Alan B. Pearce wrote :

> I get the impression, Peter, that you are 
setting out to do 
> exactly what I did using Olins development 
environment...

You might have to study this thread a bit more closely.

What Peter are doing is to (and this could have been
any code, not 
just some ISRs) let the linker "link"
(or rather "chain") code in such 
a way that when one
code just runs "out into the blue", another code 
snipped
(from another source code file) just happens to have been
put 
*physicaly* right after the first, so the PIC
can just flow from one to 
the other without any
call/return code.

Peter thought this was clever, 
I (and I think some
others) thought it was a bad idea for various 
reasons.

Peter is using the fact that the linker (he is not
using 
MPLINK) puts code in the same segment
from different .o files, in the 
same order that the .o
files are specified on the linker command line.
And by this, "chaining" code hard together, head
to tail.

I don't know 
anything about that other linker,
but it would surpice me a lot if this 
fact actualy
is documented in that linkers documentation.

A supported 
way of doing this, would be to
have the ISR's as source files and then 
use the
INCLUDE directive to chain them together. One
could then be 
sure that the code would be an
continous segment. Something like :

INCLUDE   ISR_TOP ;context save
INCLUDE   ISR_A
INCLUDE   ISR_B
INCLUDE   ISR_C
INCLUDE   ISR_D
INCLUDE   ISR_END ; context restore and 
return

Any of the ISR_X could be commented out
without breaking the 
(ISR) code.


Regards,
Jan-Erik.



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