Re: [PATCH v4 1/4] dt-bindings: soc: ti: Document the ti,tda54-mcu-scm-conf
From: Krzysztof Kozlowski
Date: Mon Nov 16 2026 - 07:22:31 EST
On 16/11/2026 12:33, Moteen Shah wrote:
> Hey Krzysztof,
>
> On 13/11/26 13:26, Krzysztof Kozlowski wrote:
>> On Thu, Nov 12, 2026 at 02:40:24PM +0530, Moteen Shah wrote:
>>> + reg:
>>> + maxItems: 1
>>> +
>>> + ti,proxy-id:
>>> + $ref: /schemas/types.yaml#/definitions/uint32
>>> + minimum: 0
>>> + maximum: 7
>>> + description:
>>> + Proxy region (0-7) through which this node's clients access
>>> + claimable registers and their fault status registers.
>> Partially continuing discussion from v2, IOW, device cannot access
>> entire address space and you need to limit the address space to specific
>> window of it? At least this is how it looks in driver code.
>>
>> This looks like the MMIO space is wrong. Listen, if you claim that the
>> DTB represents the host (or VM) view on hardware and based on that you
>> add proxy-id property to restrict the MMIO, then why do you claim that
>> entire MMIO is correct for this host (or VM) in the first place?
>
>
> Thanks for the review, I got your point.
> One thing I wanted to clarify before sending a revision: a nonzero proxy
> sets a bit in its CLAIM register to claim a register but for proxy 0 it
> shouldn't be doing that as per the hardware. If we choose to use proxy
> 0's reg window we just have to check if any other proxy has claimed it
> previously or no. Are you fine with having a property in the binding
> something like "proxy-0" or alternatively "ti,claim-required", because
> the driver won't be able to makeout which proxy it has the view of.
Dedicated property for proxy0 is fine. Maybe not calling it proxy-0 but
express that this one is shared with others so first you need to check
if it was not claimed.
But OTOH, can't the driver make the claim if if no one has claimed it in
every case? IOW, why for proxy1 you take the claim unconditionally, but
for proxy0 you check first?
Best regards,
Krzysztof