> Unfortunately, for the PIC16f886, the only access to the stack happens wi=
th the call,
 > interrupt and return instructions. There is no other access.
 > I think the only option is to instrument round the instructions in
 > ones own code.

In a previous life (before Jal) I experimented with the MPLAB macro=20
facilities. It turned out to be possible (using some weird combination=20
of #define and .macro, I forgot which) to have MPASM replace assembler=20
instructions like CALL X with an expansion of my choice. A usefully=20
expansion could be

incf level, f
CALL X
decf level, f

(The devil is in the details: preventing the CALL inside the macro from=20
being re-expanded. An easier option would be to insert the hex code for=20
the CALL.)

There are of course some nasty details: call and CALL, LCALL, and=20
coditionally skipping the call. Those are left as an exercise to the read :=
)

Getting excited, you could insert code checks for a stack overflow, and=20
if so, writes the line number to some error logging channel.

A problem: a stack overflow is harmless. It is the stack *underflow*=20
that kills your application. Instrumenting RETURN is probably no=20
problem, but instrumenting RETLW instructions might cause problems with=20
jump tables.

An if you use a HLL you will need access to the sources of the compiler.=20
With anything but Jal that might be a problem :)

--=20

Wouter van Ooijen

-- -------------------------------------------
Van Ooijen Technische Informatica: www.voti.nl
consultancy, development, PICmicro products
docent Hogeschool van Utrecht: www.voti.nl/hvu

--=20
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist
.