CW 7.1.1a on M52259DEMOKIT -- BDM, breakpoints, timer ISRs...

cancel
Showing results for 
Show  only  | Search instead for 
Did you mean: 

CW 7.1.1a on M52259DEMOKIT -- BDM, breakpoints, timer ISRs...

1,660 Views
RichTestardi
Senior Contributor II
Sorry, I'm reposting from the ColdFire forum, hoping someone here might know a workaround...
 
Hi all,
 
I am wondering about something I saw while debugging my program on CW 7.1.1a on the M52259DEMOKIT board...
 
This board apparently uses a new BDM, as I have to specify "ColdFire v2-v4 JM60 OSBDM" to debug it.
 
In general, the new BDM seems to work, but it has one strange behavior I am wondering if I can work around.
 
With the old PEMICRO USB BDM, when I hit a breakpoint in non-ISR code (with ISRs enabled) and then single stepped, I stepped following the expected code path -- i.e., I didn't instantly step into my 1ms PIT timer interrupt service routine.
 
With the new ColdFire v2-v4 JM60 OSBDM, on the other hand, if I break into my code anywhere other than in my PIT timer interrupt service routine, the first single step puts me right there!
 
I managed to get all my debugging done by setting breakpoints, rather than single stepping, but this seems like something that should be configurable, at least...
 
Is there a way to tell the debugger not to just step me into my PIT timer ISR every time I step?  Or might I just have something misconfigured?
 
As a related issue, when I insert a breakpoint in my code, I always hit it *twice* now -- once apparently on entry to the PIT timer ISR, and once apparently on return.  (I've gotten used to this and just always continue once when I hit a breakpoint.)
 
Thanks for any help!
 
Other than this, the port to the MCF52259 took less than a day, and it seems fast -- about 5% faster than a MCF52221 clock-for-clock, which I can't really explain but will not complain about!
 
-- Rich
Labels (1)
0 Kudos
2 Replies

407 Views
J2MEJediMaster
Specialist I
It appears that the timer interrupt fires while you're at the breakpoint and of course the debugger gets hauled away into the timer ISR. Can you try to inhibit the timer ISR to prove that this is the case?

---Tom

0 Kudos

407 Views
jhmrd
Contributor III
I have also run into this problem, and by setting the IMRL0 and IMRH0 registers to all FFFFFFFF's the debugger will single step. Has there been a more elegant solution found for this problem?
0 Kudos