Path: utzoo!utgpu!news-server.csri.toronto.edu!mailrus!wuarchive!zaphod.mps.ohio-state.edu!rpi!uupsi!sunic!tut!mk59200 From: mk59200@etana.tut.fi (Kolkka Markku Olavi) Newsgroups: comp.os.minix Subject: Re: 68070 chip. Message-ID: <12989@etana.tut.fi> Date: 9 May 90 14:30:20 GMT References: <5452562@hase_2.mbx.sub.org> Sender: News@tut.fi Lines: 35 From article <5452562@hase_2.mbx.sub.org>, by hase@hase_2.mbx.sub.org (Hartmut Semken): > In article jac@paul.rutgers.edu (Jonathan A. Chandross) writes: >> The 68070 is a high integration chip put out by Signetics (Philips). It >> has: a 68000, MMU, a serial port, and a few other things. > > True, but make that a 68010 and the "MMU" is called BMAC (Basic Memory No, the CPU core is more like a 68000, there is no VBR register, and it hasn't the 68010 instructions (like RTD). The exception stack frames have two formats: short frames are similar to 68010 $0 frames and long frames (format $F) are unique to 68070. > Access Controller), cause it is not really comparable to the 851 PMMU or > the like (but still: once upon a time there was an Altos machine running > Unix, suppoting up to 16 users and the work was done by a segmentation > MMU and a 68000... :-) Well, I have the "16/32-Bit Highly-integrated Microprocessor SC68070 User Manual Part 1 - Hardware" here in front of me, and it really calls the "MMU" MMU. It's compatible with some pre-68851 Signetics chip. I think you could implement a reasonable virtual memory system with it (at least as good as with a 80286). The on-chip two channel DMA controller is software compatible with 684[345]0 devices. > -- > Hartmut Semken hase@hase_2 (home) > hase_2 in der Testphase > Hartmut Semken, Lupsteiner Weg 67 1000 Berlin 37, Germany > "Sounds aweful" (Marvin, the paranoid android) -- Markku Kolkka mk59200@tut.fi