Report on the current state of h4100 vs h4300 detection

Hello h4100-port,

  Well, h4300 keyboard support becomes more pressing problem, so I've
made some experimentation towards proper detection of specific model.
To remind, the whole idea I had is very detection of presence of the
keyboard, as there doesn't appear to be any other architectural
differences between models (but I'd appreciate if someone
double-check).

  The idea in technical terms was to try to use ASIC3 SPI port and
hope that on h4100 that would lead to timeout or clearly "N/A" value
(like 0xff). This idea was based on the fact that loading h4300_kbd on
h4100 causes "ASIC3 interrupt processing overruns", which usually
among other things can happen when ports lack proper signals.

  Alas, bad news - on h4100 SPI works pretty well, and reads 0, which
as far as I could imagine, would be what it reads w/o key pressed in
h4300 too. Anyway, let's see how h4300 does. h4300 users, please build
CVS HEAD, and send dmesg after boot. We're looking for following line:

h4000_machine_is_h4300: read 0 from spi, counter: 133

(that's from h4100). Anyway, please send complete log, maybe I'll see
something interesting there.

-- 
Best regards,
 Paul                          mailto:pmiscml_at_gmail.com