On 11/06/24 14:24, Siddharth Vadapalli wrote:
> On Mon, Jun 10, 2024 at 10:31:07PM +0300, Roger Quadros wrote:
>
> [...]
>
>>> Based on your suggestion, you seem to propose the following hierarchy:
>>> k3-am62p-{main,mcu,thermal,wakeup}.dtsi = AM62P specific data
>>> k3-am62p.dtsi = k3-am62p-j722s-common-main.dtsi +
>>> k3-am62p-j722s-common-mcu.dtsi +
>>> k3-am62p-j722s-common-wakeup.dtsi +
>>> k3-am62p-j722s-common-thermal.dtsi +
>>> k3-am62p-{main,mcu,thermal,wakeup}.dtsi +
>>> <delta-5>
>>> k3-am62p5.dtsi = k3-am62p.dtsi + <delta-2>
>>> k3-j722s-{main,mcu,thermal,wakeup}.dtsi = J722S specific data
>>> k3-j722s.dtsi = k3-am62p-j722s-common-main.dtsi +
>>> k3-am62p-j722s-common-mcu.dtsi +
>>> k3-am62p-j722s-common-wakeup.dtsi +
>>> k3-am62p-j722s-common-thermal.dtsi +
>>> k3-j722s-{main,mcu,thermal,wakeup}.dtsi +
>>> <delta-6>
>>
>> What is the equivalent of k3-am62p5.dtsi here?
>> That should contain k3-j722s.dtsi + CPU and OPP stuff.
>>
>> I suppose it should be named specific to the SoC variant part number?
>
> AM62P (https://www.ti.com/product/AM62P) has two variants:
> 1. 2 Arm Cortex-A53 => AM62P3
> 2. 4 Arm Cortex-A53 => AM62P5
> Both variants will share the common k3-am62p.dtsi
>
> J722S (https://www.ti.com/product/TDA4VEN-Q1) has only one variant:
> 4 Arm Cortex-A53 => J722S
> Which is currently identical to AM62P5 w.r.t. the number of A53s.
>
> So there isn't an equivalent of AM62P5/k3-am62p5.dtsi for J722S.
> k3-j722s.dtsi is a combination of k3-am62p.dtsi and k3-am62p5.dtsi.
>
Historically AM6xx devices have had CPUs in separte file as there are OPNs with different number of CPU cores Hence, how about
k3-am62p5.dtsi => k3-am62p.dtsi + k3-am62p-j722s-common-{main,mcu,wakeup}.dtsi + k3-am62p-main.dtsi (USB2 and other deltas specific to AM62P)
and since J722s has no variants with less than 4 cores (and along the lines of rest of J7xx devices):
k3-j722s.dtsi => k3-am62p-j722s-common-{main,mcu,wakeup}.dtsi + k3-j722s-main.dtsi (USB3, C7x and other deltas specific to J722s;
--
Regards
Vignesh
