RESTART_CAUSE() gives a number not in the device header

miketwo · Joined: 04 Aug 2010 Posts: 24

When I poll restart_cause(), it gives me back 16. However, in the device header file there is no 16 -- the highest is 15 which is a trap conflict.

First, versions of everything:
Compiler -- PCD 4.109
MPLAB -- 8.53.00.00
MPLAB Plug-in -- 2.0.0.12
MPICD3 -- 1.0.3.21

I'm programming to a PIC24FJ256GA110.

I can't quite figure out where the error is yet. It's not always repeatable but I think it's a compiler issue with indexing or sizing a struct whose fields land on odd memory locations. That's just a guess though. I was hoping this question has an easy answer that might point me in the right direction.

Cheers,
Michael

FvM · Joined: 27 Aug 2008 Posts: 2337 Location: Germany

I don't think, that restart_cause() works correctly with PCD.
You may want to read RCON directly.

miketwo · Joined: 04 Aug 2010 Posts: 24

Thanks. I did that and it seems to work. I get a bizarre reset -- none of the bits are set except for the 5th one -- SWDTEN: Software Enable/Disable of WDT bit.

I'm guessing restart_cause() is as confused as I am over this. Weird.

Thanks though! That helped a bunch!

FvM · Joined: 27 Aug 2008 Posts: 2337 Location: Germany

I noticed, that you are also using ICD3. Why don't simply check the RCON state after the restart, before it's modified by the C startup code.

miketwo · Joined: 04 Aug 2010 Posts: 24

I'm still not sure how the PIC manages to reboot without setting one of the RCON bits, but I found the problem that causes it -- it's kind of crazy. I'll post it here in case anyone else gets a weird restart_cause().

In some autocoded library I was including, there was the following function call

FvM · Joined: 27 Aug 2008 Posts: 2337 Location: Germany

Did you actually observe a reboot? Or just a jump to an unexpected location?
The difference would reveal when debugging the code.