The mechanism behind keymap quirks and input handling

[Ozan Çağlayan]

Hi,

I'm having troubles understanding the mechanism behind the keymap quirks
and input handling. I read the keymap quirks tutorial written by Richard
Hughes multiple times but I still have a lot of questions.

* First of all how all that stuff is handled bottom-up from kernel
keyboard driver to hal and the userspace utilities?

* Do all of the keys on a keyboard should *absolutely* generate an event
or are there some exceptions that will never work?

* Say that I'm having "Unknown key pressed" messages on dmesg every time
I press the brightnessup key. Using dmidecode or lshal output, I can
find out my system properties and add a keymap quirk to the relevant
file to map the scancode X to keycode Y. But after mapping it to Y(say
brightnessup), should I settle a polling daemon for listening to that
event or *should* hald addons handle those events once the fdi matches
my DMI data?

* Say that I don't have any output on dmesg when I press my brightnessup
key. But if I launch xev, I can see that it correctly generates
XF86BrightnessUp events. Are there still anything to do with fdi files
and keymap quirks, etc. after that point? Even if it generates that
event, the button is still non-functional say at least on my
MacBookPro5,1 and on a brand new sony vaio tz* laptop. I know that
there's an addon called hald-addon-input which listens /dev/input*
interfaces. Every time I press that brightness key on my MacBookPro5,1,
I can see a line on hald --verbose=yes --daemon=no output that it
actually generates the event on the HAL part too. There are a lot of
other addons for brightness stuff, shouldn't that input addon pass the
information to those addons/scripts to adjust the brightness?

I'm really really confused about all that stuff, if anyone can provide
whitepapers, tutorials, faq's, about HAL and input event handling, I'd
be very glad.

I also didn't have an answer to another question that I've asked here
"Keymap quirks & input handling". The question there was tightly coupled
to the questions above.

Thanks,
Ozan Caglayan