SCE-MI 2 wd updates - revision 1.10

Oops, I just remembered that this will most likely not reach
anyone with the attachment.

I'm sending separately to a smaller group. Anyone else wanting
it can request it to the reflector to directly to myself.


Greetings ITC Techies,

As promised last week, I've made updates to the SCE-MI 2 working document.

Mainly this was just updated with the merged pipes/fifos proposals along
with the TLM proxy example in the appendix to use the new interface.

I've also removed the HDL non-blocking interface for now. If we decide
to we can add this later to the appendix as a suggested reference
implementation. I heard opinions which I generally agree with that
we might want to keep the HDL side of pipes blocking only for now.

BTW, this update to the spec does show the support for arbitrarily
sized buffers to be passed to and from the C side of transaction
pipes that Per and I have been working on.

Along with that is an example of a reference implementation of
the blocking dpi_pipe_c_send() using this mechanism.

At some point it might be nice to update the ref. implemenation of
dpi_pipe_c_try_send/receive() to use the new unlimited buffer support
enhancements.

-- johnS
______________________________/\/            \     \
John Stickley                   \             \     \
Mgr., Acceleration Methodologies \             \________________
Mentor Graphics - MED             \_
________________________________________________________________