Re: [Qemu-devel] [Patch 1/2][PXA27x] initial keypad support

[Armin <akuster@kama-aina.net>]

andrzej zaborowski wrote:
> Hi,
>
> On 12/12/2007, Armin <akuster@kama-aina.net> wrote:
>   
>> Here is an attempt to add PXA27x keypad support. It currently only
>> supports the matrix type interface. It still needs direct and
>> mulitswitch support added.
>>
>> Just wanted to get something out there for folks to pound on.
>>
>> Comment and feedback welcome.
>>     
>
> Here are some comments:
>
> * Several bits in KPC and KPREC should be reset on read. The IRQ
> should be reset when it becomes masked in KPC, and potentially
> asserted when it becomes unmasked.
>   
Done, I think I got the reset on read .
The MI and DI interrupt bits are set after a scan has completed and  
then the IRQ is asserted. I think I got it.
> * It's wasteful to go through all keys in the 8x8 matrix on every
> event. It can be avoided easily by mapping qemu keycode to matrix
> position, rather than position to keycode, for example the PalmT|E
> matrix keypad does that.
>   
Yeah, that was a bit overkill ; )
> * Please convert the indentation in the file to use spaces, before
> considering for inclusion.
>
>   
sure thing
> * The bit definitions and pxa2xx_keypad_s struct are local to the
> keypad module so they shouldn't be in the header file exposed to the
> whole world.
>   
Ok.
> * Please set a copyright owner.
>   
Ok.
> Regarding mainstone keypad patch:
>
> * Probably not worth to stick the keymap in a file of its own since it
> is not shared between different machines - it's an integral part of
> mainstone.
>   
Done

Thanks for the feedback.

New patches to follow.

-Armin