Re: [Qemu-devel] [PATCH v2] PS/2 keyboard Scancode Set 3 support

[Kevin Wolf <kwolf@redhat.com>]

Am 14.02.2011 13:49, schrieb Roy Tam:
> 2011/2/14 Kevin Wolf <kwolf@redhat.com>:
>> Am 13.02.2011 11:07, schrieb Roy Tam:
>>> The following patch adds PS/2 keyboard Scancode Set 3 support.
>>>
>>> Sign-off-by: Roy Tam <roytam@gmail.com>
>>> --
>>> v2: checkpatch.pl style fixes
>>>
>>> diff --git a/hw/ps2.c b/hw/ps2.c
>>> index 762bb00..6bea0ef 100644
>>> --- a/hw/ps2.c
>>> +++ b/hw/ps2.c
>>> @@ -143,13 +143,85 @@ static void ps2_put_keycode(void *opaque, int keycode)
>>>  {
>>>      PS2KbdState *s = opaque;
>>>
>>> -    /* XXX: add support for scancode sets 1 and 3 */
>>> -    if (!s->translate && keycode < 0xe0 && s->scancode_set == 2)
>>> -      {
>>> +    /* XXX: add support for scancode sets 1 */
>>> +    if (!s->translate && keycode < 0xe0 && s->scancode_set > 1) {
>>>          if (keycode & 0x80)
>>>              ps2_queue(&s->common, 0xf0);
>>>          keycode = ps2_raw_keycode[keycode & 0x7f];
>>> -      }
>>> +        if (s->scancode_set == 3) {
>>> +            switch (keycode) {
>>> +            case 0x1:
>>> +                keycode = 0x47;
>>> +                break;
>>> +            case 0x3:
>>> +                keycode = 0x27;
>>> +                break;
>>> +            case 0x4:
>>> +                keycode = 0x17;
>>> +                break;
>>> +            case 0x5:
>>> +                keycode = 0x7;
>>> +                break;
>>> +            case 0x6:
>>> +                keycode = 0xf;
>>> +                break;
>>> +            case 0x7:
>>> +                keycode = 0x5e;
>>> +                break;
>>> +            case 0x9:
>>> +                keycode = 0x4f;
>>> +                break;
>>> +            case 0xa:
>>> +                keycode = 0x3f;
>>> +                break;
>>> +            case 0xb:
>>> +                keycode = 0x2f;
>>> +                break;
>>> +            case 0xc:
>>> +                keycode = 0x1f;
>>> +                break;
>>> +            case 0x11:
>>> +                keycode = 0x19;
>>> +                break;
>>> +            case 0x14:
>>> +                keycode = 0x11;
>>> +                break;
>>> +            case 0x58:
>>> +                keycode = 0x14;
>>> +                break;
>>> +            case 0x5d:
>>> +                keycode = 0x5c;
>>> +                break;
>>> +            case 0x76:
>>> +                keycode = 0x8;
>>> +                break;
>>> +            case 0x77:
>>> +                keycode = 0x76;
>>> +                break;
>>> +            case 0x78:
>>> +                keycode = 0x56;
>>> +                break;
>>> +            case 0x79:
>>> +                keycode = 0x7c;
>>> +                break;
>>> +            case 0x7b:
>>> +                keycode = 0x84;
>>> +                break;
>>> +            case 0x7c:
>>> +                keycode = 0x7e;
>>> +                break;
>>> +            case 0x7e:
>>> +                keycode = 0x5f;
>>> +                break;
>>> +            case 0x83:
>>> +                keycode = 0x37;
>>> +                break;
>>> +            case 0x84:
>>> +                keycode = 0x57;
>>> +                break;
>>> +            }
>>> +        }
>>> +    }
>>>      ps2_queue(&s->common, keycode);
>>>  }
>>
>> Wouldn't a second table like ps2_raw_keycode be better than a huge
>> switch block that translates from scancode set 2 to 3?
> 
> Yeah, but I hate fixing old coding style to newer one. But still, I
> will do it for this time. I just wonder why not all sources are
> converted to new coding style when new coding style was announced.

Not sure what you're referring to, how is this related to coding style?

Kevin