Re: ukbd patch advice

On Mon, Jul 16, 2007 at 06:54:07PM +1200, Sam Banks wrote:
> Hey all,
> 
> I have been having a problem with my particular usb keyboard
> (0x03 was always being written into the first element of
> ukbd_data->keycode basically). I have tracked it down to a
> problem somewhere in between the uhci chipset (same problem
> with other cuts of uhci chipset), the uhci and/or ukbd
> drivers and the keyboard.
> 
> A fix to the problem is to reorder members of the struct
> ukbd_data. Originally, the members are ordered as modifiers,
> reserved and keycode[6] (minus a bunch of #define's). If I
> change this order to reserved, modifiers and keycode[6], my
> keyboard starts to function as it should (minus lighting up
> the LED's but that's another email all together :)). With
> this reordering, it stops other usb keyboards which function
> with the original code from working.

Can you give a diff so people can see the different order.

> 
> I'm wanting submit a patch for this fix (as other people are
> experiencing the same problems) but I'm not sure how to do
> this. I was thinking along the lines of the usb quirks
> function but it appears outside of a function body, you
> cannot have the normal if() type statements, only the
> preprocessor #ifdef types. To me, having a kernel config
> option for a single keyboard on a single driver seems quite
> overkill.
> 
> Does anyone have any suggestions on what I should do or can
> anyone point me to some code that deals with a similar
> problem?
> 
> On a side note, is anyone able to shed any light into why
> they think the above fix works? I am drawing a bit of a
> blank to be honest. Is it possible that my fix is only
> masking the problem?


cheers,
Andrew