Aucbvax.5466 fa.works utzoo!decvax!ucbvax!works Sun Dec 13 17:06:10 1981 WorkS Digest V1 #44 >From JSol@RUTGERS Sun Dec 13 17:04:42 1981 Works Digest Monday, 14 Dec 1981 Volume 1 : Issue 44 Today's Topics: Limiting 68000 To Permit Demand Paging (Very Short Digest) ---------------------------------------------------------------------- Date: 9 Dec 1981 20:06:22-PST From: decvax!watmath!idallen at Berkeley To: decvax!ucbvax!works@Berkeley Subject: Limiting 68000 to permit demand paging. Re: Ian Merritt's comment on a "real implementation" of paging. The observations of hardware deficiencies published in this digest lead me to believe that a single 68000 can not be involved in a non-restartable instruction when a page fault occurs, since there is no way to preserve the internal state of the chip, service the page fault, and resume. I am led to think that a "real implementation" must involve more than one processor, essentially leaving the 68000 stuck in mid-instruction while the second processor brings in the missing page(s). Some comments seemed to suggest that, even with this dual-processor kludge, it was still necessary to restrict the use of certain instructions. This is the point I am unsure of. I can not see this restriction as being a good thing if it means that one must avoid auto[increment|decrement] modes. I guess I am asking, What precisely is a "real implementation"? -IAN! ------------------------------ End of WorkS Digest ******************* ------- ----------------------------------------------------------------- gopher://quux.org/ conversion by John Goerzen of http://communication.ucsd.edu/A-News/ This Usenet Oldnews Archive article may be copied and distributed freely, provided: 1. There is no money collected for the text(s) of the articles. 2. The following notice remains appended to each copy: The Usenet Oldnews Archive: Compilation Copyright (C) 1981, 1996 Bruce Jones, Henry Spencer, David Wiseman.