James:
I would like a response to this issue from Keil. I posted to their forum (no replies yet) and also sent an email to them via their web site. (The mail version where you has to enter your serial number).

Comments from Jon Ward will also be welcome!!

Some how I do not think there is a bug here....in the technical sense of a bug. The linker sees a function pointer with access tree links to the function pointer itself and notes that the value in the function pointer is in the same bank and this it feels that this is a legitimate "within same bank" function pointer. At least that is what the resulting code is doing. And in addition to that, when the function pointers are loaded with values outside the current bank the linker does a slapdash job of invoking the interbank transfer table entry.

I just have to figure out how to solve the problem in the long run. It does not look like the kind of thing that can be fixed with the OVERLAY parameter of the linker.

Mike Karas