Re: [avr-gcc-list] r-engineering the JTAG-ICE protocol

[Torsten Mohr]

Hi,

great!  I also wrote some JTAG software for the parallel port to
debug my ARM7TDMI.  When i read that the ARM controllers also
have a JTAG port but Atmel won't release its specification
i thought about reverse engineering it.

Building some hardware that implements a JTAG TAP controller
could help.  This hardware would get the JTAG signals TMS, TDI,
TCK, TRST and would behave like a JTAG controller.

It could be build with some logic like 74xxx or with programmable
logic and should be put in parallel to the JTAG-ICE and the AVR
controller that is being debugged.

When in state Shift-IR or Shift-DR TDI could be shifted into
a shift register that could be read from an external micro
when the Update-IR or Update-DR state would be reached.

This way, the speed of the JTAG-ICE shouldn't be a problem.
Has anybody measured the speed the JTAG-ICE works with?
What is the time period of a Clock pulse?

This software for the parallel port i wrote runs completely in
user space, though it depends on the /dev/parport devices of
Linux.  It can:
- test the parallel port (reading/writing pins with a user interface)
- read/write several EEPROMs (24Cxx, 93X6, M95XXX) (batched or interactive)
- read Gamepads (Playstation)
- debug an ARM7TDMI (read/write registers/memory, download S-Records
  and start them)

I'll release that library under a GPL soon, at the moment i'm still
fighting with my fresh installed Linux system.


Best regards,
Torsten.

avr-gcc-list at http://avr1.org