[PATCH v2] pinctrl: baytrail: lock IRQs when starting them

[PATCH v2] pinctrl: baytrail: lock IRQs when starting them

[Linus Walleij]

This uses the new API for tagging GPIO lines as in use by
IRQs. This enforces a few semantic checks on how the underlying
GPIO line is used.

Cc: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Cc: Mika Westerberg <mika.westerberg@linux.intel.com>
Cc: Mathias Nyman <mathias.nyman@linux.intel.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
---
ChangeLog v1->v2:
- Actually make the change compile, doh.
---
 drivers/pinctrl/pinctrl-baytrail.c | 22 ++++++++++++++++++++++
 1 file changed, 22 insertions(+)

diff --git a/drivers/pinctrl/pinctrl-baytrail.c b/drivers/pinctrl/pinctrl-baytrail.c
index 1174ea86b6e9..998dbb0c4c93 100644
--- a/drivers/pinctrl/pinctrl-baytrail.c
+++ b/drivers/pinctrl/pinctrl-baytrail.c
@@ -372,11 +372,33 @@ static void byt_irq_mask(struct irq_data *d)
 {
 }
 
+static unsigned int byt_irq_startup(struct irq_data *d)
+{
+	struct byt_gpio *vg = irq_data_get_irq_handler_data(d);
+
+	if (gpio_lock_as_irq(&vg->chip, irqd_to_hwirq(d)))
+		dev_err(vg->chip.dev,
+			"unable to lock HW IRQ %lu for IRQ\n",
+			irqd_to_hwirq(d));
+	byt_irq_unmask(d);
+	return 0;
+}
+
+static void byt_irq_shutdown(struct irq_data *d)
+{
+	struct byt_gpio *vg = irq_data_get_irq_handler_data(d);
+
+	byt_irq_mask(d);
+	gpio_unlock_as_irq(&vg->chip, irqd_to_hwirq(d));
+}
+
 static struct irq_chip byt_irqchip = {
 	.name = "BYT-GPIO",
 	.irq_mask = byt_irq_mask,
 	.irq_unmask = byt_irq_unmask,
 	.irq_set_type = byt_irq_type,
+	.irq_startup = byt_irq_startup,
+	.irq_shutdown = byt_irq_shutdown,
 };
 
 static void byt_gpio_irq_init_hw(struct byt_gpio *vg)
-- 
1.8.3.1

[PATCH v2] pinctrl: baytrail: lock IRQs when starting them

[Mathias Nyman]

On 12/03/2013 05:00 PM, Linus Walleij wrote:
> This uses the new API for tagging GPIO lines as in use by
> IRQs. This enforces a few semantic checks on how the underlying
> GPIO line is used.
>
> Cc: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
> Cc: Mika Westerberg <mika.westerberg@linux.intel.com>
> Cc: Mathias Nyman <mathias.nyman@linux.intel.com>
> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
> ---
> ChangeLog v1->v2:
> - Actually make the change compile, doh.
> ---
>   drivers/pinctrl/pinctrl-baytrail.c | 22 ++++++++++++++++++++++
>   1 file changed, 22 insertions(+)
>
> diff --git a/drivers/pinctrl/pinctrl-baytrail.c b/drivers/pinctrl/pinctrl-baytrail.c
> index 1174ea86b6e9..998dbb0c4c93 100644
> --- a/drivers/pinctrl/pinctrl-baytrail.c
> +++ b/drivers/pinctrl/pinctrl-baytrail.c
> @@ -372,11 +372,33 @@ static void byt_irq_mask(struct irq_data *d)
>   {
>   }
>
> +static unsigned int byt_irq_startup(struct irq_data *d)
> +{
> +	struct byt_gpio *vg = irq_data_get_irq_handler_data(d);

This should probably be:

struct byt_gpio *vg = irq_data_get_irq_chip_data(d);

As chip methods use irq_data->chip_data, not irq_data->handler_data
(Mika tested that it works on baytrail)

> +
> +	if (gpio_lock_as_irq(&vg->chip, irqd_to_hwirq(d)))
> +		dev_err(vg->chip.dev,
> +			"unable to lock HW IRQ %lu for IRQ\n",
> +			irqd_to_hwirq(d));
> +	byt_irq_unmask(d);
> +	return 0;
> +}
> +
> +static void byt_irq_shutdown(struct irq_data *d)
> +{
> +	struct byt_gpio *vg = irq_data_get_irq_handler_data(d);

Same here.

-Mathias

[PATCH v2] pinctrl: baytrail: lock IRQs when starting them

[Mika Westerberg]

On Tue, Dec 03, 2013 at 06:02:18PM +0200, Mathias Nyman wrote:
> On 12/03/2013 05:00 PM, Linus Walleij wrote:
> >This uses the new API for tagging GPIO lines as in use by
> >IRQs. This enforces a few semantic checks on how the underlying
> >GPIO line is used.
> >
> >Cc: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
> >Cc: Mika Westerberg <mika.westerberg@linux.intel.com>
> >Cc: Mathias Nyman <mathias.nyman@linux.intel.com>
> >Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
> >---
> >ChangeLog v1->v2:
> >- Actually make the change compile, doh.
> >---
> >  drivers/pinctrl/pinctrl-baytrail.c | 22 ++++++++++++++++++++++
> >  1 file changed, 22 insertions(+)
> >
> >diff --git a/drivers/pinctrl/pinctrl-baytrail.c b/drivers/pinctrl/pinctrl-baytrail.c
> >index 1174ea86b6e9..998dbb0c4c93 100644
> >--- a/drivers/pinctrl/pinctrl-baytrail.c
> >+++ b/drivers/pinctrl/pinctrl-baytrail.c
> >@@ -372,11 +372,33 @@ static void byt_irq_mask(struct irq_data *d)
> >  {
> >  }
> >
> >+static unsigned int byt_irq_startup(struct irq_data *d)
> >+{
> >+	struct byt_gpio *vg = irq_data_get_irq_handler_data(d);
> 
> This should probably be:
> 
> struct byt_gpio *vg = irq_data_get_irq_chip_data(d);
> 
> As chip methods use irq_data->chip_data, not irq_data->handler_data
> (Mika tested that it works on baytrail)

Yes, if it uses irq_data_get_irq_handler_data() I get NULL vg pointer and
the kernel crashes. Using irq_data_get_irq_chip_data() makes it work.

[PATCH v2] pinctrl: baytrail: lock IRQs when starting them

[Gerhard Sittig]

On Tue, Dec 03, 2013 at 16:00 +0100, Linus Walleij wrote:
> 
> --- a/drivers/pinctrl/pinctrl-baytrail.c
> +++ b/drivers/pinctrl/pinctrl-baytrail.c
> @@ -372,11 +372,33 @@ static void byt_irq_mask(struct irq_data *d)
>  {
>  }
>  
> +static unsigned int byt_irq_startup(struct irq_data *d)
> +{
> +	struct byt_gpio *vg = irq_data_get_irq_handler_data(d);
> +
> +	if (gpio_lock_as_irq(&vg->chip, irqd_to_hwirq(d)))
> +		dev_err(vg->chip.dev,
> +			"unable to lock HW IRQ %lu for IRQ\n",
> +			irqd_to_hwirq(d));
> +	byt_irq_unmask(d);
> +	return 0;
> +}

Just a thought:  If failure to lock is non-fatal, should the
message be a warning then?

I do agree that failure to lock the GPIO should be non-fatal, as
there was debate and still might be concerns about how strict
locking should be.


virtually yours
Gerhard Sittig
-- 
DENX Software Engineering GmbH,     MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich, Office: Kirchenstr. 5, D-82194 Groebenzell, Germany
Phone: +49-8142-66989-0 Fax: +49-8142-66989-80  Email: office at denx.de