Lpc2000

When I tested bitbanging on a LPC2294, I were able to toggle bits at around 
15mhz - each access to a gpio set/clear register took 2 cycles. That's 
without serving usb (no idea how expensive this is on lpc devices), and 
without any processing, loop handling, and so on. I wouldn't use bitbanging - 
use a small cpld as a serial shift register, that would be way faster.

Emulator -> Target: Totally target dependant, the ARM7 TRM has information on 
how to debug these devices.
Usb-driver -> Usb-based-emulator: Totally dependant on the design. You could 
bitbang, which would be awfully slow at ms latencies. You could define a 
high-level jtag interface, in that case you would transfer your jtag 
commands. Or you implement all debugging inside the emulator, then your 
traffic is whatever protocol you use, GDB, RDI, Telnet.
Everything else also depends on these choices.

Regards,

Dominic

On Wednesday 14 December 2005 16:31, householder_lpc wrote:
> The spec is a level of performance that is significantly higher than
> Wiggler-based debugging.
>
> It would seem interesting to use an LPC with a USB controller as JTAG
> emulator, LPC2142 or so.
>
> The thing I'm not sure about is what the reference points are.
>
> What kind of traffic is there between emulator and target?
> What kind of traffic is there between host usb driver and usb-based
> emulator?
> What kind of software is needed between host usb driver and the debugger?
> What kind of traffic is there between the debugger and this program?
>
> With "what kind of traffic", I mean, what protocol is used between
> these "application layers"...
>
>