> I'm continuing to try to find the cause of this. It appears the original
> cause was a collision between the stack and the heap. I now write a known
> pattern between the stack and heap at startup and look at it after the
> system has been running for several hours to make sure the known pattern
> is still there. It looks like I'm not now getting the stack/heap
> collision.
>
> But, about midnight local time last night, I got a half dozen bus error
> exceptions, all at the same program address (0x9d031a38). Scanning up fro=
m
> there in the code view of memory, I see several jr ra instructions above
> with no labels before, so I assume this code is generated by the compiler
> and not from some library that I have source for (I think there'd be
> labels if this was compiled from source during my recent compile). So,
> it's difficult to determine what the function of the offending code is.
>
> Looking at the specific address that's generating the exception, the
> instruction is beq zero, zero, 0x1d031b1c. I assume 0x1d031b1c is the
> destination address of the branch (not the offset to the destination
> address). If so, this would be a relatively short branch, which would be
> within the branch range.
>
> I have a global called ProcessNum that I update in my main loop before
> going off to handle different tasks. I write the value of ProcessNum alon=
g
> with the exception info to a uart when I get an exception. The general
> exception is consistently happening in StackApplications(). I've added
> more ProcessNum assignments within StackApplications() to try to determin=
e
> which one is causing the problem. I'm also running WireShark to see what
> network activity happens around the time of the exception.
>
> Any other ideas as to what to try?
>
> THANKS!
>
> Harold
>
>
> --
> FCC Rules Updated Daily at http://www.hallikainen.com - Advertising
> opportunities available!


Looks like I finally found it! I left a system and WireShark running
overnight. My system crashed about 3:00am. Wireshark showed an HTTP GET
coming in that had an invalid parameter value. My code should have caught
that, but did not. The incoming parameter is allowed to be 1 to 10, which
I then drop to 0 to 9 to index into an array. The parameter coming in was
0, which I dropped to -1, which was not a valid array index. My range
check was looking for >=3D0 instead of >0. That fixed it!

These things can be hard to catch!

Harold



--=20
FCC Rules Updated Daily at http://www.hallikainen.com - Advertising
opportunities available!
--=20
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist
.