On 2011-12-21 14:52, Jeremy Bennett wrote:
On Tue, 2011-12-20 at 17:04 +0100, Steve Battazzo wrote:
Hi all,
I have connected a new JTAG debugger based on FT2232D to our OR1200
using OpenOCD.
This basically works OK, but the two LED indicators are always
alternately flashing after I connect.
Then, when I make a connection to GDB, I see the following messages:
Warn : acknowledgement received, but no packet pending
undefined debug reason 6 - target needs reset
It's a bit inconsistent, but I also get strange behavior in GDB. It
appears that loading the elf files is working very well, but about half
of the time when I run "c" I get something like:
"The target is not in the halted nor runing stated, stepi/continue
ignored."
followed by:
"Program received signal SIGINT, Interrupt."
At this point, the execution is at address 0x104 so it's happening on
the first instruction.
Anybody seen something like this?
Or does anyone have any recommendation of what I can do to start looking
at why this might be happening?
Hi Steve,
You might want to turn on the GDB flag to trace the debug packets being
sent and received. From memory I think it is
set debug remote 1
That may give you a clue. You'll need to look at the section on the
Remote Serial Protocol in the GDB manual. But it is a simple protocol,
so it should be clear if you are getting nonsense in a reply packet.
HTH,
Jeremy
Hi Jeremy,
Thanks for the tip. And yeah, that syntax is right.
It does look like I'm getting back strange things in the reply packets,
regardless of which debugger I use.
Here are the different behaviors I'm seeing:
ORSoC debugger, or-proxy: garbage-looking replies in GDB.. one JTAG
status LED comes on and it stays in that state.
ORSoC debugger, OpenOCD: garbage-looking replies in GDB, no JTAG status
LEDs, SIGINT at 0x104 on first one or two tries to load and continue
(inconsistent).
New debugger, OpenOCD: garbage-looking replies in GDB, JTAG status LEDs
always flashing alternately, SIGINT at 0x104 on first one or two tries
to load/continue (inconsistent).
Also sometimes on the very first run some garbage shows up on the serial
terminal of the UART coming through the debugger.
I wonder if maybe some of this could be fixed if I update to a newer
version of OpenOCD though?
Steve
_______________________________________________
OpenRISC mailing list
[email protected]
http://lists.openrisc.net/listinfo/openrisc
