Re: [PATCH 04/21] dt-bindings: gpu: img: Allow dma-coherent

[Andrew Davis <afd@ti.com>]

On 11/6/24 12:30 PM, Conor Dooley wrote:
> On Wed, Nov 06, 2024 at 10:18:06AM +0000, Matt Coster wrote:
>> On 05/11/2024 18:06, Conor Dooley wrote:
>>> On Tue, Nov 05, 2024 at 03:58:10PM +0000, Matt Coster wrote:
>>>> This attribute will be required for the BXS-4-64 MC1 and will be enabled in
>>>> the DTS for the TI k3-j721s2 in a subsequent patch; add it now so
>>>> dtbs_check doesn't complain later.
>>>
>>> Sounds like the property should be made required for that integration.
>>
>> This is something I went back and forth on. Where is the line drawn
>> between things that should be enforced in bindings and things that only
>> ever need to be specified once, so should just be left to the dt itself
>> to be the source of truth?
>>
>> Having said that, I realise TI could spin a new SoC with a new dt but
>> use the same compatible string for the GPU;
> 
> No they can't. New SoC, new compatible.
> 

We don't need to make a new SoC for that, our bus infra allows us to
selectively include or exclude peripherals from coherency. The GPU
being `dma-coherent` is selectable at boot time.

Andrew

>> the "single" source of truth
>> then wouldn't be so single anymore. I guess by making this property
>> required for this compatible string, we're saying any use of it must
>> behave in exactly the same way, right?