Re: [PATCH RFC v2 02/11] dt-bindings: gpu: Add PowerVR Series5 SGX GPUs

[Andrew Davis <afd@ti.com>]

On 1/9/24 5:32 AM, Krzysztof Kozlowski wrote:
> On 08/01/2024 19:32, Andrew Davis wrote:
>> The Imagination PowerVR Series5 "SGX" GPU is part of several SoCs from
>> multiple vendors. Describe how the SGX GPU is integrated in these SoC,
>> including register space and interrupts. Clocks, reset, and power domain
>> information is SoC specific.
>>
>> Signed-off-by: Andrew Davis <afd@ti.com>
>> ---
>>   .../bindings/gpu/img,powervr-sgx.yaml         | 124 ++++++++++++++++++
>>   MAINTAINERS                                   |   1 +
>>   2 files changed, 125 insertions(+)
>>   create mode 100644 Documentation/devicetree/bindings/gpu/img,powervr-sgx.yaml
>>
>> diff --git a/Documentation/devicetree/bindings/gpu/img,powervr-sgx.yaml b/Documentation/devicetree/bindings/gpu/img,powervr-sgx.yaml
>> new file mode 100644
>> index 0000000000000..bb821e1184de9
>> --- /dev/null
>> +++ b/Documentation/devicetree/bindings/gpu/img,powervr-sgx.yaml
>> @@ -0,0 +1,124 @@
>> +# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
>> +# Copyright (c) 2023 Imagination Technologies Ltd.
> 
> Your email has @TI domain, are you sure you attribute your copyrights to
> Imagination?
> 

The file started as a copy/paste from a IMG copyrighted file, even
though it is now almost completely re-written I've left their (c)
for good measure. I'll add an additional TI (c).

> ...
> 
>> +
>> +  reg:
>> +    maxItems: 1
>> +
>> +  interrupts:
>> +    maxItems: 1
>> +
>> +  clocks: true
> 
> Missing min/maxItems
> 

These are set in the allOf/if/then blocks below, seems
if I don't set them to at least something here then I get
a warning:

    'clock-names', 'clocks' do not match any of the regexes: 'pinctrl-[0-9]+'

even if I define them in the allOf block below. I don't
know what the min/max should be until I check the compatible
in the allOf block.

>> +
>> +  clock-names:
>> +    minItems: 1
>> +    items:
>> +      - const: core
>> +      - const: mem
>> +      - const: sys
>> +
>> +  power-domains:
>> +    maxItems: 1
>> +
>> +required:
>> +  - compatible
>> +  - reg
>> +  - interrupts
>> +
>> +additionalProperties: false
> 
> This goes after allOf: block.
> 

ACK

>> +
>> +allOf:
>> +  - if:
>> +      properties:
>> +        compatible:
>> +          contains:
>> +            const: ti,am6548-gpu
>> +    then:
>> +      required:
>> +        - power-domains
>> +    else:
>> +      properties:
>> +        power-domains: false
>> +  - if:
>> +      properties:
>> +        compatible:
>> +          contains:
>> +            enum:
>> +              - allwinner,sun6i-a31-gpu
>> +              - ingenic,jz4780-gpu
>> +    then:
>> +      allOf:
>> +        - if:
> 
> I don't understand why do you need to embed allOf inside another allOf.
> The upper (outer) if:then: looks entirely useless.
> 

It is so that both compatibles falls through to having
clock being required.

Logic in YAML always seems messy to me, here it is in pseudo C:

if (compatible == allwinner,sun6i-a31-gpu ||
     compatible == ingenic,jz4780-gpu) {
	if (compatible == allwinner,sun6i-a31-gpu)
		clocks: ...
	if (compatible == ingenic,jz4780-gpu)
		clocks: ...
	required:
		- clocks
		- clock-names
} else { /* disallow for all others */
	properties:
		clocks: false
		clock-names: false
}

Now if I had an "else if" that didn't force the indention to keep
growing I would have used that. (does one exist?) I also cannot
simply add the clock properties only for the two compats need
them for the reasons above and so must add them unconditionally
before then explicitly disable them in a catch-all else path.

Andrew

>> +            properties:
>> +              compatible:
>> +                contains:
>> +                  const: allwinner,sun6i-a31-gpu
>> +          then:
>> +            properties:
>> +              clocks:
>> +                minItems: 2
>> +                maxItems: 2
>> +              clock-names:
>> +                minItems: 2
>> +                maxItems: 2
> 
> 
> Best regards,
> Krzysztof
>