Heiko Panther wrote:
It looks to me as if the current serial jtag firmware does not allow
debugging of the MSP target. Is that correct?
yes
> Is any work going on to that end?
not right now. i was thinking about implementing the GDB protocol into
the box instead of my own protocol. that would open up the road to
debugging with GDB. i even started with some code (python modules to
talk the GDB protocol which are used to write unit tests against GDB
targets (gdbproxy/my own code), some C code that runs on the PC and is
nearly portable to the MSP, supporting memory and register read/write...)
however there is a small problem.. the debugging part is not open
sourced by TI. an open serialJTAG firmware could only support what
msp430-jtag does.
on the other hand, if the firmware where ready and just not implement
the breakpoint and single stepping, then a binary only version with
debugging would become possible. (this also implies that such a
serialJTAG firmware must not be GPL, it should be BSD style licensed)
a few technical problems are there too. some MSP430 silicons have a bug
in the JTAG logic that destroys RAM contents under certain conditions.
the F123 and some F4xx have that problem. the RAM has to be read and
cached between the breakpoints. the simple workaround is just not
supporting chips with that bug...
an other point to be considered is that there are USB<->JTAG addapters
out there. they are faster and support debugging (at least on some host
platforms)
chris
