp-pas
> -          - qcom,hawi-mpss-pas
> -          - qcom,maili-mpss-pas
> -          - qcom,sdx75-mpss-pas
>            - qcom,sm8550-adsp-pas

[ ... ]

> @@ -49,6 +46,12 @@ properties:
>        - items:
>            - const: qcom,sm8750-cdsp-pas
>            - const: qcom,sm8650-cdsp-pas
> +      - items:
> +          - enum:
> +              - qcom,hawi-mpss-pas
> +              - qcom,maili-mpss-pas
> +              - qcom,sdx75-mpss-pas
> +          - const: qcom,sm8650-mpss-pas

[Severity: Medium]
Does removing the standalone compatible strings and enforcing them only as 
part of a 2-item array break backward compatibility for existing device trees?

If older device trees, such as arch/arm64/boot/dts/qcom/sdx75.dtsi, use 
qcom,sdx75-mpss-pas without the newly required fallback, will they start 
failing dtbs_check validation?

>  
>    reg:
>      maxItems: 1

[ ... ]

> @@ -205,11 +208,9 @@ allOf:
>    - if:
>        properties:
>          compatible:
> -          enum:
> -            - qcom,hawi-mpss-pas
> -            - qcom,maili-mpss-pas
> -            - qcom,sdx75-mpss-pas
> -            - qcom,sm8650-mpss-pas
> +          contains:
> +            enum:
> +              - qcom,sm8650-mpss-pas
>      then:
>        properties:
>          interrupts:

[Severity: Medium]
Since older device trees might not include the fallback string, does
removing the standalone compatible strings from these allOf conditional 
blocks mean those device trees will silently bypass validation for 
interrupts, power-domains, and memory-regions constraints?

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260803180021.20368-1-mukesh.ojha@oss.qualcomm.com?part=1

Re: [PATCH 1/2] dt-bindings: remoteproc: qcom,sm8550-pas: Make SDX75, Hawi and Maili MPSS fallback to SM8650

[Krzysztof Kozlowski]

On Mon, Aug 03, 2026 at 11:30:20PM +0530, Mukesh Ojha wrote:
> SDX75, Hawi and Maili MPSS are compatible with the SM8650 MPSS as
> they share the same set of interrupts, memory regions and power domains.
> Rather than describing them as standalone compatibles, model them as
> fallbacks to qcom,sm8650-mpss-pas.
> 
> Signed-off-by: Mukesh Ojha <mukesh.ojha@oss.qualcomm.com>
> ---
>  .../bindings/remoteproc/qcom,sm8550-pas.yaml  | 29 +++++++++----------
>  1 file changed, 14 insertions(+), 15 deletions(-)
> 
> diff --git a/Documentation/devicetree/bindings/remoteproc/qcom,sm8550-pas.yaml b/Documentation/devicetree/bindings/remoteproc/qcom,sm8550-pas.yaml
> index a062188861ff..f722c55aadf4 100644
> --- a/Documentation/devicetree/bindings/remoteproc/qcom,sm8550-pas.yaml
> +++ b/Documentation/devicetree/bindings/remoteproc/qcom,sm8550-pas.yaml
> @@ -18,9 +18,6 @@ properties:
>      oneOf:
>        - enum:
>            - qcom,eliza-cdsp-pas
> -          - qcom,hawi-mpss-pas

There is no known user of this.

Why do we even document this?

This is unfortunately one more example why I should accept early
bindings without users.

Best regards,
Krzysztof

Re: [PATCH 1/2] dt-bindings: remoteproc: qcom,sm8550-pas: Make SDX75, Hawi and Maili MPSS fallback to SM8650

[Bjorn Andersson]

On Tue, Aug 04, 2026 at 10:41:19AM +0200, Krzysztof Kozlowski wrote:
> On Mon, Aug 03, 2026 at 11:30:20PM +0530, Mukesh Ojha wrote:
> > SDX75, Hawi and Maili MPSS are compatible with the SM8650 MPSS as
> > they share the same set of interrupts, memory regions and power domains.
> > Rather than describing them as standalone compatibles, model them as
> > fallbacks to qcom,sm8650-mpss-pas.
> > 
> > Signed-off-by: Mukesh Ojha <mukesh.ojha@oss.qualcomm.com>
> > ---
> >  .../bindings/remoteproc/qcom,sm8550-pas.yaml  | 29 +++++++++----------
> >  1 file changed, 14 insertions(+), 15 deletions(-)
> > 
> > diff --git a/Documentation/devicetree/bindings/remoteproc/qcom,sm8550-pas.yaml b/Documentation/devicetree/bindings/remoteproc/qcom,sm8550-pas.yaml
> > index a062188861ff..f722c55aadf4 100644
> > --- a/Documentation/devicetree/bindings/remoteproc/qcom,sm8550-pas.yaml
> > +++ b/Documentation/devicetree/bindings/remoteproc/qcom,sm8550-pas.yaml
> > @@ -18,9 +18,6 @@ properties:
> >      oneOf:
> >        - enum:
> >            - qcom,eliza-cdsp-pas
> > -          - qcom,hawi-mpss-pas
> 
> There is no known user of this.
> 
> Why do we even document this?
> 
> This is unfortunately one more example why I should accept early
> bindings without users.
> 

Thanks for catching this, Krzysztof. As written, this clearly should
have been accompanied by a driver patch.

Regards,
Bjorn

> Best regards,
> Krzysztof
> 

Re: [PATCH 1/2] dt-bindings: remoteproc: qcom,sm8550-pas: Make SDX75, Hawi and Maili MPSS fallback to SM8650

[Mukesh Ojha]

On Tue, Aug 04, 2026 at 10:41:19AM +0200, Krzysztof Kozlowski wrote:
> On Mon, Aug 03, 2026 at 11:30:20PM +0530, Mukesh Ojha wrote:
> > SDX75, Hawi and Maili MPSS are compatible with the SM8650 MPSS as
> > they share the same set of interrupts, memory regions and power domains.
> > Rather than describing them as standalone compatibles, model them as
> > fallbacks to qcom,sm8650-mpss-pas.
> > 
> > Signed-off-by: Mukesh Ojha <mukesh.ojha@oss.qualcomm.com>
> > ---
> >  .../bindings/remoteproc/qcom,sm8550-pas.yaml  | 29 +++++++++----------
> >  1 file changed, 14 insertions(+), 15 deletions(-)
> > 
> > diff --git a/Documentation/devicetree/bindings/remoteproc/qcom,sm8550-pas.yaml b/Documentation/devicetree/bindings/remoteproc/qcom,sm8550-pas.yaml
> > index a062188861ff..f722c55aadf4 100644
> > --- a/Documentation/devicetree/bindings/remoteproc/qcom,sm8550-pas.yaml
> > +++ b/Documentation/devicetree/bindings/remoteproc/qcom,sm8550-pas.yaml
> > @@ -18,9 +18,6 @@ properties:
> >      oneOf:
> >        - enum:
> >            - qcom,eliza-cdsp-pas
> > -          - qcom,hawi-mpss-pas
> 
> There is no known user of this.
> 
> Why do we even document this?
> 
> This is unfortunately one more example why I should accept early
> bindings without users.

Yes, it was a miss and wanted to send revision of [1] after Yijie comments
I thought, I acked his comment somewhere later realized I acked his comment
internally.

[1]
https://lore.kernel.org/lkml/20260713093237.691117-1-mukesh.ojha@oss.qualcomm.com/

-- 
-Mukesh Ojha

Re: [PATCH 1/2] dt-bindings: remoteproc: qcom,sm8550-pas: Make SDX75, Hawi and Maili MPSS fallback to SM8650

[Bjorn Andersson]

On Mon, Aug 03, 2026 at 11:30:20PM +0530, Mukesh Ojha wrote:
> SDX75, Hawi and Maili MPSS are compatible with the SM8650 MPSS as
> they share the same set of interrupts, memory regions and power domains.
> Rather than describing them as standalone compatibles, model them as
> fallbacks to qcom,sm8650-mpss-pas.
> 

This will introduce a dt-validation error in
arch/arm64/boot/dts/qcom/sdx75.dtsi, which you fix in patch 2.

But while the binding says qcom,sdx75-mpss-pas alone isn't a valid
compatible, we support that in the driver, so we're maintaining
backwards compatibility with current DT - until someone cleans up the
driver to match the contract defined by the binding.

It might be reasonable to accept both of these hiccups, but you need to
document the upside of it.


But if I read this correctly, it seems I merged qcom,hawi-mpss-pas and
qcom,maili-mpss-pas prematurely. Based on the commit message in [1] I
thought they would fall back to sm8650, but this patch says that was
wrong of me - meaning that [1] should have contained a driver change as
well. I.e. you didn't test [1], right?

As I said in the reply to Krzysztof, I have dropped [1] again from my
queue.

[1] https://lore.kernel.org/all/20260713093237.691117-1-mukesh.ojha@oss.qualcomm.com/

Regards,
Bjorn

> Signed-off-by: Mukesh Ojha <mukesh.ojha@oss.qualcomm.com>
> ---
>  .../bindings/remoteproc/qcom,sm8550-pas.yaml  | 29 +++++++++----------
>  1 file changed, 14 insertions(+), 15 deletions(-)
> 
> diff --git a/Documentation/devicetree/bindings/remoteproc/qcom,sm8550-pas.yaml b/Documentation/devicetree/bindings/remoteproc/qcom,sm8550-pas.yaml
> index a062188861ff..f722c55aadf4 100644
> --- a/Documentation/devicetree/bindings/remoteproc/qcom,sm8550-pas.yaml
> +++ b/Documentation/devicetree/bindings/remoteproc/qcom,sm8550-pas.yaml
> @@ -18,9 +18,6 @@ properties:
>      oneOf:
>        - enum:
>            - qcom,eliza-cdsp-pas
> -          - qcom,hawi-mpss-pas
> -          - qcom,maili-mpss-pas
> -          - qcom,sdx75-mpss-pas
>            - qcom,sm8550-adsp-pas
>            - qcom,sm8550-cdsp-pas
>            - qcom,sm8550-mpss-pas
> @@ -49,6 +46,12 @@ properties:
>        - items:
>            - const: qcom,sm8750-cdsp-pas
>            - const: qcom,sm8650-cdsp-pas
> +      - items:
> +          - enum:
> +              - qcom,hawi-mpss-pas
> +              - qcom,maili-mpss-pas
> +              - qcom,sdx75-mpss-pas
> +          - const: qcom,sm8650-mpss-pas
>  
>    reg:
>      maxItems: 1
> @@ -205,11 +208,9 @@ allOf:
>    - if:
>        properties:
>          compatible:
> -          enum:
> -            - qcom,hawi-mpss-pas
> -            - qcom,maili-mpss-pas
> -            - qcom,sdx75-mpss-pas
> -            - qcom,sm8650-mpss-pas
> +          contains:
> +            enum:
> +              - qcom,sm8650-mpss-pas
>      then:
>        properties:
>          interrupts:
> @@ -257,13 +258,11 @@ allOf:
>    - if:
>        properties:
>          compatible:
> -          enum:
> -            - qcom,hawi-mpss-pas
> -            - qcom,maili-mpss-pas
> -            - qcom,sdx75-mpss-pas
> -            - qcom,sm8550-mpss-pas
> -            - qcom,sm8650-mpss-pas
> -            - qcom,sm8750-mpss-pas
> +          contains:
> +            enum:
> +              - qcom,sm8550-mpss-pas
> +              - qcom,sm8650-mpss-pas
> +              - qcom,sm8750-mpss-pas
>      then:
>        properties:
>          power-domains:
> -- 
> 2.53.0
> 

Re: [PATCH 1/2] dt-bindings: remoteproc: qcom,sm8550-pas: Make SDX75, Hawi and Maili MPSS fallback to SM8650

[Mukesh Ojha]

On Tue, Aug 04, 2026 at 11:11:29AM -0500, Bjorn Andersson wrote:
> On Mon, Aug 03, 2026 at 11:30:20PM +0530, Mukesh Ojha wrote:
> > SDX75, Hawi and Maili MPSS are compatible with the SM8650 MPSS as
> > they share the same set of interrupts, memory regions and power domains.
> > Rather than describing them as standalone compatibles, model them as
> > fallbacks to qcom,sm8650-mpss-pas.
> > 
> 
> This will introduce a dt-validation error in
> arch/arm64/boot/dts/qcom/sdx75.dtsi, which you fix in patch 2.

Right, but it will come at some point if we want to fix this.

> 
> But while the binding says qcom,sdx75-mpss-pas alone isn't a valid
> compatible, we support that in the driver, so we're maintaining
> backwards compatibility with current DT - until someone cleans up the
> driver to match the contract defined by the binding.
> 
> It might be reasonable to accept both of these hiccups, but you need to
> document the upside of it.
>

Thanks.

> 
> But if I read this correctly, it seems I merged qcom,hawi-mpss-pas and
> qcom,maili-mpss-pas prematurely. Based on the commit message in [1] I
> thought they would fall back to sm8650, but this patch says that was
> wrong of me - meaning that [1] should have contained a driver change as
> well. I.e. you didn't test [1], right?

Yes, it was wrong of me. I should have acknowledged Yijie's comments
publicly, saying it along with the sdx75 fix, and I did that internally with
him. I did test it against Hawi, but not for Maili, which is what Yijie
commented to take care of Maili along with Hawi.

> 
> As I said in the reply to Krzysztof, I have dropped [1] again from my
> queue.

Sure, will rebase this series accordingly with some documentation updates
as per your comment on this one.

> 
> [1] https://lore.kernel.org/all/20260713093237.691117-1-mukesh.ojha@oss.qualcomm.com/
>

-Mukesh