Agilex5 introduces changes in how reserved memory is mapped and accessed
compared to previous SoC generations. Agilex5 compatible allows stratix10-
FPGA manager driver to handle these changes.

Fallback is added for driver probe and init that rely on matching of table
and DT node.

Reviewed-by: Xu Yilun <yilun.xu@intel.com>
Signed-off-by: Khairul Anuar Romli <khairul.anuar.romli@altera.com>
---
Changes in v4:
	- Remove redundant "items - enum" as suggested in v3.
	- Simplify compatible property to use contains instead of oneOf.
	- Validate fallback and non-fallback DT. Also validate binding with
          dt_binding_check.
Changes in v3:
	- Add description for Agilex5 Device
	- Add and define fallback to "intel,agilex-soc-fpga-mgr"
	- Validate against Agilex and Stratix10 (non-fallback) and Agilex5
	  (fallback)
Changes in v2:
	- No changes in this patch
---
 .../bindings/fpga/intel,stratix10-soc-fpga-mgr.yaml       | 8 +++++---
 1 file changed, 5 insertions(+), 3 deletions(-)

diff --git a/Documentation/devicetree/bindings/fpga/intel,stratix10-soc-fpga-mgr.yaml b/Documentation/devicetree/bindings/fpga/intel,stratix10-soc-fpga-mgr.yaml
index 6e536d6b28a9..65c737a4734c 100644
--- a/Documentation/devicetree/bindings/fpga/intel,stratix10-soc-fpga-mgr.yaml
+++ b/Documentation/devicetree/bindings/fpga/intel,stratix10-soc-fpga-mgr.yaml
@@ -20,9 +20,11 @@ description:
 
 properties:
   compatible:
-    enum:
-      - intel,stratix10-soc-fpga-mgr
-      - intel,agilex-soc-fpga-mgr
+    contains:
+      enum:
+        - intel,stratix10-soc-fpga-mgr
+        - intel,agilex-soc-fpga-mgr
+        - intel,agilex5-soc-fpga-mgr
 
 required:
   - compatible
-- 
2.43.7

On Tue, Nov 18, 2025 at 09:11:42AM +0800, Khairul Anuar Romli wrote:
> Agilex5 introduces changes in how reserved memory is mapped and accessed
> compared to previous SoC generations. Agilex5 compatible allows stratix10-
> FPGA manager driver to handle these changes.
> 
> Fallback is added for driver probe and init that rely on matching of table
> and DT node.
> 
> Reviewed-by: Xu Yilun <yilun.xu@intel.com>
> Signed-off-by: Khairul Anuar Romli <khairul.anuar.romli@altera.com>
> ---
> Changes in v4:
> 	- Remove redundant "items - enum" as suggested in v3.
> 	- Simplify compatible property to use contains instead of oneOf.
> 	- Validate fallback and non-fallback DT. Also validate binding with
>           dt_binding_check.

What are you doing?? You've been told several times exactly what to do
and yet you keep conjuring up completely random different things.
Using contains instead of the oneOf construct that was being done before
is *not* a simplification, it is functionally different.

NAK. Go do what you were told to do.

pw-bot: changes-requested

> Changes in v3:
> 	- Add description for Agilex5 Device
> 	- Add and define fallback to "intel,agilex-soc-fpga-mgr"
> 	- Validate against Agilex and Stratix10 (non-fallback) and Agilex5
> 	  (fallback)
> Changes in v2:
> 	- No changes in this patch
> ---
>  .../bindings/fpga/intel,stratix10-soc-fpga-mgr.yaml       | 8 +++++---
>  1 file changed, 5 insertions(+), 3 deletions(-)
> 
> diff --git a/Documentation/devicetree/bindings/fpga/intel,stratix10-soc-fpga-mgr.yaml b/Documentation/devicetree/bindings/fpga/intel,stratix10-soc-fpga-mgr.yaml
> index 6e536d6b28a9..65c737a4734c 100644
> --- a/Documentation/devicetree/bindings/fpga/intel,stratix10-soc-fpga-mgr.yaml
> +++ b/Documentation/devicetree/bindings/fpga/intel,stratix10-soc-fpga-mgr.yaml
> @@ -20,9 +20,11 @@ description:
>  
>  properties:
>    compatible:
> -    enum:
> -      - intel,stratix10-soc-fpga-mgr
> -      - intel,agilex-soc-fpga-mgr
> +    contains:
> +      enum:
> +        - intel,stratix10-soc-fpga-mgr
> +        - intel,agilex-soc-fpga-mgr
> +        - intel,agilex5-soc-fpga-mgr
>  
>  required:
>    - compatible
> -- 
> 2.43.7
> 

On 18/11/2025 20:07, Conor Dooley wrote:
> On Tue, Nov 18, 2025 at 09:11:42AM +0800, Khairul Anuar Romli wrote:
>> Agilex5 introduces changes in how reserved memory is mapped and accessed
>> compared to previous SoC generations. Agilex5 compatible allows stratix10-
>> FPGA manager driver to handle these changes.
>>
>> Fallback is added for driver probe and init that rely on matching of table
>> and DT node.
>>
>> Reviewed-by: Xu Yilun <yilun.xu@intel.com>
>> Signed-off-by: Khairul Anuar Romli <khairul.anuar.romli@altera.com>
>> ---
>> Changes in v4:
>> 	- Remove redundant "items - enum" as suggested in v3.
>> 	- Simplify compatible property to use contains instead of oneOf.
>> 	- Validate fallback and non-fallback DT. Also validate binding with
>>           dt_binding_check.
> 
> What are you doing?? You've been told several times exactly what to do
> and yet you keep conjuring up completely random different things.
> Using contains instead of the oneOf construct that was being done before
> is *not* a simplification, it is functionally different.
> 
> NAK. Go do what you were told to do.

There is no single file with that syntax, so I really do not understand
why completely different syntax is reasonable for the author :/

Best regards,
Krzysztof

On Tue, Nov 18, 2025 at 08:28:46PM +0100, Krzysztof Kozlowski wrote:
> On 18/11/2025 20:07, Conor Dooley wrote:
> > On Tue, Nov 18, 2025 at 09:11:42AM +0800, Khairul Anuar Romli wrote:
> >> Agilex5 introduces changes in how reserved memory is mapped and accessed
> >> compared to previous SoC generations. Agilex5 compatible allows stratix10-
> >> FPGA manager driver to handle these changes.
> >>
> >> Fallback is added for driver probe and init that rely on matching of table
> >> and DT node.
> >>
> >> Reviewed-by: Xu Yilun <yilun.xu@intel.com>
> >> Signed-off-by: Khairul Anuar Romli <khairul.anuar.romli@altera.com>
> >> ---
> >> Changes in v4:
> >> 	- Remove redundant "items - enum" as suggested in v3.
> >> 	- Simplify compatible property to use contains instead of oneOf.
> >> 	- Validate fallback and non-fallback DT. Also validate binding with
> >>           dt_binding_check.
> > 
> > What are you doing?? You've been told several times exactly what to do
> > and yet you keep conjuring up completely random different things.
> > Using contains instead of the oneOf construct that was being done before
> > is *not* a simplification, it is functionally different.
> > 
> > NAK. Go do what you were told to do.
> 
> There is no single file with that syntax, so I really do not understand
> why completely different syntax is reasonable for the author :/

There are actually a handful of bindings that do it (14). Some are
probably wrong, others are the generic portions of snps etc IP schemas.
Run this if you wanna quickly see what's what:

rg --multiline "^properties:\n.*compatible:\n.*contains" --context=5

I'd say that about half of them should be enums or const.