Large function causes program to fail [DONE]

nilsener · Joined: 06 Dec 2005 Posts: 59

PCH 4.026
18F2620

Dear,

I have a large function with approx. 1000 code lines. All is working fine but after inserting an other approx. 15 code lines then the program fails. It compiles without problems and even if the function is not called from main(), the program fails. I got one watchdog reset after an other, that means the main() function is no more working. Restart_Wdt(); is already included. If I set NOWDT, I get watchdog reset, too!!!

Is it possible that the function with 1000 lines is too large and that this is the problem?

Thanks for any help!

nilsener

Foppie · Joined: 16 Sep 2005 Posts: 138 Location: The Netherlands

Maybe it can be solved with a different compiler version.

The CCS site gives the following bugfix for release 4.031. This might have to do something with your problem.

PCM programmer · Joined: 06 Sep 2003 Posts: 21708

icesynth · Joined: 03 Sep 2007 Posts: 32 Location: Edmonton, Alberta

What is the memory usage reported in your .lst file when you run the compiler? should look something like this:

Guest · Posted: Tue Sep 04, 2007 2:40 am

Dear icesynth,

I have this report in my lst file, do you think it is ok?:

CCS PCH C Compiler, Version 4.026, 31622 03-Sep-07 17:05

Filename: DFS73_001.lst

ROM used: 33026 bytes (50%)
Largest free fragment is 32506
RAM used: 781 (20%) at main() level
907 (23%) worst case
Stack: 10 worst case (6 in main + 4 for interrupts)

Guest · Posted: Tue Sep 04, 2007 7:17 am

PCH 4.014
18F2620
ICD2
MPLAB 7.22.00.00

Dear PCM Programmer,

thanks for your answer.

In the meantime I found, that the problem occurs if my program reaches exactly the size of >= 32768 byte = 50% of total program memory space. I use ICD2 and MPLAB 7.22.00.00. The Program Memory is programmed from 0x0 - 0x7FFF (32767 byte) even if the program is larger and so there will be a lack of code in the PIC. I think this is a strange error and I have no idea what to do. By the way, the reset is not coming from the watchdog. Restart_Cause() gives me a 15 that means MCLR_FROM_RUN but there is nothing like that.

Do you have any suggestions?

Best Regards
nilsener

nilsener · Joined: 06 Dec 2005 Posts: 59

PostPosted: Tue Sep 04, 2007 7:17 am Post subject:
PCH 4.014
18F2620
ICD2
MPLAB 7.22.00.00

I found in MPLAB - Programmer-Settings-Program that the Program Memory is set to 0x0 - 0x7FFF (32kb). If I change to 0xFFFF (64kb) MPLAB changes it automatically back to 0x7FFF. How can I change to 0xFFFF?

Best Regards
nilsener

Ttelmah · Guest

'15', can have a variety of meanings, dependant on which bits CCS are actually checking, and whether some bits have been reset since the last boot.
The one that I'd be particularly suspicious about, is that a 'stack full' reset, also gives the same pattrn, but with the 'STKFUL' bit added. CCS, don't include this in their testing, so it'd be worth adding a check on this bit, and seeing if this might be the cause.

Best Wishes

icesynth · Joined: 03 Sep 2007 Posts: 32 Location: Edmonton, Alberta

Try upgrading MPLAB to the latest version (7.60), I tried programming the device here, and had no trouble.

http://ww1.microchip.com/downloads/en/DeviceDoc/mp760a.zip
_________________
Programming for the the real world.
--Chris Burchett
Sylver Technologies Inc.

nilsener · Joined: 06 Dec 2005 Posts: 59

Dear,

problem solved. I have upgraded from 18F2520 to 18F2620 and forgot to change settings in MPLAB to 18F2620 and so it programs only to 0x7FFF (32kb) instead of 0xFFFF (64kb). I thought that it is enough to change it in my CCS Code. So it was an very easy problem and I feel shame.

Thank you all for your help!!!

nilsener

icesynth · Joined: 03 Sep 2007 Posts: 32 Location: Edmonton, Alberta

We all have those days Laughing

I'm happy you got it fixed.
_________________
Programming for the the real world.
--Chris Burchett
Sylver Technologies Inc.