mpc5xxx keyboard BUG found

mpc5xxx keyboard BUG found

[Igor Luri]

We have an MPC5200LiteB board with a Microsoft USB keyboard atached and 
we have realized that some keys don't work properly, like up, down, lef 
and right arrows, PageUp, PageDown etc.

We have found that when those keys are pressed, first it sends a special 
scancode 0xe0 and then pressed key scancode.

Modifiying mpc5xxx_translate function to ignore the first scancode
 

                 // special prefix scancodes..
                  if (scancode == 0xe0 || scancode == 0xe1) {
                        return 0;
                  } 

our keyboard works well. Other cpus ignorates this special scancode or  
it evaluates.

Now, mpc5xxx_translate looks like this:

int mpc5xxx_translate(unsigned char scancode, unsigned char *keycode,
               char raw_mode)
{
    
  // special prefix scancodes..
  if (scancode == 0xe0 || scancode == 0xe1) {
        return 0;
  }  

  //0xFF is sent by a few keyboards, ignore it. 0x00 is error
  if (scancode == 0x00 || scancode == 0xff) {
    return 0;
  }
  *keycode = scancode;
  return 1;
}

Re: mpc5xxx keyboard BUG found

[Jarno Manninen]

On Friday 17 March 2006 13:38, Igor Luri wrote:

> We have an MPC5200LiteB board with a Microsoft USB keyboard atached and
> we have realized that some keys don't work properly, like up, down, lef
> and right arrows, PageUp, PageDown etc.

Hi!

Actually that Microsoft keyboard is working fine. A rare exception. ,-)
For more information see chapter 1.3 from here:

http://www.win.tue.nl/~aeb/linux/kbd/scancodes.html

So basically you will have to handle those in the application level code.

- Jarno

RE: mpc5xxx keyboard BUG found

[IGOR LURI]

The problem is that from aplication can't handle 0xe0 scancode because =
mpc5xxx_translate function translates the scancode to a keycode, and to =
the aplication arrives for example 0x0a1b5b41 (up arrow) and inmediatly =
another 0x0a. For this efect I think mpc5xxx_translate function needs to =
evaluate 0xe0 scancode correctly.


-----Mensaje original-----
De: Jarno Manninen [mailto:jarno.manninen@tut.fi]
Enviado el: viernes, 17 de marzo de 2006 19:50
Para: linuxppc-embedded@ozlabs.org
CC: IGOR LURI
Asunto: Re: mpc5xxx keyboard BUG found


On Friday 17 March 2006 13:38, Igor Luri wrote:

> We have an MPC5200LiteB board with a Microsoft USB keyboard atached =
and
> we have realized that some keys don't work properly, like up, down, =
lef
> and right arrows, PageUp, PageDown etc.

Hi!

Actually that Microsoft keyboard is working fine. A rare exception. ,-)
For more information see chapter 1.3 from here:

http://www.win.tue.nl/~aeb/linux/kbd/scancodes.html

So basically you will have to handle those in the application level =
code.

- Jarno

RE: mpc5xxx keyboard BUG found

[Jarno Manninen]

Lainaus IGOR LURI <iluri@fagorautomation.es>:

> The problem is that from aplication can't handle 0xe0 scancode 
> because mpc5xxx_translate function translates the scancode to a 
> keycode, and to the aplication arrives for example 0x0a1b5b41 (up 
> arrow) and inmediatly another 0x0a. For this efect I think 
> mpc5xxx_translate function needs to evaluate 0xe0 scancode correctly.

Hi!

I propably should have read your mail with more attention.

I'm not sure about keyboard handling on 2.6, but on 2.4 I just used 
translation function from pckeyb.c and it worked just fine.

- Jarno

> -----Mensaje original-----
> De: Jarno Manninen [mailto:jarno.manninen@tut.fi]
> Enviado el: viernes, 17 de marzo de 2006 19:50
> Para: linuxppc-embedded@ozlabs.org
> CC: IGOR LURI
> Asunto: Re: mpc5xxx keyboard BUG found
>
>
> On Friday 17 March 2006 13:38, Igor Luri wrote:
>
>> We have an MPC5200LiteB board with a Microsoft USB keyboard atached and
>> we have realized that some keys don't work properly, like up, down, lef
>> and right arrows, PageUp, PageDown etc.
>
> Hi!
>
> Actually that Microsoft keyboard is working fine. A rare exception. ,-)
> For more information see chapter 1.3 from here:
>
> http://www.win.tue.nl/~aeb/linux/kbd/scancodes.html
>
> So basically you will have to handle those in the application level code.
>
> - Jarno
>