Re: [PATCH v3 3/5] i2c: mux: pca954x: Add interrupt controller support

quoted

BTW, what does client->irq == 0 represent?

This one I'm fairly confident on. Initially I had "client->irq < 0"
Which caused problems with an i2c mux that doesn't have irq's defined in the dt.

This commit seems to confirm that.
commit dab472eb931b ("i2c / ACPI: Use 0 to indicate that device does not have interrupt assigned")

In the Linux Kernel, 0 means no irq so one can write:

	if (!irq) ...

There used to be NO_IRQ per arch but this is converted over, although it
takes a lot of time... If there is a HW IRQ0, it needs to be remapped.
But since we rely on irq_desc these days and not irq numbers, this is
easy.

At least this was the transition a few years ago, but I don't think
something essential has changed. Please enlighten me, if so.

Regards,

   Wolfram