home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!olivea!gossip.pyramid.com!penguin.eng.pyramid.com!bms
- From: bms@penguin.eng.pyramid.com (Bruce Schlobohm)
- Newsgroups: comp.sys.mac.hardware
- Subject: Re: Motorola's 68060 for the Mac
- Message-ID: <1jpi0gINN1ui@gossip.pyramid.com>
- Date: 22 Jan 93 19:30:24 GMT
- References: <D2150056.o6gh64@erics.infoserv.com> <1993Jan22.054617.25775@bilby.cs.uwa.edu.au>
- Organization: Pyramid Technology, San Jose, Ca
- Lines: 38
- NNTP-Posting-Host: penguin.eng.pyramid.com
-
- In article <1993Jan22.054617.25775@bilby.cs.uwa.edu.au>, Quinn <quinn@cs.uwa.edu.au> writes:
- | In article <D2150056.o6gh64@erics.infoserv.com> Eric S. Smith,
- | erics@infoserv.com writes:
- | >The 68000 contains no MMU, nor is it possible to implement virtual memory
- | >with the 68000 - there are some instructions that could not be restarted
- | >following a page fault.
- |
- | The Lisa had an 68000 and implemented virtual memory but only for
- | code segments. Therefore I assume you can always manage to restart
- | instruction fetches on the 68000. I guess that putting data in
- | code segments was frowned upon (and didn't work).
-
- Eric is correct.
-
- The existance of an on chip MMU has nothing to do (in this case) with the
- ability to implement Virtual Memory (VM). The 68020 doesn't have an MMU
- *on chip* but many systems have been built (Mac and others) that have
- a VM implementation. The MMU just needs to be there, either in discrete
- parts or in the form of the 68851. The problem with VM and the 68000 is
- that the bus fault frame didn't allow for restarting of most instructions
- after a bus fault or page fault.
- One of the major differences between the 68000 and the 68010 was that
- the 68010 had a new bus fault frame that allowed for restarting
- instructions. A 68000 system can be designed (and was) that can provide
- the appearance of VM by doing full process swapping (instead of paging),
- and thus allowing for several programs combined to consume more memory than
- was physically present. But in my book, this isn't the same as virtual
- memory, in that a single process/program can't reference more memory
- than is physically present.
-
- I don't know how the Lisa was designed, nor if/how it implemented VM.
- And yes, data in code segments is usually a problem, most especially so
- with processors like the 68040 that have separate paths and caches for
- instructions and data.
-
- --
- bruce schlobohm
- bms@pyramid.com
-