In message <BAY17-F17C3D220CEF5F1DDAB4767C36E0@phx.gbl>, "Jared Momose" writes:
>Kamal,
>
>I have been thinking about this problem for a while, considering the 
>Blackfin line of DSPs from Analog Devices. These processors are similar to 
>what you describe in that they have address protection but not translation. 
>Without translation, you are doomed some nasty page swapping madness. You 
>might be able to minimize it by attempting to locate your binaries to 
>different addresses such that concurrently running binaries occupy different 
>sections of your address space, but you still have severe limitations 
>(multiple instances of the same programe, i.e. getty, anything multiuser, 
>fork needs to be reworked?!). Personally, I have concluded that although it 
>may be possible, the result would be terribly inefficient and not very 
>useful for anything academic or commercial.
>

Let me dredge up some ancient technology....  OS/360, vintage 1965 or 
so, had memory protection but no virtual memory of any sort.  However, 
executables were relocated at program load time.  This imposes a 
requirement on both the linker and the program loader, of course.  
Beyond that, NetBSD may have more sophisticated demands on the MMU; I'm 
not sure, and I'll leave that to others.

		--Prof. Steven M. Bellovin, http://www.cs.columbia.edu/~smb