Re: [PATCH 1/4] drm/{i915,xe}/display: move irq calls to parent interface

[Jani Nikula <jani.nikula@intel.com>]

On Mon, 10 Nov 2025, Ville Syrjälä <ville.syrjala@linux.intel.com> wrote:
> On Mon, Nov 10, 2025 at 09:31:36PM +0200, Jani Nikula wrote:
>> @@ -146,7 +145,7 @@ void ilk_update_display_irq(struct intel_display *display,
>>  	new_val |= (~enabled_irq_mask & interrupt_mask);
>>  
>>  	if (new_val != display->irq.ilk_de_imr_mask &&
>> -	    !drm_WARN_ON(display->drm, !intel_irqs_enabled(dev_priv))) {
>> +	    !drm_WARN_ON(display->drm, !display->parent->irq->enabled(display->drm))) {
>
> Can't we keep intel_irqs_enabled()/etc. as wrapper functions instead of
> open coding all that pointer chasing everywhere?

I opted for a compromise in v2, keeping the static wrappers inside
intel_display_irq.c, the most common caller of the interface, and
pointer chasing everywhere else.

I'm still a bit undecided how to deal with this in general. Currently
intel_display_rpm.c wraps the calls for rpm. But should we have
dedicated file(s) for the wrappers, with uniform naming conventions for
both the files and the wrappers? And do they *all* need a wrapper, even
when there's just one caller?

BR,
Jani.


-- 
Jani Nikula, Intel