Thu Jun 4 20:04:47 CEST 2009

multiple devices

Before getting into other busses, I need a way to access 2 separate
chips from the same debugging session.  I tried communication
protocols before, but without a proper way to control _both_ sender
and receiver testing becomes quite difficult.

Some observations:

  * Async serial ports simply work.  There is no substitute for
    interfacing to a PC.

  * Staapl monitor communication is host directed rpc / half duplex.
    Slaves are quiet unless addressed.  This means all slave outputs
    could be wired together.

  * Bit-banging serial _output_ data is a lot simpler than input.  The
    debugger could send out a protocol extension which sends the slave
    mask before the message.

  * The router could perform the reply OR in software as its inner
    loop.  It doesn't need to understand the protocol, just combine

  - build the router.
  - use it to bootstrap monitor protocol over I2C.