Re: [Openocd-development] Problem with LPC2294, OpenOCD and GDB

[Andreas S. <ad...@co...>]

Am 25.04.2010 15:21, schrieb Andreas Schmidt:
> Hello,
>
> at the moment I'm trying to get my Olimex LPC-E2294-8MB Board running 
> with OpenOCD and GDB. Unfortunalty I had only very little success. 
> I've compiled the current OpenOCD version from GIT with FTDI drivers. 
> Everything worked well for the first time I start debugging with GDB 
> (Version 6.8 and Version 7.0). But in a non-deterministic manner I get 
> problems with stopping at breakpoints when starting the debug session 
> and after pressing halt or stop I see that the debugger is halting in 
> the data-abort handler. The strange behavior is, that the same code 
> runs very well the first time and without any recompiling or code 
> modification the debugger doenst work right and stops in data-abort 
> handler. sometimes, after 5-8 trys it works again. But restarting the 
> hardware and restarting openocd doenst change it. If I connect with 
> telnet to openOCD it seems that every thing is working well. Checking 
> PC and Stepping works and my code does not hang is the data-abort 
> handler.
>
> I would assume, the the code works well an no illegal memory reagion 
> is accessed. After googling around the web I found this mailinglist 
> thread form Michael Fischer 
> http://lists.berlios.de/pipermail/openocd-development/2008-May/002068.html 
> talking about a similar problem with LPC and GDB. But I'm not sure 
> whether this could be the same or similar problem because the thread 
> is from 2008 an OpenOCD has been developed further.
>
> If some can give me some hints about this I would be very happy....I 
> don't know what do to anymore.
>
> Thanks
>
> Beste regards
> Andy
> _______________________________________________
> Openocd-development mailing list
> Ope...@li...
> https://lists.berlios.de/mailman/listinfo/openocd-development

Hi all,

it seems I've solved the problem. Well, I'm not sure whether I solved it 
or is my solution only a workaround. I use Eclipse with Zylin Embedded 
CDT. In my Debugger Run-script I had this:

break main
continue

With this code, the strange behaviour occured. After changing it to

break main
monitor soft_reset_halt
continue


everythin worked fine. The typically cycle of 
compiling-loading-debugging-compiling-... worked even after 20. loops 
without restart. So I would assume this works stable. Also a restart of 
Eclipse and OpenOCD has no effect on stability.

Beste regards
Andy