Re: [avrdude-dev] Re: How to talk to second device in JTAG chain?

[Joerg Wunsch]

As Grant Edwards wrote:

> > The attached patch adds a -j option that allows adjusting JTAG
> > daisy-chain parameters.  [...]

> > I cannot test it except by passing -j 0,0,0,0, and cross-checking that
> > specifying any other parameters results in the ICE being unable to
> > communicate with the target.

> Brilliant!  I'll give it a try.

Meanwhile, I've been able to setup a real scan chain, consisting of
one STK500 with an ATmega16, and a second one with an ATmega1281.  I
can confirm that the implementation actually works.  The tricky part
was to find out how many bits each AVR unit is shifting in the chain:
it's 4 bits.  So I could select the first AVR with -j 1,0,4,0, and
the second one with -j 0,1,0,4.

However, I somewhat changed my mind about the option name.  It doesn't
make much sense to eat up the entire alphabet for option letters that
often only apply to one or to programmer types.  Instead, I changed
the implementation to provide a generic -x option that acceptes an
"extended parameter" that is subject to interpretation by the
programmer backend.  That way, each programmer can implement its own
set of specific options.  Besides the JTAG scan chain parameters,
things like the STK500 target and AREF voltage come to mind here as
well.  (A good number of the existing options should probably have
been implemented that way already before.)

So find attached a new patch including documentation.  If there are
no objections, I'll commit that one to CVS.
-- 
cheers, J"org               .-.-.   --... ...--   -.. .  DL8DTL

http://www.sax.de/~joerg/                        NIC: JW11-RIPE
Never trust an operating system you don't have sources for. ;-)

daisychain.patch
Description: Text document