On Thursday, May 7, 2015 at 10:00:04 AM UTC+2, Maxime Ripard wrote:
>
> ...
> > any advice for further investigation? 
>
> There's more to it. 
>
>
thanx for the quick reply!
 

> I don't think that ever supposed to work that way, most notably 
> because it just wouldn't work with SoCs that multiple interrupt banks 
> (like the A31, A23, etc.) 
>
>
sorry, i'm slowly wading my way on this topic, so i'm not very competent
I  supposed this single bank interrupt setup, as it's simpler (and older: 
A10 A20..) case were already working, and someway it works! :-)

then the more general case, multiple banks and so on.. will follow adding 
over.
 

> I think it worked by accident, because it's somewhat compatible with 
> the default interrupt property parsing code, but that probably need 
> some additional development. 
>
> What's mainly missing would be an irq_xlate function in the pinctrl 
> driver that would parse the interrupt property with the same format 
> than we have for gpios (phandle excluded, of course), which would mean 
> something like that in your case. 
>
> interrupt-parent = <&pio>; 
> interrupts = <8 10 flags>; 
>
>
in this example, there was a "interrupt-extended" entry that's i suppose 
it's more or less similar..

http://linux-sunxi.org/External_interrupts  

is that a fake/stub entry? amazing..

That function would then need to lookup in our pin array that it's 
> indeed an interrupt, and would return the number of the interrupt 
> within the interrupt controller (which would be 22 in your case). 
>
> I can try to cook something during the weekend, and send you a patch 
> for you to test. 
>

that would be more than i expect.. surely i'll test!

thanx

Andrea
 

>
> Maxime 
>
> -- 
> Maxime Ripard, Free Electrons 
> Embedded Linux, Kernel and Android engineering 
> http://free-electrons.com 
>

-- 
You received this message because you are subscribed to the Google Groups 
"linux-sunxi" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to linux-sunxi+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Reply via email to