From: Tom Van Vleck Newsgroups: alt.folklore.computers Subject: Memory protection (was Re: Powerful Technique 5604) Date: Sun, 21 Jan 2001 17:26:56 -0500 Organization: Multicians Lines: 29 Message-ID: <3A6B6230.4F856ADC@multicians.org> References: <3a6881e9.1@news2.tm.net.my> <3A69B58D.ED6B9A52@uchicago.edu> <94eia4$gct$5@bob.news.rcn.net> <3A6B3448.8B5E776E@uchicago.edu> <94fei7$9i1$1@bob.news.rcn.net> NNTP-Posting-Host: 04.36.29.37 Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Server-Date: 21 Jan 2001 22:25:10 GMT X-Mailer: Mozilla 4.76 [en] (Win98; U) X-Accept-Language: en,pdf Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news.ip-plus.net!news.tesion.net!news.belwue.de!news-stu1.dfn.de!news-mue1.dfn.de!news-was.dfn.de!news-spur1.maxwell.syr.edu!news.maxwell.syr.edu!newsfeed.direct.ca!look.ca!newsfeed1.earthlink.net!newsfeed2.earthlink.net!newsfeed.earthlink.net!news.mindspring.net!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:73942 Eric Fischer wrote, describing memory protection: > I think the first actual implementation of the idea was in 1961 in > MIT's Compatible Time Sharing System, which isolated the OS from user > programs by keeping them in separate memory banks, and programs from > each other by keeping only one in memory at a time. Some minor points. 1. The 7090 had a "transfer trapping mode" that limited the running program to 32K and trapped transfers into the other half of memory. This was documented in the regular manual. I was told at the time (1962 probably) that this mode was there to support the 709 emulator. The set of CPU mods to the 7094 to support CTSS included ideas from this mode, but the CTSS mods were an RPQ (request price quotation, that is, a one-off engineering special). 2. CTSS kept location 0 of one program at a time in memory. The modified 7094 had an address limit register that trapped references greater than the limit. If you ran a big program and then a little one, only enough of the big program to make room for the little one was swapped to drum. The code that kept track of the pieces of various users' core images on drum and reassembled them when it came time to run was called the "onion skin" algorithm. I have always thought that it was very sad that the initial PC operating systems for the Mac and IBM PC lacked memory protection. Many hours have been wasted. It's not as if it was a new idea. ###### From: pjb@imaginet.fr (Pascal J. Bourguignon) Newsgroups: alt.folklore.computers Subject: Re: Memory protection (was Re: Powerful Technique 5604) Date: 22 Jan 2001 01:59:51 +0100 Lines: 27 Sender: pascal@triton.local.net Message-ID: <873decqu48.fsf@triton.local.net> References: <3a6881e9.1@news2.tm.net.my> <3A69B58D.ED6B9A52@uchicago.edu> <94eia4$gct$5@bob.news.rcn.net> <3A6B3448.8B5E776E@uchicago.edu> <94fei7$9i1$1@bob.news.rcn.net> <3A6B6230.4F856ADC@multicians.org> NNTP-Posting-Host: 193-153-126-22.uc.nombres.ttd.es (193.153.126.22) X-Trace: fu-berlin.de 980125203 14233161 193.153.126.22 (16 [41911]) X-Newsreader: Gnus v5.7/Emacs 20.7 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news.ip-plus.net!Amsterdam.Infonet!News.Amsterdam.UnisourceCS!skynet.be!fu-berlin.de!uni-berlin.de!193-153-126-22.uc.nombres.ttd.ES!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:73981 Tom Van Vleck writes: > Eric Fischer wrote, describing memory protection: > > I think the first actual implementation of the idea was in 1961 in > > MIT's Compatible Time Sharing System, which isolated the OS from user > > programs by keeping them in separate memory banks, and programs from > > each other by keeping only one in memory at a time. > [...] > I have always thought that it was very sad that the initial > PC operating systems for the Mac and IBM PC lacked memory > protection. Many hours have been wasted. It's not as if > it was a new idea. Right. And a fundamental flaw that won't have been corrected until next March, when MacOSX will at least be delivered, for the Mac; and that has been corrected on PCs with MMU as soon as Linux and other unix used it ten years ago. -- __Pascal Bourguignon__ PGP Key ID: 0xEF5E9966 (o_ mailto:pjb@imaginet.fr PGP fingerprint: 00 F5 7B DB CA 51 8A AD 04 5B //\ http://informatimago.free.fr/index 6C DE 32 60 16 8E EF 5E 99 66 V_/ () Join the ASCII ribbon campaign against html email and Microsoft attachments. /\ Software patents are endangering the computer industry all around the world. Join the LPF: http://lpf.ai.mit.edu/ http://petition.eurolinux.org/ ###### From: Eric Fischer Newsgroups: alt.folklore.computers Subject: Re: Memory protection (was Re: Powerful Technique 5604) Date: 22 Jan 2001 02:44:20 GMT Lines: 39 Message-ID: <94g6q4$6v0$1@bob.news.rcn.net> References: <3a6881e9.1@news2.tm.net.my> <3A6B3448.8B5E776E@uchicago.edu> <94fei7$9i1$1@bob.news.rcn.net> <3A6B6230.4F856ADC@multicians.org> X-Trace: UmFuZG9tSVYDlImpgTpzbfgPm1Od1zilCwbncQpju6a9pfYHxMGuMQcM3gpvhJ5E X-Complaints-To: abuse@rcn.com NNTP-Posting-Date: 22 Jan 2001 02:44:20 GMT X-Newsreader: trn 4.0-test72 (19 April 1999) Originator: enf@enteract.com (Eric Fischer) Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news.ip-plus.net!news.tesion.net!news.belwue.de!news-stu1.dfn.de!news-koe1.dfn.de!news-was.dfn.de!news-spur1.maxwell.syr.edu!news.maxwell.syr.edu!europa.netcrusader.net!207.172.3.44!feed2.news.rcn.net!feed1.news.rcn.net!rcn!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:73925 Tom Van Vleck describes the way the Compatible Time Sharing System handled memory allocation: > 2. CTSS kept location 0 of one program at a time in memory. > The modified 7094 had an address limit register that trapped > references greater than the limit. If you ran a big program > and then a little one, only enough of the big program to > make room for the little one was swapped to drum. The code > that kept track of the pieces of various users' core images > on drum and reassembled them when it came time to run was > called the "onion skin" algorithm. Thanks for the clarification. I've just reread "An Experimental Time Sharing System," and it looks like it was only the original version for the IBM 709 that swapped entire programs out the way I had claimed: To avoid memory allocation clashes, protect users from one another, and simplify the initial 709 system organization, only one user was kept in core memory at a time. However, with the special memory protection and relocation features of the 7090, more sophisticated storage allocation procedures are being implemented. The paper also makes it clear that, contrary to what I wrote, in the 709 version, the OS and user programs were not kept in separate address spaces -- instead, the OS occupied the bottom 5KW of memory and the user program the remaining 27KW. > I have always thought that it was very sad that the initial > PC operating systems for the Mac and IBM PC lacked memory > protection. Many hours have been wasted. It's not as if > it was a new idea. Agreed. It seems to be one of those things (like hierarchical file systems) that have to be reinvented for each new generation of hardware. Do palmtops have protected memory yet? eric ###### Sender: root@fdn.com Newsgroups: alt.folklore.computers Subject: Re: Memory protection (was Re: Powerful Technique 5604) References: <3a6881e9.1@news2.tm.net.my> <3A6B3448.8B5E776E@uchicago.edu> <94fei7$9i1$1@bob.news.rcn.net> <3A6B6230.4F856ADC@multicians.org> <94g6q4$6v0$1@bob.news.rcn.net> From: gcash Message-ID: Lines: 20 X-Newsreader: Gnus v5.5/Emacs 20.3 Date: 21 Jan 2001 23:28:58 -0500 NNTP-Posting-Host: 216.199.46.32 X-Trace: news1.fdn.com 980137728 216.199.46.32 (Sun, 21 Jan 2001 23:28:48 EST) NNTP-Posting-Date: Sun, 21 Jan 2001 23:28:48 EST Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news.ip-plus.net!news.tesion.net!news.belwue.de!news-stu1.dfn.de!news-mue1.dfn.de!news-was.dfn.de!news-spur1.maxwell.syr.edu!news.maxwell.syr.edu!newsfeed.skycache.com!Cidera!news1.fdn.com!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:74061 Eric Fischer writes: > Do palmtops have protected memory yet? Palm OS ones may. The O'Reilly "Palm Programming" book says memory is "hardware write-protected" and the Motorola MC68328 & MC68EZ328 Dragonball processors are 68000 derivatives. The Dragonball site at http://www.motorola.com/SPS/WIRELESS/pda/index.html, didn't mention an MMU, and none of the PDF files were readable. I wrote some quickie programs and accessed some intentionally null pointers and other memeory faux pas, and it seemed to catch it and pop up a fatal error dialog. -gc -- Missed the braking marker, missed the turn-in, missed the apex, and now fervently hoping to miss the tree ###### From: Adam Sampson Newsgroups: alt.folklore.computers Subject: Re: Memory protection (was Re: Powerful Technique 5604) Date: 22 Jan 2001 22:36:45 +0000 Organization: The Campaign For The Writing Of "a lot" As Two Words Lines: 20 Sender: azz@cartman.azz.net Message-ID: <871ytvusci.fsf@cartman.azz.net> References: <3a6881e9.1@news2.tm.net.my> <3A6B3448.8B5E776E@uchicago.edu> <94fei7$9i1$1@bob.news.rcn.net> <3A6B6230.4F856ADC@multicians.org> <94g6q4$6v0$1@bob.news.rcn.net> NNTP-Posting-Host: stue407.ukc.ac.uk Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII X-Trace: athena.ukc.ac.uk 980203235 20597 129.12.228.7 (22 Jan 2001 22:40:35 GMT) X-Complaints-To: news@ukc.ac.uk NNTP-Posting-Date: Mon, 22 Jan 2001 22:40:35 +0000 (UTC) User-Agent: Gnus/5.0807 (Gnus v5.8.7) XEmacs/21.1 (Carlsbad Caverns) Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newsfeeds.belnet.be!news.belnet.be!btnet-peer1!btnet-peer0!btnet-peer!btnet!news2.vas-net.net!server2.netnews.ja.net!athena.ukc.ac.uk!cartman.us-lot.org!nobody Xref: chonsp.franklin.ch alt.folklore.computers:73985 gcash writes: > > Do palmtops have protected memory yet? > > Palm OS ones may. The O'Reilly "Palm Programming" book says memory is > "hardware write-protected" and the Motorola MC68328 & MC68EZ328 Dragonball > processors are 68000 derivatives. Well, not terribly. The 68000 will complain if you access unaligned memory, so the PalmOS will hand you unaligned pointers unless you've unlocked the bit of memory you want first. There's no real MMU, though. Several other models of palmtop (with somewhat more powerful processors, like the ARM) have protected memory, though. -- Adam Sampson azz@gnu.org ###### From: cbh@REMOVE_THIS.teabag.fsnet.co.uk (Chris Hedley) Newsgroups: alt.folklore.computers Subject: Re: Memory protection (was Re: Powerful Technique 5604) Date: 22 Jan 2001 23:34:15 GMT Organization: teabag Message-ID: <94ig1n$but$6@teabag.cbhnet> References: <3a6881e9.1@news2.tm.net.my> <3A6B3448.8B5E776E@uchicago.edu> <94fei7$9i1$1@bob.news.rcn.net> <3A6B6230.4F856ADC@multicians.org> <94g6q4$6v0$1@bob.news.rcn.net> NNTP-Posting-Host: teabag.cbhnet X-NNTP-Posting-Host: teabag.demon.co.uk:193.237.4.110 X-Trace: news.demon.co.uk 980206459 nnrp-10:2987 NO-IDENT teabag.demon.co.uk:193.237.4.110 X-Complaints-To: abuse@demon.net Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii NNTP-Posting-Date: 22 Jan 2001 23:34:15 GMT X-Newsreader: knews 1.0b.0 Lines: 9 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news.ip-plus.net!news.tesion.net!news.belwue.de!news-stu1.dfn.de!news-mue1.dfn.de!news-nue1.dfn.de!uni-erlangen.de!newsfeeds.belnet.be!news.belnet.be!btnet-peer1!btnet-peer0!btnet!dispose.news.demon.net!news.demon.co.uk!demon!teabag.demon.co.uk!teabag.cbhnet!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:73913 In article , gcash writes: > Palm OS ones may. The O'Reilly "Palm Programming" book says memory is > "hardware write-protected" and the Motorola MC68328 & MC68EZ328 Dragonball > processors are 68000 derivatives. I thought that the DragonBall was an ARM derivitive, but I could be wrong. Chris. ###### Sender: eric@ruckus.brouhaha.com From: Eric Smith Newsgroups: alt.folklore.computers Subject: Re: Memory protection (was Re: Powerful Technique 5604) References: <3a6881e9.1@news2.tm.net.my> <3A69B58D.ED6B9A52@uchicago.edu> <94eia4$gct$5@bob.news.rcn.net> <3A6B3448.8B5E776E@uchicago.edu> <94fei7$9i1$1@bob.news.rcn.net> <3A6B6230.4F856ADC@multicians.org> <873decqu48.fsf@triton.local.net> X-Disclaimer: Everything I write is false. Organization: Eric Conspiracy Secret Labs X-Eric-Conspiracy: There is no conspiracy. Date: 23 Jan 2001 17:55:18 -0800 Message-ID: Lines: 14 User-Agent: Gnus/5.0807 (Gnus v5.8.7) Emacs/20.7 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii NNTP-Posting-Host: ruckus.brouhaha.com X-Trace: 23 Jan 2001 17:56:46 -0800, ruckus.brouhaha.com Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news.ip-plus.net!news.tesion.net!news.belwue.de!news-stu1.dfn.de!news-koe1.dfn.de!news-was.dfn.de!news-spur1.maxwell.syr.edu!news.maxwell.syr.edu!telocity-west!TELOCITY!enews.sgi.com!news.sgi.com!news.spies.com!ruckus.brouhaha.com Xref: chonsp.franklin.ch alt.folklore.computers:74112 Tom Van Vleck writes: > I have always thought that it was very sad that the initial > PC operating systems for the Mac and IBM PC lacked memory > protection. Many hours have been wasted. It's not as if > it was a new idea. pjb@imaginet.fr (Pascal J. Bourguignon) writes: > Right. And a fundamental flaw that won't have been corrected until > next March, when MacOSX will at least be delivered, for the Mac; and > that has been corrected on PCs with MMU as soon as Linux and other > unix used it ten years ago. And you don't count A/UX on the Mac? It seemed to support protected memory just fine more than ten years ago. ###### From: rsteiner@isis.visi.com (Richard C. Steiner) Newsgroups: alt.folklore.computers Subject: Re: Memory protection (was Re: Powerful Technique 5604) References: <3a6881e9.1@news2.tm.net.my> <3A69B58D.ED6B9A52@uchicago.edu> <94eia4$gct$5@bob.news.rcn.net> <3A6B3448.8B5E776E@uchicago.edu> <94fei7$9i1$1@bob.news.rcn.net> <3A6B6230.4F856ADC@multicians.org> <873decqu48.fsf@triton.local.net> Organization: Vector Internet Services, Inc. Message-ID: User-Agent: slrn/0.9.5.4 (UNIX) Lines: 19 Date: Wed, 24 Jan 2001 18:16:24 GMT NNTP-Posting-Host: 209.98.98.8 X-Complaints-To: abuse@visi.com X-Trace: ruti.visi.com 980360184 209.98.98.8 (Wed, 24 Jan 2001 12:16:24 CST) NNTP-Posting-Date: Wed, 24 Jan 2001 12:16:24 CST Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!isdnet!newsfeed.icl.net!news.maxwell.syr.edu!hermes.visi.com!news-out.visi.com!ruti.visi.com!rsteiner Xref: chonsp.franklin.ch alt.folklore.computers:74115 In article , Eric Smith wrote: > >pjb@imaginet.fr (Pascal J. Bourguignon) writes: >> Right. And a fundamental flaw that won't have been corrected until >> next March, when MacOSX will at least be delivered, for the Mac; and >> that has been corrected on PCs with MMU as soon as Linux and other >> unix used it ten years ago. > >And you don't count A/UX on the Mac? It seemed to support protected >memory just fine more than ten years ago. Not to mention the fact that many Linux flavors (Yellow Dog, Linux PPC, etc.) exist for the Mac as well, as do OpenBSD and NetBSD. -- -Rich Steiner >>>---> rsteiner@visi.com >>>---> Eden Prairie, MN Written online using slrn 0.9.5.4! The Theorem Theorem: If If, Then Then. ###### Newsgroups: alt.folklore.computers From: ehrice@his.com (Edward Rice) Subject: Re: Memory protection (was Re: Powerful Technique 5604) Message-ID: Organization: NDS Date: Mon, 12 Feb 2001 17:41:25 -0500 References: <3a6881e9.1@news2.tm.net.my> <3A69B58D.ED6B9A52@uchicago.edu> <94eia4$gct$5@bob.news.rcn.net> <3A6B3448.8B5E776E@uchicago.edu> <94fei7$9i1$1@bob.news.rcn.net> <3A6B6230.4F856ADC@multicians.org> NNTP-Posting-Host: pm2v-65.his.com X-Trace: 12 Feb 2001 17:41:32 -0500, pm2v-65.his.com Lines: 61 X-Authenticated-User: ehrice Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news-fra1.dfn.de!news.tele.dk!205.231.236.10!newspeer.monmouth.com!news.lightlink.com!news1.his.com!vienna7.his.com!user Xref: chonsp.franklin.ch alt.folklore.computers:75423 In article <3A6B6230.4F856ADC@multicians.org>, Tom Van Vleck wrote: > Some minor points. > 1. The 7090 had a "transfer trapping mode" that limited the > running program to 32K and trapped transfers into the other > half of memory. This was documented in the regular manual. > I was told at the time (1962 probably) that this mode was > there to support the 709 emulator. The set of CPU mods > to the 7094 to support CTSS included ideas from this mode, > but the CTSS mods were an RPQ (request price quotation, > that is, a one-off engineering special). Sort-a, but I think amplification will make that "more correct." Transfers didn't have to cross to another 32KW bank to be trapped. And it looks to me like the mode would also be excellent for debugging as well as catching inter-bank transfers. (Could execution just "roll off" the top of one bank into the other, with no transfer instruction involved, Tom?) Here's what the 7094 Principles of Operation says about Transfer Trap Mode, p. 7 of A22-6703-4: _Transfer Trap Mode_ The computer can be operated in a special transfer trap mode. The major use of this mode is in program testing. Operation in the trap mode per- mits the program to run at normal speed with interruptions of normal operation only at transfer points. At such points the location of the last sequential instruction is saved, and a transfer of control is made to a fixed locatuion. Beginning at this fixed location, a special moni- toring program may aid the programmer in control of his store program even in the event of an incorrect transfer. When the computer is operating in this mode, the location of each transfer instruction replaces the address part of core storage location 00000 before the instruction is executed. All functions of the transfer instruction (index register modification, testing, etc.) are performed normally. However, program control is transferred to location 00001 for all unconditional transfer instructions and conditional transfer instruc- tions for which the transfer conditions are met. Two instructions, trap transfer (TTR) and enter storage nullification and transfer (ESNT), are immune to the trapping mode. Their locations are never stored in location 00000, and control is always transferred to the location specified y the address of these instructions. The instructions ENTER TRAPPING MODE and LEAVE TRAPPING MODE are used to enter or leave this special mode. Depression of the clear or reset keys on the console also causes the computer to exit from this mode. It looks to me, btw, like ESNT is the dispatch instruction which the system would execute, just after setting trap mode on, in order to return control to the user. This would work fine in the "enhanced 7094" with two banks, or the normal one with just 32KW. -- Edward "off to explore www.ibsys.com" R -- Bring Dejanews back to life! See the petition at http://www2.petitiononline.com/dejanews/petition.html and SIGN IT, please. Help bring back a valuable Usenet resource! DejaNews was the net's recorded history, and we'd like to see it available online again.