Re: SuperH DTF connection protocol

[Andrew Stubbs <Andrew dot Stubbs at st dot com>]

On Mon, 20 Aug 2001 tm@kloo.net wrote:

> Andrew Stubbs wrote:
> ...
> 
> > The standard setup for newlib on the sh (I don't know about other
> > platforms) is to communicate with the outside world (read, write, open,
> > ...) using a trap. As I understand it this is something understood by the
> > gdb simulator - it does not work on an actual sh chip and board.
> 
> It can work with an actual SH chip and board.
> 
> I've ported the debug stub to the SH4 twice, and I redirected the traps
> to print the printfs inside the debugger using the "O" command (iirc).
> 
> Also, I think GDB supports an encapsulation protocol whereby you can
> send
> private data through the GDB link. It's been a while since I've looked
> at the source, so you may want to verify this.

Yes, but as I understand it you have to define this protocol yourself - at
least you have to define how it uses the hardware - gdb just uses it.
Maybe I have got this wrong, but what is certain is that the debugger guys
here have a gdb server working with DTF as the protocol. This is why we
want newlib plumbed into DTF somehow (in a more official way than we
currently do).

------------------------------------------------------------------------
|  Andrew Stubbs,                   |  E-mail: Andrew.Stubbs@st.com    |
|  STMicroelectronics Ltd.,         |  or      stubbsa@bristol.st.com  |
|  1000 Aztec West,                 |                                  |
|  Almondsbury,                     |  Tel:    +44 (0)1454 462325      |
|  Bristol, BS32 4SQ, U.K.          |  Fax:    +44 (0)1454 617910      |
------------------------------------------------------------------------