PCA GPIO interrupt triggers

PCA GPIO interrupt triggers

[Peter Rusko]

Hi,

I'm using a GPIO chip (PCA9554) for a matrix-keypad. The keypad driver 
itself needs interrupts for both rising and falling edges. The PCA 
driver (gpio-pca953x.c) doesn't support it, but I don't understand why.

I think, this patch should work:

diff --git a/drivers/gpio/gpio-pca953x.c b/drivers/gpio/gpio-pca953x.c
index 147df8a..be9aff4 100644
--- a/drivers/gpio/gpio-pca953x.c
+++ b/drivers/gpio/gpio-pca953x.c
@@ -356,12 +356,6 @@ static int pca953x_irq_set_type(struct irq_data *d, 
unsigned int type)
         uint16_t level = d->irq - chip->irq_base;
         uint16_t mask = 1 << level;

-       if (!(type & IRQ_TYPE_EDGE_BOTH)) {
-               dev_err(&chip->client->dev, "irq %d: unsupported type %d\n",
-                       d->irq, type);
-               return -EINVAL;
-       }
-
         if (type & IRQ_TYPE_EDGE_FALLING)
                 chip->irq_trig_fall |= mask;
         else

I think that the other parts of the code checks for rising and falling 
edges correctly. Am I missing something?

-- 
Ruskó Péter
Fejlesztőmérnök

Prolan Zrt. / Prolan Co.
Hungary 2011 Budakalász, Szentendrei út 1-3.
Tel./Phone: +36 20 954 3118
Fax: +36 26 540 420
E-mail: rusko.peter@prolan.hu
Web: www.prolan.hu
Timezone:CET

Re: PCA GPIO interrupt triggers

[Grant Likely]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain, Size: 1541 bytes --]

On Thu, 22 Mar 2012 16:58:09 +0100, Peter Rusko <rusko.peter@prolan.hu> wrote:
> Hi,
> 
> I'm using a GPIO chip (PCA9554) for a matrix-keypad. The keypad driver 
> itself needs interrupts for both rising and falling edges. The PCA 
> driver (gpio-pca953x.c) doesn't support it, but I don't understand why.
> 
> I think, this patch should work:

Can someone with hardware review this patch please?  David perhaps?

g.

> 
> diff --git a/drivers/gpio/gpio-pca953x.c b/drivers/gpio/gpio-pca953x.c
> index 147df8a..be9aff4 100644
> --- a/drivers/gpio/gpio-pca953x.c
> +++ b/drivers/gpio/gpio-pca953x.c
> @@ -356,12 +356,6 @@ static int pca953x_irq_set_type(struct irq_data *d, 
> unsigned int type)
>          uint16_t level = d->irq - chip->irq_base;
>          uint16_t mask = 1 << level;
> 
> -       if (!(type & IRQ_TYPE_EDGE_BOTH)) {
> -               dev_err(&chip->client->dev, "irq %d: unsupported type %d\n",
> -                       d->irq, type);
> -               return -EINVAL;
> -       }
> -
>          if (type & IRQ_TYPE_EDGE_FALLING)
>                  chip->irq_trig_fall |= mask;
>          else
> 
> I think that the other parts of the code checks for rising and falling 
> edges correctly. Am I missing something?
> 
> -- 
> Ruskó Péter
> Fejlesztőmérnök
> 
> Prolan Zrt. / Prolan Co.
> Hungary 2011 Budakalász, Szentendrei út 1-3.
> Tel./Phone: +36 20 954 3118
> Fax: +36 26 540 420
> E-mail: rusko.peter@prolan.hu
> Web: www.prolan.hu
> Timezone:CET

-- 
Grant Likely, B.Sc, P.Eng.
Secret Lab Technologies,Ltd.

Re: PCA GPIO interrupt triggers

[Wolfram Sang]

[-- Attachment #1: Type: text/plain, Size: 620 bytes --]


> I'm using a GPIO chip (PCA9554) for a matrix-keypad. The keypad
> driver itself needs interrupts for both rising and falling edges.
> The PCA driver (gpio-pca953x.c) doesn't support it, but I don't
> understand why.

? I fail to see why having EDGE_FALLING together with EDGE_RISING is not
supported. Are you sure you didn't use some other IRQ_TYPE here?

> I think, this patch should work:

Think? Can't you test it :)

Regards,

   Wolfram

-- 
Pengutronix e.K.                           | Wolfram Sang                |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 198 bytes --]

Re: PCA GPIO interrupt triggers, matrix-keypad

[Peter Rusko]

> ? I fail to see why having EDGE_FALLING together with EDGE_RISING is not
> supported. Are you sure you didn't use some other IRQ_TYPE here?
>
Yes, you're right, the problem is not there.

>> I think, this patch should work:
>
> Think? Can't you test it :)
I wish that could be that simple :)

I can test it with the matrix-keypad driver. But there's an inverter
(+delay) in the circuit (to prevent contact bounce) and I don't know how
it affects the signals. But in the end sometimes I don't get e.g. the
key-release event.

Another change in the code was with the IRQ request. I could only make
the matrix-keypad driver working with threaded interrupts:

diff --git a/drivers/input/keyboard/matrix_keypad.c 
b/drivers/input/keyboard/matrix_keypad.c
index e2ae657..f047502 100644
--- a/drivers/input/keyboard/matrix_keypad.c
+++ b/drivers/input/keyboard/matrix_keypad.c
@@ -48,7 +48,7 @@ struct matrix_keypad {
  static void __activate_col(const struct matrix_keypad_platform_data 
*pdata,
                            int col, bool on)
  {
-       bool level_on = !pdata->active_low;
+       bool level_on = (!pdata->active_low)^(!!pdata->invert_output);

         if (on) {
                 gpio_direction_output(pdata->col_gpios[col], level_on);
@@ -341,7 +341,8 @@ static int __devinit init_matrix_gpio(struct 
platform_device *pdev,
                 }
         } else {
                 for (i = 0; i < pdata->num_row_gpios; i++) {
-                       err = request_irq(gpio_to_irq(pdata->row_gpios[i]),
+                       err = 
request_threaded_irq(gpio_to_irq(pdata->row_gpios[i]),
+                                       NULL,
                                         matrix_keypad_interrupt,
                                         IRQF_TRIGGER_RISING |
                                         IRQF_TRIGGER_FALLING,

Is this okay, or is there another way?

Regards,
Peter

Re: PCA GPIO interrupt triggers, matrix-keypad

[Wolfram Sang]

[-- Attachment #1: Type: text/plain, Size: 1573 bytes --]

On Mon, Mar 26, 2012 at 12:27:28PM +0200, Peter Rusko wrote:

> >? I fail to see why having EDGE_FALLING together with EDGE_RISING is not
> >supported. Are you sure you didn't use some other IRQ_TYPE here?
> >
> Yes, you're right, the problem is not there.

It was the most likely answer looking at how IRQ_TYPE_EDGE_BOTH is
defined.

> 
> >>I think, this patch should work:
> >
> >Think? Can't you test it :)
> I wish that could be that simple :)
> 
> I can test it with the matrix-keypad driver. But there's an inverter
> (+delay) in the circuit (to prevent contact bounce) and I don't know how
> it affects the signals. But in the end sometimes I don't get e.g. the
> key-release event.

It would be great if you could find out, since it might result in
arbitrary patches otherwise which would be a waste of time to review.

> Another change in the code was with the IRQ request. I could only make
> the matrix-keypad driver working with threaded interrupts:

[...]

> Is this okay, or is there another way?

The change itself is basically correct, but you need to make a proper
patch out of it. Please read Documentation/SubmittingPatches (also
Documentation/development-process/ if you haven't already) how to
generate one. Look at other threaded irq users how to convert properly
(you are missing something). Please CC me to the generated patch.

Regards,

   Wolfram

-- 
Pengutronix e.K.                           | Wolfram Sang                |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 198 bytes --]