Re: [PATCH 9/9] ARM: Layerscape: LS1046a: add PSCI node

[Ahmad Fatoum <a.fatoum@pengutronix.de>]

On 21.10.24 09:57, Sascha Hauer wrote:
> Hi Ahmad,
> 
> On Mon, Oct 21, 2024 at 09:33:23AM +0200, Ahmad Fatoum wrote:
>> Hello Sascha,
>>
>> On 21.10.24 09:21, Sascha Hauer wrote:
>>> Unlike other SoCs the LS1046a upstream dtsi files do not have a PSCI
>>> node. Add one for barebox and Linux to find the PSCI support.
>>>
>>> Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
>>
>> I am wondering how powering CPUs on and off worked before?
>> I see no enable-method = "spin-table" in the kernel DT for the SoC.
> 
> The PPA is PSCI compatible. The barebox PPA adapter called
> of_psci_fixup() to register a PSCI node.

Ah, I see.

>> Anyhow, barebox should probably fixup enable-method = "psci" into each
>> CPU in /cpus in the kernel DT.
> 
> This is already done in of_psci_fixup().
> 
> The PSCI node in the barebox dts is only needed to probe the psci-client
> driver which in turn creates/fixups the PSCI node for the kernel. We
> already know that we started the TF-A. I thought about using this
> knowledge to register the PSCI client driver without explicitly creating
> a device node for it, but on the other hand there isn't too much gain in
> doing so, so I just created a node for now.

Sounds perfectly sensible.

Thanks for the clarification,
Ahmad

> 
> Sascha
> 


-- 
Pengutronix e.K.                           |                             |
Steuerwalder Str. 21                       | http://www.pengutronix.de/  |
31137 Hildesheim, Germany                  | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |