Re: [avr-chat] Avrdude and JTAG / Dragon...

[Vincent Trouilliez]

On Mon,  2 Mar 2009 17:05:30 +0100 (MET)
address@hidden (Joerg Wunsch) wrote:

> Yes, sure.  The limitation is only that the JTAG ICE is the JTAG
> master, and there can only be one master in the chain.  Since each
> JTAG ICE (or AVR Dragon, for that matter) can only talk to one AVR at
> a time, it can only debug one.  The other one is simply acting as a
> shift register, passing the other device's JTAG commands through,
> without being addressed by them.

Ah great :-)

> Just remember, while for plain JTAG programming, you can drive the
> JTAG clock as fast as possible (subject of the limitations of the
> passive circuitry around, see David Kelly's cable length issues), for
> debugging, you have to obey the fCPU/4 rule again for the JTAG clock.

Well I guess that's not a problem, since when debugging you are not
transfering big chuncks of data no ? You are just sending commands and
reading/writing variables. So no matter how slow the link, it will
still be pretty much real-time since pressing on a button on my
keyboard to step the program, or pressing the enter key to read/write a
variable, will take much more time than JTAG needs to perform the
operation...

--
Vince, just opened his mail box... received 3 brand new Mega32, in case
I blow one again ! ;-)