DockStar JTAG

[mrbill]

I've put this pinout, along with TIAO example wire colors and their diagram, up at

http://www.mrbill.net/dockstar/tiao/

[ygator]

I have just unbricked a dockstar with the steps in the wiki.
Thanks for everyone for doing the leg work on this.

@mrbill
Thanks. You might just want to add the purple as being no connection as well since people might think if was forgotten.

[mrbill]

Updated the diagram to reflect NOCONNECT for orange and purple, and added a snippet showing pin numbers for the Dockstar JTAG header.

[ygator]

@mrbill

Looks good! Thanks

[klingon]

Glad it worked out for you. I struggled for 2 days thinking I got a bad JTAG and was about to exchange it! Until I contacted Kragorn for his wiring.

[bzboi]

@mrbill - would you mind updating the Wiki to include your diagram and findings? Thanks!

[mrbill]

Done, added to http://Arch Linux ARM.com/index.php5?title=Ha ... n_via_JTAG

[jtagfun]

$this->bbcode_second_pass_quote('klingon', 'T')his is what I learnt from Kragorn if you're using Tiao's universal JTAG set to buffered mode and wiring instructions from http://www.tiaowiki.com/w/Image:Univers ... er.597.jpg

- Green wire -> pin 4 (of Dockstar)
- Yellow -> pin 5
- Red -> pin 3
- White -> pin 2
- Blue -> pin 6
- Black -> pin 10

Do NOT connect the orange wire (n_SRST in the JTAG) or openocd will give you the errors you're seeing. Also make sure you supply power to the JTAG using the mini-USB plug. Hope this works for you.

[bzboi]

I believe that the folks using the wiggler have not had to do it even without SRST#. I think that in this case, what matters is whether TRST# is connected. Try without the workaround first and if that doesn't work try with the workaround.

http://code.google.com/p/openrd/issues/detail?id=7

From what I understand, the cause of the problem is that the JTAG reset line is driven push-pull instead of open drain. Whether the reset actually happens depends on the initial (out of reset) state of the SoC I/O and the reset threshold of the chip, so it is likely to vary from board to board. I did not need the workaround when I had a working uBoot in mtd0, but after I erased mtd0, I needed the workaround to continue to load uboot to RAM.

It may very well be that I'm the only person that encountered this, so I welcome more feedback on who else had to do this so we can definitively keep/remove the workaround from the Wiki.

[klingon]

$this->bbcode_second_pass_quote('jtagfun', '
')When n_SRST is not used, do you need to push the Reset button at all (the "work around" suggested by wiki) when starting OPENOCD?