lkml.org 
[lkml]   [2012]   [Feb]   [10]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: [PATCH v4 4/4] gpio: pl061: enable interrupts with DT style binding
On Thu, Feb 09, 2012 at 06:17:44PM -0600, Rob Herring wrote:
> On 02/09/2012 05:58 PM, Shawn Guo wrote:
> > On Thu, Feb 09, 2012 at 04:03:10PM -0600, Rob Herring wrote:
> >>
> >> On 02/09/2012 02:04 PM, Shawn Guo wrote:
> >>> On Fri, Feb 03, 2012 at 04:35:12PM -0600, Rob Herring wrote:
> >>> ...
> >>>> @@ -126,18 +127,16 @@ static void pl061_set_value(struct gpio_chip *gc, unsigned offset, int value)
> >>>> static int pl061_to_irq(struct gpio_chip *gc, unsigned offset)
> >>>> {
> >>>> struct pl061_gpio *chip = container_of(gc, struct pl061_gpio, gc);
> >>>> -
> >>>> - if (chip->irq_base <= 0)
> >>>> - return -EINVAL;
> >>>> -
> >>>> - return chip->irq_base + offset;
> >>>> + if (!chip->irq_gc)
> >>>> + return -ENXIO;
> >>>> + return irq_find_mapping(chip->irq_gc->domain, offset);
> >>>
> >>> If I understand the driver correctly, it will add a linear domain for
> >>> dt case. Do you have code somewhere creating the mapping before this
> >>> irq_find_mapping gets called here? The reason I'm asking this is I
> >>> have to call irq_create_mapping rather than irq_find_mapping here to
> >>> get imx gpio driver working with linear domain, otherwise the
> >>> irq_find_mapping call will fail.
> >>>
> >>
> >> Right, the user has to call irq_of_parse_and_map (which calls
> >> irq_create_mapping ultimately). Interrupts are allocated on demand. The
> >> dts needs to declare the gpio controller as an interrupt-controller and
> >> the node using the gpio line needs to set its interrupt parent and
> >> interrupt connection
> >>
> > Yes, that's how dt users use irq. But since I'm trying to make the
> > imx gpio irq_domain as linear for both non-dt and dt users. Calling
> > irq_create_mapping here may make sense for me, since it will not
> > require all these non-dt users change the way they use gpio irq.
> >
>
> I wouldn't try to use linear for non-DT at this point.
>
Sure, I'm not convincing you to do that but just explaining what I'm
doing here.

--
Regards,
Shawn


\
 
 \ /
  Last update: 2012-02-10 17:41    [W:0.040 / U:2.768 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site