Progress update:

Now the daemon gets the interrupt from keyboard chip, and reads input report from it over i2c...

debug dump of single keypress and release: (6sec long)

2014-01-14 22:22:39 :: TOH Interrupt reached interrupt handler routine.
2014-01-14 22:22:39 :: Input report: 0b 00 ee 00 00 9a 00 00 00 00 00
2014-01-14 22:22:45 :: TOH Interrupt reached interrupt handler routine.
2014-01-14 22:22:45 :: Input report: 0b 00 ee 00 00 00 00 00 00 00 00


Something strange still with multiple simultaneous keypresses.

Then some level of functionality in getting keypresses to the applications gained by using uinput device..

I can create virtual input device!
[root@localhost input]# cat /sys/devices/virtual/input/input36/name
tohkbd


Simple test sends keystrokes.
/dev/uinput opened succesfully
tohkbd created succesfully
sending test
send key 0 == 37
send key 1 == 23
send key 2 == 50
send key 3 == 50
send key 4 == 24
send key 5 == 57
send key 6 == 20
send key 7 == 18
send key 8 == 31
send key 9 == 20
send key 10 == 30
send key 11 == 30
tohkbd destroyed succesfully


And this appears to screen only after i close/destroy the device (so not key-by-key as expected). And also the onscreen keyboard needs to be active.