Hans Nieuwenhuis writes: > On Thu, 10 Jun 2010 11:33:42 +1200 > [email protected] wrote: > > > Hans Nieuwenhuis writes: > > > On Wed, 9 Jun 2010 14:30:29 +1200 > > > [email protected] wrote: > > > > > > > > Hi Hans, > > > > > > > > I (the author) would be interested in the Snoopy logs :). > > > > > > > > There is a change you might like to try. I has worked for some > > > > people but I haven't committed it yet as I'm still trying to figure > > > > out if it's going to break things for others: > > > > > > > > diff --git a/fet.c b/fet.c > > > > index 0907f87..1c745fd 100644 > > > > --- a/fet.c > > > > +++ b/fet.c > > > > @@ -812,7 +812,7 @@ device_t fet_open(transport_t transport, int > > > > proto_flags, int vcc_mv, /* configure: Spy-Bi-Wire or JTAG */ > > > > if (xfer(dev, C_CONFIGURE, NULL, 0, > > > > 2, FET_CONFIG_PROTOCOL, > > > > - (proto_flags & FET_PROTO_SPYBIWIRE) ? 1 : 0) < 0) { > > > > + (proto_flags & FET_PROTO_SPYBIWIRE) ? 1 : 2) < 0) { > > > > fprintf(stderr, "fet: configure failed\n"); > > > > goto fail; > > > > } > > > > > > > > This appears to be a slightly different JTAG mode, and apparently > > > > it's used by the new IAR software when querying devices on startup. > > > > > > Hi Daniel, > > > > > > Thanks, that one seems to do the trick for me. I briefly tested it > > > and now it starts up and initialises correctly. Haven't been able to > > > do any further testing due to time contraints. > > > > Hi Hans, > > > > That's good to know. > > > > This seems to fix it for most others with an 5xxx series too, but I'm > > a little hesitant to apply the change until I'm sure it's not going to > > break support for other chips. > > > > Perhaps the best idea for now would be to add a command-line flag to > > select mode 2 -- perhaps "-J" as an alternative to "-j". It's a little > > messy and possibly unnecessary but I don't have any better ideas for > > now. > > > > Any thoughts? > > Maybe you can parse the option -j with an optional argument which > specifies the method you use. -j would be the old way of doing things and > -j2 could use the fix for newer FETs. What happens when you use this > method on very old FET firmware? Do they understand this option?
Hi Hans, I have an older FET here, but haven't had time yet to try it out. I'd guess that the older FETs don't recognise this option though. As for the flag, parsing -j/-j2 is a slight problem for getopt, so it might have to be -j/-J. Another method that has occured to me is to try mode 2 if identification with mode 0 fails. - Daniel
