On Tue, Aug 07, 2012 at 06:03:34PM +0100, Mark Brown wrote:

On Tue, Aug 07, 2012 at 06:01:30PM +0100, Lee Jones wrote:

quoted

Okay, so I've just spent a small amount of time looking at this. I think
the best place for this would be in *_get_virq(), using the same logic that
selected a *_legacy or *_linear domain in the first place. The only thing 
the domain can test for is the 'type' of domain and the requested IRQ. This
is where the issue lies. If a hwirq to virq conversion is requested, but a
virq is passed (which happens in the non-DT case) a WARN() is triggered
because the irq passed is bigger than first_irq + size. I think *_get_virq()
should ensure that only a hwirq is passed to irq_create_mapping().

quoted

Let me know if you had other ideas.

I'd expect your driver to always pass a hwirq into _get_virq() here.

Okay, actually this isn't so easy. Currently we have:

During DT boot:
 - No platform data is passed, hence no IRQ base for AB8500 is either
 - No IRQ base means we register a Linear IRQ Domain
 - MFD sees there is no base and leaves the IRQ resource as a hwirq
 - AB8500 child devices use *_get_virq() to convert the hwirq to a virq

During non-DT boot:
 - Platform data is passed, which contains an IRQ base
 - If an IRQ base is requested we use it to register a Legacy IRQ Domain
 - MFD adds the IRQ base to the hwirq and registers it as a virq
 - AB8500 child devices use *_get_virq() to convert virq to virq - *ERROR*

I guess my suggestion falls-back to placing logic in *_get_virq() to only
call irq_create_mapping() when when !ab8500->irq_base.

-- 
Lee Jones
Linaro ST-Ericsson Landing Team Lead
Linaro.org │ Open source software for ARM SoCs
Follow Linaro: Facebook | Twitter | Blog
--
To unsubscribe from this list: send the line "unsubscribe linux-input" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html