matrix_keyboard based driver

matrix_keyboard based driver

[Christian Gmeiner]

Hi all,

today I started with a basic matrix_keyboard based keypad driver. It
seems to work quite
well but I have some questions.

Is it possible to define following values from user space?

* debounce time in [ms]
* auto repeat time in [ms]
* multi click abort time in [ms]

I am also not sure about the default keymap I should use in the
platform_data. At
the moment it looks like:

static const uint32_t ot200_keymap[] = {
	KEY(0, 1, 0x01),
	KEY(0, 2, 0x02),
	KEY(0, 3, 0x03),
	KEY(0, 4, 0x04),
	KEY(0, 5, 0x05),
	KEY(0, 6, 0x06),
	KEY(0, 7, 0x07),
	KEY(0, 8, 0x08),
	KEY(0, 9, 0x09),
...
...
	KEY(9, 6, 0x96),
	KEY(9, 7, 0x97),
	KEY(9, 8, 0x98),
	KEY(9, 9, 0x99),
};

I need to configure the mapping from userspace and it is be done via loadkeys
if I am not wrong. Does the input core handle cases like SHIFT key and
CAPS lock?

thanks
--
Christian Gmeiner, MSc

Re: matrix_keyboard based driver

[Dmitry Torokhov]

Hi CHristian,

On Mon, Nov 21, 2011 at 02:52:17PM +0100, Christian Gmeiner wrote:
> Hi all,
> 
> today I started with a basic matrix_keyboard based keypad driver. It
> seems to work quite
> well but I have some questions.
> 
> Is it possible to define following values from user space?
> 
> * debounce time in [ms]

No as it is mostly hardware property, not user preference.

> * auto repeat time in [ms]

Yes: EVIOCSREP (kbdrate or similar). Note that X ignores kernel
autorepeat and implements its own.

> * multi click abort time in [ms]

I have no idea what that is, so apparently no ;)

> 
> I am also not sure about the default keymap I should use in the
> platform_data. At
> the moment it looks like:
> 
> static const uint32_t ot200_keymap[] = {
> 	KEY(0, 1, 0x01),
> 	KEY(0, 2, 0x02),
> 	KEY(0, 3, 0x03),
> 	KEY(0, 4, 0x04),
> 	KEY(0, 5, 0x05),
> 	KEY(0, 6, 0x06),
> 	KEY(0, 7, 0x07),
> 	KEY(0, 8, 0x08),
> 	KEY(0, 9, 0x09),
> ...
> ...
> 	KEY(9, 6, 0x96),
> 	KEY(9, 7, 0x97),
> 	KEY(9, 8, 0x98),
> 	KEY(9, 9, 0x99),

The last items in triplets should be KEY_* constant defined in
include/linux/input.h

> };
> 
> I need to configure the mapping from userspace and it is be done via loadkeys
> if I am not wrong.

To adjust mapping of the input device itself one should use 'setkeycode'
or some other utility issuing EVIOCGKEYCODE/EVIOCSKEYCODE ioctls.

loadkeys adjusts higher level keymap used by the console.

> Does the input core handle cases like SHIFT key and
> CAPS lock?

No, that is the task for higher layers (console, X, etc...).

Thanks.

-- 
Dmitry

Re: matrix_keyboard based driver

[Christian Gmeiner]

Hi Dmitry,

2011/11/21 Dmitry Torokhov <dmitry.torokhov@gmail.com>:
> Hi CHristian,
>
> On Mon, Nov 21, 2011 at 02:52:17PM +0100, Christian Gmeiner wrote:
>> Hi all,
>>
>> today I started with a basic matrix_keyboard based keypad driver. It
>> seems to work quite
>> well but I have some questions.
>>
>> Is it possible to define following values from user space?
>>
>> * debounce time in [ms]
>
> No as it is mostly hardware property, not user preference.

Okay... lets see If i need to do this in hardware.

>
>> * auto repeat time in [ms]
>
> Yes: EVIOCSREP (kbdrate or similar). Note that X ignores kernel
> autorepeat and implements its own.

Fine.

>
>> * multi click abort time in [ms]
>
> I have no idea what that is, so apparently no ;)

At the moment I have a uinput based user space program,
which gets the matrix coordinates and send back the pressed
key. And there it is possible to enter a handy-like input method.

press 3 times in a short time the letter 1 on the keypad and the
uinput based program sends a 'c' to the kernel.


>
>>
>> I am also not sure about the default keymap I should use in the
>> platform_data. At
>> the moment it looks like:
>>
>> static const uint32_t ot200_keymap[] = {
>>       KEY(0, 1, 0x01),
>>       KEY(0, 2, 0x02),
>>       KEY(0, 3, 0x03),
>>       KEY(0, 4, 0x04),
>>       KEY(0, 5, 0x05),
>>       KEY(0, 6, 0x06),
>>       KEY(0, 7, 0x07),
>>       KEY(0, 8, 0x08),
>>       KEY(0, 9, 0x09),
>> ...
>> ...
>>       KEY(9, 6, 0x96),
>>       KEY(9, 7, 0x97),
>>       KEY(9, 8, 0x98),
>>       KEY(9, 9, 0x99),
>
> The last items in triplets should be KEY_* constant defined in
> include/linux/input.h
>

Okay.. this should be doable... my company offers a way to have their
own mapping of some universal keys on the keypad. for instance it is
possible to assign to a key the action to sending ALT + TAB.

>> };
>>
>> I need to configure the mapping from userspace and it is be done via loadkeys
>> if I am not wrong.
>
> To adjust mapping of the input device itself one should use 'setkeycode'
> or some other utility issuing EVIOCGKEYCODE/EVIOCSKEYCODE ioctls.
>

So these two ioctls are changing the last value in the triplets of the keymap?

>> static const uint32_t ot200_keymap[] = {
>>       KEY(0, 1, 0x01),

In this case 0x01 will get replaced?

> loadkeys adjusts higher level keymap used by the console.
>

ah okay.

>> Does the input core handle cases like SHIFT key and
>> CAPS lock?
>
> No, that is the task for higher layers (console, X, etc...).
>

Need to play a little bit with I think.

thanks
--
Christian Gmeiner, MSc
--
To unsubscribe from this list: send the line "unsubscribe linux-input" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html