Hello Peter
I thought the logic behind the original driver was:
1) Create gpiochip so it can be used by other modules/userland
2) if there is also a irq available for it, create a irqdomain.
otherwise continue.
I can change the code so it does not continue, and exit cleanly if the
irqdomain fails.
You are the original developer of the driver. So: do you think it
makes sense to continue only as a gpiochip? :P
If I need to change it I rather add it as a another patch to the
patchset, so if it causes an issue somewhere it can be found easier
via bisect.
Regards!
On Mon, Jan 19, 2015 at 11:48 PM, Peter Korsgaard <peter@xxxxxxxxxxxxx> wrote:
>>>>>> "Ricardo" == Ricardo Ribalda Delgado <ricardo.ribalda@xxxxxxxxx> writes:
>
> > This way we do not need to transverse the device tree manually.
> > Cc: Linus Walleij <linus.walleij@xxxxxxxxxx>
> > Cc: Alexandre Courbot <gnurou@xxxxxxxxx>
> > Cc: Grant Likely <grant.likely@xxxxxxxxxx>
> > Cc: Rob Herring <robh+dt@xxxxxxxxxx>
> > Cc: Peter Korsgaard <jacmet@xxxxxxxxxx>
> > Cc: devicetree@xxxxxxxxxxxxxxx
> > Signed-off-by: Ricardo Ribalda Delgado <ricardo.ribalda@xxxxxxxxx>
> > ---
> > drivers/gpio/gpio-mpc8xxx.c | 48 +++++++++++++++++++++------------------------
> > 1 file changed, 22 insertions(+), 26 deletions(-)
>
> > diff --git a/drivers/gpio/gpio-mpc8xxx.c b/drivers/gpio/gpio-mpc8xxx.c
> > index d1ff879..57eb794 100644
> > --- a/drivers/gpio/gpio-mpc8xxx.c
> > +++ b/drivers/gpio/gpio-mpc8xxx.c
> > @@ -15,6 +15,7 @@
> > #include <linux/of.h>
> > #include <linux/of_gpio.h>
> > #include <linux/of_irq.h>
> > +#include <linux/of_platform.h>
> > #include <linux/gpio.h>
> > #include <linux/slab.h>
> > #include <linux/irq.h>
> > @@ -342,8 +343,9 @@ static struct of_device_id mpc8xxx_gpio_ids[] __initdata = {
> > {}
> > };
>
> > -static void __init mpc8xxx_add_controller(struct device_node *np)
> > +static int mpc8xxx_probe(struct platform_device *pdev)
> > {
> > + struct device_node *np = pdev->dev.of_node;
> > struct mpc8xxx_gpio_chip *mpc8xxx_gc;
> > struct of_mm_gpio_chip *mm_gc;
> > struct gpio_chip *gc;
> > @@ -351,11 +353,9 @@ static void __init mpc8xxx_add_controller(struct device_node *np)
> > unsigned hwirq;
> > int ret;
>
> > - mpc8xxx_gc = kzalloc(sizeof(*mpc8xxx_gc), GFP_KERNEL);
> > - if (!mpc8xxx_gc) {
> > - ret = -ENOMEM;
> > - goto err;
> > - }
> > + mpc8xxx_gc = devm_kzalloc(&pdev->dev, sizeof(*mpc8xxx_gc), GFP_KERNEL);
> > + if (!mpc8xxx_gc)
> > + return -ENOMEM;
>
> > spin_lock_init(&mpc8xxx_gc->lock);
>
> > @@ -375,16 +375,16 @@ static void __init mpc8xxx_add_controller(struct device_node *np)
>
> > ret = of_mm_gpiochip_add(np, mm_gc);
> > if (ret)
> > - goto err;
> > + return ret;
>
> > hwirq = irq_of_parse_and_map(np, 0);
> > if (hwirq == NO_IRQ)
> > - goto skip_irq;
> > + return 0;
>
> Actually, looking closer - Isn't this leaking mm_gc? It was as well
> before, but it would be good to get it fixed.
>
>
> > mpc8xxx_gc-> irq = irq_domain_add_linear(np, MPC8XXX_GPIO_PINS,
> > &mpc8xxx_gpio_irq_ops, mpc8xxx_gc);
> > if (!mpc8xxx_gc->irq)
> > - goto skip_irq;
> > + return 0;
>
> And here as well?
>
>
> --
> Bye, Peter Korsgaard
--
Ricardo Ribalda
--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
