Well after struggling with this most of saturday night, sunday and  
parts of Monday. I reflected back on all the
registers and was fortune enough to catch ctr set to a location in  
gldModifyTexture.

Hunting on the web gldModifyTexture finds nothing and I wasted quite a  
bit of time thinking/hunting for an openGL issue,
checking incoming data etc etc.

However a breakpoint on gldModifyTexture and a stack trace shows this

(gdb) break gldModifyTexture
Breakpoint 11 at 0x27e5974
(gdb) continue
warning: ppc_frame_chain_valid: stack pointer from 0xf0182b90 to 0x2601  
grows upward; assuming invalid
(gdb) bt
#0  0x027e5974 in gldModifyTexture ()
#1  0x02748e9c in gloStaticProfilingData ()
#2  0x026d5a94 in gloStaticProfilingData ()
#3  0x026d5b00 in gloStaticProfilingData ()
#4  0x029f7ce0 in 0_gpregs ()
#5  0x029ed7c4 in ffiCallAddressOfWithReturnType (fn=warning:  
ppc_frame_chain_valid: stack pointer from 0xf0182b90 to 0x2601 grows  
upward; assuming invalid

0, callType=0, typeSpec=0) at  
sqMacFFIPPC.c:468/Users/johnmci/Documents/SqueakFFI/
warning: ppc_frame_chain_valid: stack pointer from 0xf0182b90 to 0x2601  
grows upward; assuming invalid

You'll all note the  
sqMacFFIPPC.c:468/Users/johnmci/Documents/SqueakFFI/ which still makes  
it
my problem (I think, I've not seen the callee smalltalk code yet, I  
just found this gem 2 seconds ago).

Normally what would happen is the stack would unwind all the way back  
to ffiCallAddressOfWithReturnType and
place a data address in the stack frame and hose me with out much a  
clue. However now that I know ffiCallAddressOfWithReturnType is
broken, I'm now loading the MACFFIPlugin into the debugger and hope to  
resolve why and what the fix is later tonight.
--
======================================================================== 
===
John M. McIntosh <johnmci at smalltalkconsulting.com> 1-800-477-2659
Corporate Smalltalk Consulting Ltd.  http://www.smalltalkconsulting.com
======================================================================== 
===