Re: [avr-gcc-list] avarice and I/O ports during JTAG programming/debugging

[Theodore A. Roth]

On Tue, 18 Nov 2003, Brian Cuthie wrote:

> 
> Hi All,
> 
> I'm seeing some strange behavior from the I/O ports on an ATmega16 while
> doing JTAG debugging and programming using avarice. Specifically, I'm
> seeing a bunch of state changes on I/O lines programmed for output (from
> code already in the avr chip) when avarice reads the configuration and
> fuse bytes. This concerns me a bit, since it can create temporary I/O
> states that may be damaging to hardware interfaced to the processor.
> 
> I have two questions:
> 
> 1) Is this behavior by design? or is this some issue with avarice?

This may be by design. Doesn't jtag boundary scan let you twiddle pins to 
test for opens and shorts?

> 
> 2) Would it possible for avarice to reset the processor before doing any
> other JTAG operations so that the I/O ports are forced to be dormant?

That depends are what jtag ops you are talking about. In most cases, I don't 
think it would hurt to reset the device. The only cases I can think of where 
you wouldn't want to reset is when you are using the --capture option. Nils
just did a bunch of work on that last week

It should just take call resetProgram() at the right place. Have a look in 
main.cc and see what you come up with.

Ted Roth