Lpc2000

Hi,

I have a weird problem (hmmm, what problems aren't weird?) regarding 
the optimization flag in GCC.

I use GNUARN 4.0.1, Eclipse 3.0.1 and an board from Olimex for my 
project. The Olimex-uC talks to a PC-app via UART 1. Work is 
progreesing but then I got a Data Abort interrupt indicating 
something was wrong.

So my idea was to try to debug the uC-code from Eclipse with OCD 
Remote. If I do it with the code compiled with -03 (as it shouldn't 
be according to James P. Lynch's excellent guide) it starts up 
alright and the app talks with the PC. But when I set a breakpoint 
and try to single-step through the code I get the error "Target 
request failed: Target is not responding (timed out).".

OK, so lets compile the code using -O0 as the Lynch-guide tells me. 
Now the application no longer talks to the PC-app at all. 
Indications tells me that the UART1 clock is wrong.

Anyone that has ideas on why less optimization can cause problems 
for me? I know that the -Ox are "aliases" for a bunch of 
optimization-flag but I haven't looked into them.

And maybe someone can tell me why I get the error message when 
debugging and compiling using -O3.

/Anders

andersryl wrote:

>Hi,
>
>I have a weird problem (hmmm, what problems aren't weird?) regarding 
>the optimization flag in GCC.
>
>I use GNUARN 4.0.1, Eclipse 3.0.1 and an board from Olimex for my 
>project. The Olimex-uC talks to a PC-app via UART 1. Work is 
>progreesing but then I got a Data Abort interrupt indicating 
>something was wrong.
>
>So my idea was to try to debug the uC-code from Eclipse with OCD 
>Remote. If I do it with the code compiled with -03 (as it shouldn't 
>be according to James P. Lynch's excellent guide) it starts up 
>alright and the app talks with the PC. But when I set a breakpoint 
>and try to single-step through the code I get the error "Target 
>request failed: Target is not responding (timed out).".
>
>OK, so lets compile the code using -O0 as the Lynch-guide tells me. 
>Now the application no longer talks to the PC-app at all. 
>Indications tells me that the UART1 clock is wrong.
>
>Anyone that has ideas on why less optimization can cause problems 
>for me? I know that the -Ox are "aliases" for a bunch of 
>optimization-flag but I haven't looked into them.
>
>And maybe someone can tell me why I get the error message when 
>debugging and compiling using -O3.
>
>  
>
IIRC, -O3 is a synonym for -Os, and if you can debug the code produced 
under -Os, more power to you!  If I need to debug a section of the 
program I'm working on, I turn off all optimizations: -O0.

TomW


>/Anders
>
>
>
>
>
> 
>Yahoo! Groups Links
>
>
>
> 
>
>
>  
>


-- 
Tom Walsh - WN3L - Embedded Systems Consultant
http://openhardware.net, http://cyberiansoftware.com
"Windows? No thanks, I have work to do..."
----------------------------------------------------

> I have a weird problem (hmmm, what problems aren't weird?) 
> regarding the optimization flag in GCC.
> 
...
> 
> So my idea was to try to debug the uC-code from Eclipse with 
> OCD Remote. If I do it with the code compiled with -03 (as it 
> shouldn't be according to James P. Lynch's excellent guide) 
> it starts up alright and the app talks with the PC. But when 
> I set a breakpoint and try to single-step through the code I 
> get the error "Target request failed: Target is not 
> responding (timed out).".

I get this error message now as well in Insight (or it thinks the target
has continued running) I think it correlates with moving to the newer
OCDremote which handles hardware breakpoints. So one step forward
(hardware breakpoint support) and one back (stuffed after 1 breakpoint).
I haven't had time to investigate further; it may be a hardware problem
with the PC I'm using as someone stole the PC I was using for debugging.

Cheers,
Bruce