Re: [PATCH 0/2] change "client->irq >= 0" to "client->irq > 0"

[Lars-Peter Clausen]

On 06/03/2015 01:02 PM, Alexandre Belloni wrote:
> On 03/06/2015 at 00:34:11 +0300, Octavian Purdila wrote :
> > This fixes an issue introduces by commit dab472eb931b ("i2c / ACPI:
> > Use 0 to indicate that device does not have interrupt assigned") where
> > drivers will try to request IRQ 0 when no GpioInt is defined in ACPI.
> >
> > The same issue occurs when the device is instantiated via device tree
> > with no IRQ, or from the i2c sysfs interface, even before the patch
> > above.
> >
> > Linus, since the commit above was already merged in the GPIO tree,
> > should these fixes be merged also via the GPIO tree (with ACKs from
> > the others subsystem maintainers)?
>
> Side question, has it been considered that IRQ 0 is valid on some
> platform and that means i2c devices will not be able to be wired to that
> IRQ anymore? Though, I don't think there are any existing design that
> does so.

If IRQ 0 is valid, that's a bug. IRQ 0 has been an invalid IRQ number in the 
global IRQ namespace for a while now. Though architectures are allowed to 
have a valid IRQ 0, but it may only be used inside the architecture code 
itself and must not be used for IRQs that are potentially be used by drivers.

- Lars

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/