Ignatios Souvatzis wrote:

> I didn't look at the code yet. But:
> as the stuff depends on a standard super i/o chip, most of this should be
> suitable for many normal com devices, too, is this right?

Well, maybe!  It all depends on how you're asking the question.  The UARTs
in this chip are 16550s, and that assumption is buried through most of the
code.  Of the major bits:

- Init.  Very device-specific, but this is handled by three
         mode-driven tables, so it should be easy to replace.
- IR.    Chip-specific, but the capabilities it assumes will
         most probably be present in other IR implementations,
         such as synchronous TX/RX.  After init, the only
         really device-specific assumption is the ability
         to disable the receiver, yet have it automatically
         re-enable on subsequent IR data...
- UART   Straightforward, should work with any 16450/16550-
         class device.
- Bitstream: The 'bitstream' drivers which drive the rtscts
        and dtrdsr devices should work with just about any
        16450/16550-class UART.
- Etc:  Most of the timing, RLL-generation/decoding,
        system I/O code, etc, should be fairly device-neutral.

Hmmm...  looking at this list, I think the code is fairly device-specific.
However, since National Semiconductor drove the Super I/O market for a
while, other manufacturer's parts are often compatible - it shouldn't be
that hard to support similar "PC-class" Super I/O chips...

MarkF