Section: Keyboard Support (8)
Updated: 8 Nov 1994
The setkeycodes command reads its arguments two at a time, each pair of arguments consisting of a scancode (given in hexadecimal) and a keycode (given in decimal). For each such pair, it tells the kernel keyboard driver to map the specified scancode to the specified keycode.
This command is useful only for people with slightly unusual keyboards, that have a few keys which produce scancodes that the kernel does not recognize.
The usual PC keyboard produces a series of scancodes for each key press and key release. (Scancodes are shown by showkey -s, see showkey?(1).) The kernel parses this stream of scancodes, and converts it to a stream of keycodes (key press/release events). (Keycodes are shown by showkey.) Apart from a few scancodes with special meaning, and apart from the sequence produced by the Pause key, and apart from shiftstate related scancodes, and apart from the key up/down bit, the stream of scancodes consists of unescaped scancodes xx (7 bits) and escaped scancodes e0 xx (8+7 bits). To these scancodes or scancode pairs, a corresponding keycode can be assigned (in the range 1-127). For example, if you have a Macro key that produces e0 6f according to showkey?(1), the command
<DD CLASS="c1|setkeycodes e06f 112:
Some older kernels might hardwire a low scancode range to the equivalent keycodes; setkeycodes will fail when you try to remap these.
In 2.6 kernels key codes lie in the range 1-255, instead of 1-127. (It might be best to confine oneself to the range 1-239.)
dumpkeys (1), loadkeys (1), showkey (1), getkeycodes (8)