Atmel Studio 6.2, AVR ONE!, JTAG to AVR32 (AT32UC3B0256)
Picking up some work after a few months, I re-opened the project in Studio, connected to the AVR ONE!, read the chip sig and voltage as good, and burned the .hex image from when it was last worked on (including a chip erase). All was well and the code executed as expected.
Making some changes to the code, I attempted to use the debugger the same way as historically and attempted to set a breakpoint. The breakpoint marker was solid red. However, when I pressed the green arrow or F5, it ran past the breakpoint, which had turned open-circle/yellow exclamation mark, "The breakpoint will not currently be hit. Unable to set requested breakpoint on target." The (changed) C code ran as expected - I can tell from a flashy LED that it's compiled and executed rather than the older code in the original .hex image.
I have also tried Debug pulldown, Start Debugging and Break. The code starts and immediately stops. The green "Go/Continue" arrow will start the code running, and again the breakpoint turns open-circle and the program runs past it.
Whenever the code is stopped (e.g. with the Ctrl-Shift-F5 square) the breakpoint becomes solid red again.
In the Debug menu/pulldown, "Attach to Target" is forever greyed out. This symptom seems...important?
I have tried/checked all of the following:
- Complete un-install/re-install of Atmel Studio 6.2 and the AVR ONE! USB drivers
- Making a new Project, importing the ASF modules, coping over and adding my own .c/.h files, save, clean, and rebuild everything
- Checked Project, Properties , Build, Configuration = Debug
- Checked Project, Properties, Toolchain > Compiler, Optimization = -O1
- Checked Project, Properties, Toolchain > Assembly > Debugging, Level = Default (-g)
- Checked Project, Properties, Tool, Programming Settings = Erase entire chip
- Ensured that the breakpoints must (should) be hit in the code, right near the beginning of main()
- Tried on a completely different target (AVR32 on a different board)
- I've made many other hail-mary tweaks (and set them back again) with equal lack of effect, but won't spam this post with them unless someone suggests they may be relevant.
And NOTHING has enabled me to keep a breakpoint active and stop at it during debug/run.
I tried a second AVR ONE! with the same results as above. On a whim I went into View , Available Atmel Tools. Both claimed their firmware was up to date (5.1b, according to Tools, Device Programming). I then ran the Self-Test, and both failed with the error:
"The AVR ONE! self test failed. Be sure that the test adapter is connected as described in the help documentation.
Both the MICTOR38 connector and the 10-pin connector contains malfunctioning signals. Most likely the probe and/or the probe cable is broken."
Both failing with the same error struck me as a little suspicious, and a few elderly fora threads suggested this may be an (older) firmware issue rather than a hardware issue, although none specifically referenced the 5.1b firmware. So I don't know.
I'm now at a loss. If anyone has any insight or suggestions, they will be gratefully received, and I will report back on progress (or lack thereof).