Re: [PATCH v2 6/6] gpiolib: let the character device free its IRQs first on chip removal

[Kent Gibson]

On Mon, Oct 19, 2026 at 11:40:43AM +0300, Andy Shevchenko wrote:
> On Fri, Oct 16, 2026 at 09:40:03PM +0800, Kent Gibson wrote:
> > On Fri, Oct 16, 2026 at 12:23:38PM +0300, Adrian Ambrozewicz wrote:
> 
> ...
> 
> > > +	/* Called with gdev->chip already NULL, so gpiod_to_irq() can't be used. */
> > > +	if (gc->to_irq)
> > > +		irq = gc->to_irq(gc, gpiod_hwgpio(desc));
> > > +	else
> > > +		irq = 0;
> > > +	if (irq <= 0)
> > > +		return;
> > 
> > Perhaps I'm channelling Andy here, but how about:
> 
> It was in the previous version and I suggested to move it to the above. Seems
> like we have an opposite comments. Can you provide a rationale behind your choice?
> 

Ha, so it was, though your suggestion was not quite the above - it
didn't allow for gc->to_irq() returning 0 and it set the irq to a
negative error code.

> > 	if (!gc->to_irq)
> >         return;
> > 	irq = gc->to_irq(gc, gpiod_hwgpio(desc));
> > 	if (irq <= 0)
> > 		return;
> 

This way is clearer given gc->to_irq() can return 0 - the <= condition
there is purely because of its possible return values.
Setting irq to either 0 or an negative error code, rather than just returning
on the spot, are both more confusing and pointless.

This is my preferred form, and I obviously expected that you would prefer the
shorter form as well, but apparently not.

Cheers,
Kent.