Hi Wookey , You should try to have -d3 debug log for both your 64bits and 32bits ! Then compare the 2 log files. And send us these two log files
Note: Amontec is very interested to your balloon3 board and I am sure you could be interested in our Amontec JTAGkey-2 (Hi-Speed USB 2.0 JTAG adapter & Interface cable, including RTCK with 1.3V to 5V JTAG IO level shifter @ 32mA by output) . We could think to exchange our products. Please contact me at laurent . gauch @ amontec . com if you are interested. Best regards, Laurent http://www.amontec.com ARM CORTEX JTAG Debugger Amontec JTAGkey (Tiny) / Amontec JTAGkey-2 / Amontec JTAGkey-2P > I have been having fun with various combinations of openocd, amontec > jtagkey and olimex usb-tiny and pxa270 and chianed and unchained JTAG > interfaces. We have one JTAG port to xilinx xcr3256 (coolrunner II > CPLD) and one to pxa270. > > I have a large table of what does and doesn't work which is all very > nice, but I suspect won't generally be of much interest here because > it all related to r1613 because that was the last version on which > xsvf or svf playing into CPLDs worked. I intend to work through svn > revisions, keeping the 1613-vontage xsvf code to see how far forward > one can go until that dies due to changes elsewhere. David Bisset is > also working with latest code to see if he can get it going again on > xscale - it seems that nor programing only works in simple-minded mode > - as soon as you try to use the debug-handler code in the xscale to > program NOR at a useful speed it fails. > > Anyway. The issue we have is that we are stuck with r1613 for the time > being and we can make that work with an amontec jtagkey in > chained-JTAG setup quite well (it segfaults every odd run and works > just fine every even run - which is about as good as it gets IME :-) > > However this turned out (when we tried to deploy it to production) to > only be true on amd64 machines. On i386 machines boundary scan always > fails. This is a bit of a pain (we've gone to buy another amd64 box > for now as openocd-fixing time has run out). > > It works the same (every other time) on amd64 debian testing and > stable boxes, with both libftdi 0.13 and 0.16. It doesn't work just > the same with libftdi 0.13 and 0.16 on debian stable i386 machines. > > So, if anyone has any bright ideas for fixing the underlying xsvf > problem so we can move to newer code or why it should be different on > i386. > > Is anyone else using OpenOCD to program coolrunner CPLDs? I suspect > not otherwise there should be more grumbling that it's been broken for > quite some time. > > all our config files are available at > http://balloonboard.org/cgi-bin/viewcvs.cgi/balloon/trunk/utils/openocd/ > > The case I am talking about above is: > openocd -s utils/openocd -f balloon3-amontec.cfg -f loadloon.cfg -f > shutdown.cfg > > Wookey > -- > Principal hats: iEndian - Balloonboard - Toby Churchill - Emdebian > http://wookware.org/ _______________________________________________ Openocd-development mailing list [email protected] https://lists.berlios.de/mailman/listinfo/openocd-development
