OMAP5912 boot broken by "gpio/omap: don't create an IRQ mapping for every GPIO on DT"

[Javier Martinez Canillas]

On 29/07/2013, at 13:43, Linus Walleij <linus.walleij at linaro.org> wrote:

> On Mon, Jul 29, 2013 at 12:19 PM, Javier Martinez Canillas
> <javier.martinez at collabora.co.uk> wrote:
> 
>> I guess the safer approach is to just revert these since they are causing
>> a regression and what the patches aims to fix as been broken since the
>> initial OMAP migration to DT anyways.
> 
> I have already reverted these and pushed to linux-next.
> 

Thanks for doing that and sorry for all the inconvenience.

Too bad this regression was not found before the patches hit mainline and they even were on linux-next for some time but shit happens :-(

> Actually input-hogs was not such a good idea either, after
> meditating on this I realized that the information about what
> GPIOs are used as interrupts is a piece of information that
> already exist in the device tree, albeit a bit distributed.
> 
> So I'm drafting an RFC to indicate how I think this should
> be solved.
> 

Great, I hope we can finally have a good solution for this long standing issue.

> Yours,
> Linus Walleij

Best regards,
Javier