Olimex Support Forum

ARM => JTAG => Topic started by: afremont on January 19, 2013, 07:38:20 PM

Title: OlimexODS single stepping source problem
Post by: afremont on January 19, 2013, 07:38:20 PM
I installed the 9(k) release of the ODS and have most of it working now.  I can do everything except single step the source reliably.  I have the following statement as a delay loop:
for(j=0; j<50000; j++);

When I try to step over it with F6, the debugger shows that it is single stepping still, but the source line cursor doesn't show in the source window.  At that point, it's basically stuck.  The PC doesn't advance after this.  If I telnet to openocd, and issue "step" commands, the program is stepping around making no sense at all, but the PC changes just not predictably.  Any ideas what I need to update or do?

BTW, I'm using an LPC-2378STK Rev B board that I've owned for a few years now.  I've done this all before, but it's been a while.  It all seems to be doing what it should, but the single stepping of the source over the for loop hangs it up.
Title: Re: OlimexODS single stepping source problem
Post by: afremont on January 19, 2013, 09:42:07 PM
Ok, it looks like I just didn't wait long enough.  It's taking forever (like 30 minutes) to get thru the loop.  Is there anything I can do to speed that up?  I am currently running/debugging from flash.
Title: Re: OlimexODS single stepping source problem
Post by: ottelo on January 22, 2013, 02:31:03 PM
I've the same problems with CooCox IDE. When I add breakpoint to a line in a function (not in main) it breaks but if I click un Run to get to another breakpoint, it doesnt work and it still on the first breakpoint line and the variable havent changed! If I click on next step I get to an Interrupt ISR ...

Title: Re: OlimexODS single stepping source problem
Post by: afremont on January 22, 2013, 07:07:53 PM
I used to know how to do this, but it's been about four years now since I've tinkered with it.  I think I just need to get my GDB and OpenOCD config files tweaked a bit more and I'll have it.

I understand that there are only two hardware breakpoints.  I'm sure that's why the stepping is so slow, and when I try to do a run-to-line command it complains about being out of hardware breakpoints.  I thought there was some "feature" that allowed more efficient debugging from flash.