Nils Soderman wrote:
> Hi,
> can anybody give me a tip to solve this: No interrupt when I run in
> debug mode, when I reset the board and let it run free interrupts are
> OK!
Maybe the controller did not execute the startup-code correctly or it's
a problem with remapping.
> Do I need any change in my debug setup when running revision 717 of
> OpenOCD compared with r279??? Probably?! I am "naturally" using the very
> good tutorial by Jim Lynch, but can't find any revision for r717 on the
> net.
As far as I know Jim Lynch's tutorials describe the usage of rather old
versions of OpenOCD. OpenOCD is a very "moving target". Please see the
OpenOCD documentation.
> I have the latest Yagarto (Thankyou very much Martin!!!!!)
Yagarto is a work done by Michael Fischer not by me. So you have to
thank him.
(I'm the one who creates the "WinARM" package)
> toolbox
> installed, AND the program is running when let loose! But it is
> debugging like I did in the 70's with Z80!
> Anyone having a idéa what can be the problem?
I'm currently using this setup and debugging works as expected (hope I
have not forgotten something). Maybe you can get some useful information
for you setup from this:
- Codesourcery G++ lite arm-none-eabi Q1-2008 (GNU toolchain and gdb)
- Eclipse "Ganymede" (3.4) with CDT
- Eclipse C/C++ GDB Hardware Debugging plugin (this is not the Zylin
Plugin)
- OpenOCD SVN922 self-build with mingw/minsys from berlios svn-checkout
the "directory tree" for openocd looks like this:
c:\embeddev\tools\bin - openocd.exe here
c:\embeddev\tools\interface - the interface directory for openocd svn
c:\embeddev\tools\event - the event directory from openocd svn
c:\embeddev\tools\target - the target directory from openocd svn
(this is because openocd will start search in ".." from the exe)
- Amontec JTAGkey
- AT91SAM7SE512 on AT91SAM7SE-EK
- Openocd launched with
openocd -d1 -f interface/jtagkey.cfg -f openocd.cfg
jtagkey.cfg is from the openocd SVN (you can browse it from berlios)
for the Olimex-"cable" you can use interface/arm-usb-ocd.cfg
openocd.cfg looks like this:
debug_level 3
fast disable
jtag_speed 200
jtag_nsrst_delay 200
jtag_ntrst_delay 200
script target/eir-sam7se512.cfg
jtag_khz 6000
fast enable
gdb_flash_program enable
init
debug_level 1
air-sam7se512.cfg is again from the openocd SVN ("target-library")
- I have 3 debug configurations in Eclipse using the
"CDT GDB hardware debugging":
(a) "debug"
Tab Main:
- Project from Workspace (the only on in my workspace)
- C/C++ application: the elf-file generated during build
Tab Debugger:
- gdb-command: arm-none-eabi-gdb (comes with CS G++ lite)
- Command set Standard, Protocol MI, Verbose console checked
- Use Remote Target checked, Device generic, Host: localhost,
Port: 3333
Tab Startup:
- Halt checked
- Entry in textfield:
set mem inaccessible-by-default off
- Run Commands textfield:
stepi
(everything else unchecked)
Tab Source: did not change anything
Tab Common:
- Save as local file
- Display in favorite Debug checked
- Allocate console checked
- Launch in background checked
(b) "reset and debug"
Everything as in (a) but halt unchecked and in the Init-Command
Textfield:
monitor debug_level 3
set mem inaccessible-by-default off
monitor soft_reset_halt
monitor debug_level 1
(c) "load, reset and debug"
Everything as in (a) but halt unchecked and in the Init-Command
Textfield:
monitor debug_level 3
monitor soft_reset_halt
monitor flash probe 0
load
monitor soft_reset_halt
set mem inaccessible-by-default off
monitor debug_level 1
The used application's startup and linker-script is configured to run
code from the internal flash (not RAM or external Flash) and is based on
an example from the Atmel software-package V1.4 for the AT91SAM7SE-EK.
I'm all but sure if this is the best or even a correct configuration but
it works for me. I have to say that now that the debugging is working I
really enjoy developing with Eclipse.
Hope this helps
Martin Thomas
Edit (b) and (c): added halt unchecked