>May be I was not explicit enough... The problem does not show up in 
>console mode, it is only when running the X server that Right-Shift on 
>the LK401 does not produce shift. Hence I think the way Xcfbpmax thinks 
>of its keyboard needs to be changed...


The kernel keyboard-handling code in fb.c that maps LK-201 key events
to ASCII already deals with shift and right-shift.

It seems clear that whatever-it-is that gets done to the keyboard to
put it in a mode where it sends distinct key up/down events, which X
needs, is sufficient to confuse the X server.  Which leaves me
dubious of Matt's suggestion.

I'd guess (and it's just a guess) that the relatively old code in
X11R5 doesn't know about the LK-401 right shift.

What does xev show (both keysym and keycode) when you press
right-shift?  (I don't have an LK-401 to play with...)