[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
An interesting implementation of MC68000 demand paging...
-----BEGIN PGP SIGNED MESSAGE-----
I was just browsing through some of the m68k source in the uClinux
kernel and started researching the Apollo series after looking at some
of the support code that looked kind of interesting.
The early models where based on the MC68000 processor and did, in fact,
implement demand paging. How they did it was very cleaver. They actually
ran two 68K processors, one was the main processor and one was a sort of
MMU assitant such that when the MMU hardware detected that a page would
fault, it would force the main 68K processor into wait-state while the
second processor handled the fault and swapped in the page or whatever
it needed to do.
I though this was extreemly cleaver, and goes back to something Andrew
was saying about the 68K MMU that I've been working on that if we could
keep it sort of generic, it might be able to be used on other processors
on the S100 bus. Well, using an approach similar to this might make
demand paging possible with some of the processor that otherwise would
not be able to support demand paging.
Anyway just thought I'd share that very interesting tidbit.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.10 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/
-----END PGP SIGNATURE-----