Re: [PATCH] dt-bindings: pinctrl: samsung: Require clocks for Exynos9610 pinctrl nodes

[Alexandru Chimac]

Hello!
> On 09/23/2026 5:58 PM EEST Krzysztof Kozlowski <krzk@xxxxxxxxxx> wrote:
> 
>  
> On 22/09/2026 14:15, Alexandru Chimac wrote:
> > All Exynos9610 pin controllers (except PINCTRL_TOP) have clocks assigned
> > to them, so force the nodes to have clocks assigned to them.
> > 
> > Signed-off-by: Alexandru Chimac <alex@xxxxxxxxx>
> > ---
> > Most Samsung Exynos9610 pin controllers (except TOP) have clocks
> > designated for them, so allow the device tree node to include clock
> 
> I don't understand above. It's contradictory to commit msg which says
> ALL. Here you say most.
> 

On Exynos9610, only TOP pinctrl doesn't have a clock (as far as I'm
aware; haven't found a clock for it), while the rest of the pin controller
blocks do.

> 
> Why two commit msgs?

The first message is the commit message, and the second one is the b4
cover message. I wasn't actually sure how b4 handles that.

> 
> > information for the driver.
> 
> that's ABI break without reason.
> 

AFAIK there are no (publicly available) users of this device, so this
change shouldn't break any currently available device trees.

The best (most correct imo) implementation would be having the clock
properties be optional, pin controllers for Exynos9610 don't need
to be controlled by Linux for suspend/restore (change is more for
information(?) correctness anyway), since the CMUs should have HWACG
enabled anyway.

Regards,
Alexandru Chimac <alex@xxxxxxxxx