On Mon, Jun 20, 2011 at 6:14 PM, Øyvind Harboe <oyvind.har...@zylin.com> wrote:
> (..) I would dearly like to see SWD done before
> opening the floodgates to GPIO and other serial protocols.

SWD is packed-based half-duplex bus that defines TRN operation for bus
direction change and buffers reconfiguration. There is no SWD without
bitbanging the pins responsible for that TRN switch. It can be
implemented per-interface in a source code, or it can be more general
and bring even more use to existing code/hardware. The main reason for
writing the bitbang was driving RnW pin that is responsible for TRN
bus operation. This RnW can be implemented as one bit, or as bitmask,
and even more bitmask that require setting one bits high and other
bits low. This is must-have for transport other than JTAG. I will not
hardcore this functionality only for one interface, if it can be used
for all others with even more elegant solution, because I don't like
this approach and I find it too much in current source code - if it
was already there I would not have to write it from scratch! With
presented signal/bitbang API it is possible to easily create other
transports using smart interfaces, I bet even JTAG-only interfaces
will score, and it is totally backward compatibile. It will be
possible to quickly write custom TCL scripts that will provide lots of
more new features than hardcoding them for some specific solution with
specific hardware. Do you need anything more? Just let me know and
give some time! :-)

Best regards! :-)
Tomek

-- 
CeDeROM, SQ7MHZ, http://www.tomek.cedro.info
_______________________________________________
Openocd-development mailing list
Openocd-development@lists.berlios.de
https://lists.berlios.de/mailman/listinfo/openocd-development

Reply via email to