)]}' { "log": [ { "commit": "466126613387959c0206a8cb13b2e94382308633", "tree": "cc6b7c40c3a16a360fdd2dc710a068439bfd71f5", "parents": [ "ab95eb709f42d5dabcda86b31d35429f25a4934b" ], "author": { "name": "Linus Torvalds", "email": "Linus.Torvalds@helsinki.fi", "time": "Mon Mar 14 12:51:16 1994 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:36 2007 -0400" }, "message": "Linux 1.0---A better UNIX than Windows NT\n\nSummary: Linux 1.0 released\nKeywords: Linux Kernel 1.0 Academy Awards\nX-Moderator-Added-Keywords: universe, end of\n\nFinally, here it is. Almost on time (being just two years late is\npeanuts in the OS industry), and better than ever:\n\n\tLinux kernel release 1.0\n\nThis release has no new major features compared to the pl15 kernels, but\ncontains lots and lots of bugfixes: all the major ones are gone, the\nsmaller ones are hidden better. Hopefully there are no major new ones.\n\nThe Linux kernel can be found as source on most of the Linux ftp-sites\nunder the names\n\n\tlinux-1.0.tar.gz\t\t(full source)\n\tlinux-1.0.patch.pl15.gz\t\t(patch against linux-0.99pl15)\n\tlinux-1.0.patch.alpha.gz\t(patch from linux-pre-1.0)\n\nit should be available at least at the sites\n\n\tftp.funet.fi:\n\t\tpub/OS/Linux/PEOPLE/Linus\t(now)\n\tsunsite.unc.ed:\n\t\tpub/Linux/Incoming\t\t(now)\n\t\tpub/Linux/kernel\t\t(soon)\n\ttsx-11.mit.edu:\n\t\tpub/linux/sources/system\t(soon)\n\tftp.cs.helsinki.fi:\n\t\tpub/Software/Linux/Kernel\t(now)\n\nThis release finally moves Linux out of Beta status and is meant as a\nbase for distributions to build on. It will neither change Linux\u0027\nstatus as FreeWare under the GPL, nor will it mean the end of\ndevelopment on Linux. In fact many new features where held back for\nlater releases so that 1.0 could become a well tested and hopefully\nstable release.\n\nThe Linux kernel wouldn\u0027t be where it is today without the help of lots\nof people: the kernel developers, the people who did user-level programs\nmaking linux useful, and the brave and foolhardy people who risked their\nharddisks and sanity to test it all out. My thanks to you all.\n(Editorial note: if you think this sounds too much like the Academy\nAwards ceremony, just skip this: it\u0027s not getting any better.)\n\nThanks to people like Aaron Kushner, Danny ter Haar and the authors of\nthe AnwenderHandbuch (and others) who have helped me with hardware or\nmonetary donations (and to the Oxford Beer Trolls and others who took\ncare of the drinkware). And thanks to Dirk, who helped me write this\nannouncement despite my lazyness (\"hey, it\u0027s just another release, who\nneeds an announcement anyway?\").\n\nTo make a long and boring story a bit shorter and boring, here is at\nleast a partial list of people who have been helping make Linux what it\nis today. Thanks to you all,\n\n\tKrishna Balasubramanian \u003cbalasub@cis.ohio-state.edu\u003e\n\tArindam Banerji \u003caxb@cse.nd.edu\u003e\n\tPeter Bauer \u003c100136.3530@compuserve.com\u003e\n\tFred Baumgarten \u003cdc6iq@insu1.etec.uni-karlsruhe.de\u003e\n\tDonald Becker \u003cbecker@super.org\u003e\n\tStephen R. van den Berg \u003cberg@pool.informatik.rwth-aachen.de\u003e\n\tHennus Bergman \u003chennus@sky.nl.mugnet.org\u003e\n\tRoss Biro \u003cbir7@leland.Stanford.Edu\u003e\n\tBill Bogstad \u003cbogstad@cs.jhu.edu\u003e\n\tJohn Boyd \u003cboyd@cis.ohio-state.edu\u003e\n\tAndries Brouwer \u003caeb@cwi.nl\u003e\n\tRemy Card \u003cRemy.Card@masi.ibp.fr\u003e\n\tEd Carp \u003cecarp@netcom.com\u003e\n\tRaymond Chen \u003craymondc@microsoft.com\u003e\n\tAlan Cox \u003ciiitac@pyr.swan.ac.uk\u003e\n\tLaurence Culhane \u003cloz@holmes.demon.co.uk\u003e\n\tWayne Davison \u003cdavison@borland.com\u003e\n\tThomas Dunbar \u003ctdunbar@vtaix.cc.vt.edu\u003e\n\tTorsten Duwe \u003cTorsten.Duwe@informatik.uni-erlangen.de\u003e\n\tDrew Eckhardt \u003cdrew@cs.Colorado.EDU\u003e\n\tBjorn Ekwall \u003cbj0rn@blox.se\u003e\n\tDoug Evans \u003cdje@cygnus.com\u003e\n\tRik Faith \u003cfaith@cs.unc.edu\u003e\n\tJuergen Fischer \u003cfischer@server.et-inf.fho-emden.de\u003e\n\tJeremy Fitzhardinge \u003cjeremy@sw.oz.au\u003e\n\tRalf Flaxa \u003crfflaxa@immd4.informatik.uni-erlangen.de\u003e\n\tNigel Gamble \u003cnigel%gamble.uucp@gate.net\u003e\n\tPhilip Gladstone \u003cphilipg@onsett.com\u003e\n\tBruno Haible \u003chaible@ma2s2.mathematik.uni-karlsruhe.de\u003e\n\tAndrew Haylett \u003cajh@gec-mrc.co.uk\u003e\n\tDirk Hohndel \u003chohndel@informatik.uni-wuerzburg.de\u003e\n\tNick Holloway \u003calfie@dcs.warwick.ac.uk\u003e\n\tRon Holt \u003cron@novell.com\u003e\n\tRob W. W. Hooft \u003chooft@EMBL-Heidelberg.DE\u003e\n\tMichael K. Johnson \u003cjohnsonm@sunsite.unc.edu\u003e\n\tFred N. van Kempen \u003cwaltje@uwalt.nl.mugnet.org\u003e\n\tOlaf Kirch \u003cokir@monad.swb.de\u003e\n\tIan Kluft \u003cikluft@thunder.sbay.org\u003e\n\tRudolf Koenig \u003crfkoenig@immd4.informatik.uni-erlangen.de\u003e\n\tBas Laarhoven \u003cbas@vimec.nl\u003e\n\tWarner Losh \u003cimp@boulder.parcplace.com\u003e\n\tH.J. Lu \u003chjl@nynexst.com\u003e\n\tTuomas J. Lukka \u003cTuomas.Lukka@Helsinki.FI\u003e\n\tKai M\"akisara \u003cKai.Makisara@vtt.fi\u003e\n\tPat Mackinlay \u003cpat@it.com.au\u003e\n\tJohn A. Martin \u003cjmartin@csc.com\u003e\n\tBradley McLean \u003cbrad@bradpc.gaylord.com\u003e\n\tCraig Metz \u003ccmetz@tjhsst.edu\u003e\n\tWilliam (Bill) Metzenthen \u003cbillm@vaxc.cc.monash.edu.au\u003e\n\tRick Miller \u003crick@discus.mil.wi.us\u003e\n\tCorey Minyard \u003cminyard@wf-rch.cirr.com\u003e\n\tEberhard Moenkeberg \u003cemoenke@gwdg.de\u003e\n\tIan A. Murdock \u003cimurdock@shell.portal.com\u003e\n\tJohan Myreen \u003cjem@vipunen.hut.fi\u003e\n\tStefan Probst \u003csnprobst@immd4.informatik.uni-erlangen.de\u003e\n\tDaniel Quinlan \u003cquinlan@bucknell.edu\u003e\n\tFlorian La Roche \u003crzsfl@rz.uni-sb.de\u003e\n\tRobert Sanders \u003cgt8134b@prism.gatech.edu\u003e\n\tPeter De Schrijver \u003cstud11@cc4.kuleuven.ac.be\u003e\n\tDarren Senn \u003csinster@scintilla.santa-clara.ca.us\u003e\n\tChris Smith \u003ccsmith@convex.com\u003e\n\tDrew Sullivan \u003cdrew@lethe.north.net\u003e\n\tTommy Thorn \u003cTommy.Thorn@daimi.aau.dk\u003e\n\tJon Tombs \u003cjon@gtex02.us.es\u003e\n\tTheodore Ts\u0027o \u003ctytso@mit.edu\u003e\n\tSimmule Turner \u003csimmy@digex.com\u003e\n\tStephen Tweedie \u003csct@dcs.ed.ac.uk\u003e\n\tThomas Uhl \u003cuhl@sun1.rz.fh-heilbronn.de\u003e\n\tJuergen Weigert \u003cjnweiger@immd4.informatik.uni-erlangen.de\u003e\n\tMatt Welsh \u003cmdw@sunsite.unc.edu\u003e\n\tMarco van Wieringen \u003cmvw@mercury.mcs.nl.mugnet.org\u003e\n\tStephen D. Williams \u003csdw@lig.net\u003e\n\tG\\\"unter Windau \u003cgunter@mbfys.kun.nl\u003e\n\tLars Wirzenius \u003clars.wirzenius@helsinki.fi\u003e\n\tRoger E. Wolff \u003cwolff@dutecai.et.tudelft.nl\u003e\n\tFrank Xia \u003cqx@math.columbia.edu\u003e\n\tEric Youngdale \u003ceric@tantalus.nrl.navy.mil\u003e\n\tOrest Zborowski \u003corestz@microsoft.com\u003e\n\nA more detailed list with contact and description information can be\nfound in the CREDITS file that accompanies the kernel sources.\n" }, { "commit": "ab95eb709f42d5dabcda86b31d35429f25a4934b", "tree": "3daace3103106c322d222f261d4c784920e5e275", "parents": [ "73907e3b1b8e4719e039eb6b46b2f376e952356c" ], "author": { "name": "Linus Torvalds", "email": "torvalds@cc.helsinki.fi", "time": "Sun Mar 13 01:55:41 1994 +0200" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:36 2007 -0400" }, "message": "ALPHA-1.0\n" }, { "commit": "73907e3b1b8e4719e039eb6b46b2f376e952356c", "tree": "62dbdca66dc899710d49882ca023a950a33fa82c", "parents": [ "61bfd4ff199f5fb360c106073461e550e2ed5f0c" ], "author": { "name": "Linus Torvalds", "email": "torvalds@cc.helsinki.fi", "time": "Sun Mar 06 16:14:51 1994 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:35 2007 -0400" }, "message": "Linux 1.0-pre1\n" }, { "commit": "61bfd4ff199f5fb360c106073461e550e2ed5f0c", "tree": "9c87d7e5b63dd26cc3af920929cb656ef82e5523", "parents": [ "24497521076f7ce46c176d2dad018744e6f8cedd" ], "author": { "name": "Linus Torvalds", "email": "torvalds@cc.helsinki.fi", "time": "Wed Mar 02 17:35:20 1994 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:35 2007 -0400" }, "message": "ALPHA-pl15j\n" }, { "commit": "24497521076f7ce46c176d2dad018744e6f8cedd", "tree": "3184dc1735522b821a9d6bed05f99e070520b840", "parents": [ "eaee9b0646fdb4acad111941255378b0e882ffa2" ], "author": { "name": "Linus Torvalds", "email": "torvalds@cc.helsinki.fi", "time": "Tue Mar 01 17:29:05 1994 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:35 2007 -0400" }, "message": "ALPHA-pl15i\n" }, { "commit": "eaee9b0646fdb4acad111941255378b0e882ffa2", "tree": "cc4fab413df296412ade8f5fc041ad7b9435382a", "parents": [ "ece0168d6056dde89594a5de89133ef9b2b7097c" ], "author": { "name": "Linus Torvalds", "email": "torvalds@cc.helsinki.fi", "time": "Tue Feb 22 15:42:37 1994 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:34 2007 -0400" }, "message": "ALPHA-pl15h\n" }, { "commit": "ece0168d6056dde89594a5de89133ef9b2b7097c", "tree": "b2743632e1b80fe3f6334b691d835684fe0f2516", "parents": [ "42f89c5ded09dc700fbabdd6008c7781e99762e2" ], "author": { "name": "Linus Torvalds", "email": "torvalds@cc.helsinki.fi", "time": "Mon Feb 21 18:41:53 1994 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:34 2007 -0400" }, "message": "ALPHA-pl15g\n" }, { "commit": "42f89c5ded09dc700fbabdd6008c7781e99762e2", "tree": "94f70c69ae53701ff5d040f1ed34a1f257b17b13", "parents": [ "2fd717bbd7af8cb1dfd2f00df621b945e55e16fb" ], "author": { "name": "Linus Torvalds", "email": "torvalds@cc.helsinki.fi", "time": "Fri Feb 18 14:54:02 1994 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:33 2007 -0400" }, "message": "ALPHA-pl15f\n" }, { "commit": "2fd717bbd7af8cb1dfd2f00df621b945e55e16fb", "tree": "b590e5a3ae50b4eecfbd913c1092326fbfe069b2", "parents": [ "61119d8da02481d9744f37f62cabeec5aecac3ba" ], "author": { "name": "Linus Torvalds", "email": "torvalds@cc.helsinki.fi", "time": "Wed Feb 16 11:07:57 1994 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:32 2007 -0400" }, "message": "ALPHA-pl15e\n" }, { "commit": "61119d8da02481d9744f37f62cabeec5aecac3ba", "tree": "22bf3689b1b5559861c4c3bc5ca480c65937aa8e", "parents": [ "d9cc76127bcc137e3214b9166c439e02d2060cda" ], "author": { "name": "Linus Torvalds", "email": "torvalds@cc.helsinki.fi", "time": "Mon Feb 14 16:51:45 1994 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:32 2007 -0400" }, "message": "ALPHA-pl15d\n" }, { "commit": "d9cc76127bcc137e3214b9166c439e02d2060cda", "tree": "8825a01f3f8cdd036e4bb9737236e898ca5eef2e", "parents": [ "252058a6e7d838d1cab9512f6240f9e6b5b8b751" ], "author": { "name": "Linus Torvalds", "email": "torvalds@cc.helsinki.fi", "time": "Sun Feb 13 17:08:57 1994 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:32 2007 -0400" }, "message": "ALPHA-pl15c\n" }, { "commit": "252058a6e7d838d1cab9512f6240f9e6b5b8b751", "tree": "7fcdee3245d42fbbe8f913f4c126c2c0599d6732", "parents": [ "d5a0671f4362f508274b5d32a5cdf222e8dae5ec" ], "author": { "name": "Linus Torvalds", "email": "torvalds@cc.helsinki.fi", "time": "Wed Feb 09 10:07:53 1994 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:31 2007 -0400" }, "message": "ALPHA-pl15b\n\n[ Message-ID: \u003c2jaihl$21l@klaava.Helsinki.FI\u003e ]\n\n - pl15b mainly fixes some small tty problems.\n" }, { "commit": "d5a0671f4362f508274b5d32a5cdf222e8dae5ec", "tree": "8151d96fcd458e2da336cfb18aac0f1512673f80", "parents": [ "e2778b587aa9fb80b1ce6bf12ff05ab1f7520dde" ], "author": { "name": "Linus Torvalds", "email": "torvalds@cc.helsinki.fi", "time": "Mon Feb 07 14:58:39 1994 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:31 2007 -0400" }, "message": "ALPHA-pl15a\n\n[ Message-ID: \u003c2jaihl$21l@klaava.Helsinki.FI\u003e ]\n\n - pl15a fixes the buffer cache growing problem, adds emulation for a\n few unimportant floating point instructions (i287 instructions that\n are No-Ops on the i387, so \"emulating\" them is easy :^) and fixes a\n silly bug when mmap\u0027ing stuff write-only. It also fixes a buggy lock\n in the networking.\n" }, { "commit": "e2778b587aa9fb80b1ce6bf12ff05ab1f7520dde", "tree": "0872e8de45c468b9706a5f4031ca5c441332e713", "parents": [ "1710d32b4ba3b602008ffd3734d1f88fc5a94675" ], "author": { "name": "Linus Torvalds", "email": "torvalds@cc.helsinki.fi", "time": "Fri Feb 04 22:31:37 1994 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:31 2007 -0400" }, "message": "Linux 0.99.15 released: Codefreeze for 1.0\n\nPeople who look into my directory on ftp.funet.fi will already have\nnoticed that the latest version of linux (0.99.15) is available, and I\nassume it will be available on most other linux sites soon. As\nexplained in a previous announcement, 0.99.15 is \"it\", in that this will\nbe the base for 1.0 after about a month of testing. No further patches\nare accepted until the 1.0 release, unless they obviously fix a serious\nbug.\n\n**** NOTE 1 ****\n\n For this code-freeze to be effective yet still potential bugs be\n found, testing is needed, along with good reports of errors and\n problems. Thus, nobody should think \"hey, the *real* release will be\n out in a month, let\u0027s wait for that\", but instead think: \"hey, I\u0027d\n better test this one, so that the *real* release won\u0027t result in any\n ugly surprises for me\".\n\n In short: test it out, preferably even more than you usually do. Run\n \"crashme\" for the whole month if you have the CPU-power to spare,\n and/or just misuse your machine as badly as you can. And if there are\n problems, report them to me (and the better the report, the more\n likely I am to be able to do something about it).\n\n**** NOTE 2 ****\n\n Bumping the linux version number to 1.0 doesn\u0027t mean anything more\n than that: it\u0027s only a version number change. More explicitly, it\n does *NOT* mean that linux will become commercial (the copyright will\n remain as-is), nor does it mean that development stops here, and that\n 1.0 will be anything special in that respect.\n\n I\u0027m also afraid that just changing the version number will not make\n potential bugs magically disappear: this has been amply proven by\n various software houses over the years. This code-freeze is there in\n order to avoid most of the problems that people sometimes associate\n with \"X.0 releases\", and I hope that it will mean that we have a\n reasonably stable release that we can call 1.0 and one that I won\u0027t\n have to be ashamed of.\n\nOk, enough said, I hope. The pl15 release is hopefully good, but I\u0027ll\ncontinue to make ALPHA patches against it along the whole month as\nproblems crop up. The networking code has been much maligned, and is\nnot perfect by far yet, but it\u0027s getting its act together thanks to\nvarious developers and testers. And as wiser men than I have said (or\nif they haven\u0027t, they should have):\n\n\t\"There is life after 1.0\"\n\nAny rumors that the world is coming to an end just because I\u0027m about to\nrelease a 1.0-version are greatly exaggerated. I think.\n\n\t\t\tLinus\n----------\nThings that remained the same between 0.99.14 and 0.99.15:\n\n - I again forgot to update the README before uploading the release. In\n pl14, I talked about pl13, while the all new and improved README has\n now caught up with pl14. Remind me to buy a new brain one of these\n days.\n\nChanges between versions 0.99.14 and 0.99.15:\n\n - improved Pentium detection. Some of you may have had linux report\n your 4086DX2 as a pentium machine, but the new kernel will tell you\n the sad truth. Whee.\n - Network driver updates by Donald Becker. New drivers added, old ones\n updated.\n - FPU emulation updates by Bill Metzenthen. Various minor errors and\n misfeatures fixed (mostly error handling).\n - Support for the SoubdBlaster Pro CD-ROM driver added by Eberhard\n Moenkeberg.\n - extended support for keyboard re-definition, along with font\n re-programming (Eugene Crosser, Andries Brouwer et al).\n - tty handling fixes: true canonical mode with most features supported\n by Julian Cowley. This may make your canonical mode behave funnily\n if you happen to use old and broken programs that happened to work\n with the old and broken behaviour (this includes at least some\n \u0027getty\u0027 programs).\n - serial driver changes and tty fixes by Theodore Ts\u0027o.\n - SCSI fixes by Drew Eckhardt, Eric Youngdale, Rik Faith, Kai Mdkisara\n et al.\n - Updated sound card driver to version 2.4 (Hannu Savolainen)\n - COFF binary loading support (but you will still need the experimental\n iBCS2 patches to run non-linux i386 COFF binaries) by Al Longyear.\n - Upgraded ext2fs filesystem routines (0.4a -\u003e 0.4b), with new\n features. Read the fs/ext2/CHANGES file for details. Remy Card and\n Stephen Tweedie. Get a new fsck that knows about the new features.\n - pipe behaviour fixed in the presense of multiple writers (now\n actually conforms to POSIX specs about atomic writes). Much of the\n code by Florian Coosmann.\n - minix filesystem extended to support the clean flag: get a new fsck\n that knows about it.\n - System V filesystem (support for Xenix, Coherent and SysV\n filesystems) by Doug Evans, Paul Monday, Pascal Haible and Bruno\n Haible.\n - loadable modules (various authors, don\u0027t remember original author of\n the \"modules\" code).\n - Lots of networking fixes by various people: Alan Cox, Charles\n Hedrick, me and various other people. Non-byte-aligned networks\n work, and the networking code should be much stabler in general.\n\n + various bugfixes and enhacements here and there (mcd driver update by\n Jon Tombs, atixlmouse fix by Chris Colohan, /dev/full by XXX etc etc)\n\nAll in all, the patches come out to 1.5MB uncompressed (about 400kB\ngzip-9\u0027d), so there is little or no idea to make patches to plain pl14\navailable. Incremental patches and ALPHA-releases can be found on\nftp.funet.fi: pub/OS/Linux/PEOPLE/Linus/ALPHA-pl14.\n" }, { "commit": "1710d32b4ba3b602008ffd3734d1f88fc5a94675", "tree": "7b7a1542e60110fbc29386fb285443d3921a2be6", "parents": [ "3cfe2636ea6f3283638c0e8c14f47d7d620f072d" ], "author": { "name": "Linus Torvalds", "email": "torvalds@cc.helsinki.fi", "time": "Wed Feb 02 15:59:29 1994 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:30 2007 -0400" }, "message": "ALPHA-pl14z\n" }, { "commit": "3cfe2636ea6f3283638c0e8c14f47d7d620f072d", "tree": "d087626ffe85d00b68fd8676fc66184829a80ce3", "parents": [ "0ef15e52af53204c01d894863c7bd1bf2ee86cd6" ], "author": { "name": "Linus Torvalds", "email": "torvalds@cc.helsinki.fi", "time": "Wed Feb 02 06:45:07 1994 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:30 2007 -0400" }, "message": "ALPHA-pl14y\n" }, { "commit": "0ef15e52af53204c01d894863c7bd1bf2ee86cd6", "tree": "1b61351d0da736c9632bd44961a027652343c239", "parents": [ "5924f25a1b43d674b453dbcf9adde23d0bc8c2b2" ], "author": { "name": "Linus Torvalds", "email": "torvalds@cc.helsinki.fi", "time": "Sat Jan 29 12:57:42 1994 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:30 2007 -0400" }, "message": "ALPHA-pl14x\n" }, { "commit": "5924f25a1b43d674b453dbcf9adde23d0bc8c2b2", "tree": "f1f386c0c5e2e674e58519eb019901a790ca7b4a", "parents": [ "cdf6bc43e624291b32c55a499e40625aff5de89b" ], "author": { "name": "Linus Torvalds", "email": "torvalds@cc.helsinki.fi", "time": "Thu Jan 27 14:24:02 1994 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:29 2007 -0400" }, "message": "ALPHA-pl14w\n" }, { "commit": "cdf6bc43e624291b32c55a499e40625aff5de89b", "tree": "42ec33965866208fdfa03de1aaa1cb53aee4b90f", "parents": [ "95c5f0cce26ec795afe80ee9fb9a48f22aba2ae2" ], "author": { "name": "Linus Torvalds", "email": "torvalds@cc.helsinki.fi", "time": "Wed Jan 26 12:08:25 1994 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:29 2007 -0400" }, "message": "ALPHA-pl14v\n" }, { "commit": "95c5f0cce26ec795afe80ee9fb9a48f22aba2ae2", "tree": "06b89f116181e40e7b4eb84d76302deb9ad49bbd", "parents": [ "5ba11befea43b83546e0cefa25f2551a7bd17e07" ], "author": { "name": "Linus Torvalds", "email": "torvalds@cc.helsinki.fi", "time": "Tue Jan 25 07:37:08 1994 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:29 2007 -0400" }, "message": "ALPHA-pl14u\n" }, { "commit": "5ba11befea43b83546e0cefa25f2551a7bd17e07", "tree": "42973c35cb5d9758404c146edfd210411cde3b9a", "parents": [ "c5002faf69a59a022ae0d51a35e4836858fea1d0" ], "author": { "name": "Linus Torvalds", "email": "torvalds@cc.helsinki.fi", "time": "Fri Jan 21 13:59:51 1994 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:28 2007 -0400" }, "message": "ALPHA-pl14t\n" }, { "commit": "c5002faf69a59a022ae0d51a35e4836858fea1d0", "tree": "417358626b9343bd6008f6ec2dfed0d67f9fb1e2", "parents": [ "0b5e8609bf7e6899c1fea30aa467812d488b6c11" ], "author": { "name": "Linus Torvalds", "email": "torvalds@cc.helsinki.fi", "time": "Wed Jan 19 12:12:50 1994 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:27 2007 -0400" }, "message": "ALPHA-pl14s\n" }, { "commit": "0b5e8609bf7e6899c1fea30aa467812d488b6c11", "tree": "dc845da00e7092e55aa257acf2b49cc2ce7e5f9f", "parents": [ "e3e62943752a4c7960c04d452718a9b247fe185d" ], "author": { "name": "Linus Torvalds", "email": "torvalds@cc.helsinki.fi", "time": "Tue Jan 18 14:19:08 1994 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:27 2007 -0400" }, "message": "ALPHA-pl14r\n\n[...]\n\nFrom: torvalds@cc.helsinki.fi (Linus Torvalds)\nNewsgroups: comp.os.linux.announce\nSubject: Linux 1.0 codefreeze imminent\nDate: 18 Jan 1994 23:40:37 +0200\n\nThis is a general announcement of the imminent code-freeze that will\nhopefully make linux 1.0 a reality. The plan has been discussed a bit\nwith various developers already, and is already late, but is still in\neffect otherwise.\n\nIn short, the next version of linux (0.99.15) will be a \"full-featured\"\nrelease, and only obvious bug-fixes to existing features will be applied\nbefore calling it 1.0. If this means that your favourite feature or\nnetworking version won\u0027t make it, don\u0027t despair: there is life even\nafter beta (and it\u0027s probably not worth mailing me about it any more:\nI\u0027ve seen quite a few favourite features already ;-).\n\nIn fact, 1.0 has little \"real meaning\", as far as development goes, but\nshould be taken as an indication that it can be used for real work\n(which has been true for some time, depending on your definition of\n\"real work\"). Development won\u0027t stop or even slow down: some of it has\neven been shelved pending a 1.0 already.\n\nCalling it 1.0 will not necessarily make all bugs go away (quite the\nopposite, judging by some other programs), but I hope it will be a\nreasonably stable release. In order to accomplish this, the code-freeze\nafter 0.99.15 will be about a month, and I hope people will test out\nthat kernel heavily, instead of waiting for \"the real release\" so that\nany potential bugs can be found and fixed.\n\nAs to where we are now: as of this moment, the latest release is the \u0027r\u0027\nversion of pl14 (aka \"ALPHA-pl14r\"). I\u0027ve made ALPHA releases available\non ftp.funet.fi almost daily, and expect a final pl15 within a few more\ndays. Testing out the ALPHA releases is not discouraged either if you\nlike recompiling kernels every day or two..\n\nAnd finally: we also try to create a \"credits\" file that mentions the\ndevelopers of the kernel and essential linux utilities. The credit file\ncompilator is jmartin@opus.starlab.csc.com (John A. Martin), and if you\nfeel you have cause to be mentioned in it, please contact him.\n\n\t\tLinus\n" }, { "commit": "e3e62943752a4c7960c04d452718a9b247fe185d", "tree": "d1c19ba60a81b8f63a013f529770e9d43ba26083", "parents": [ "5f295b33cf4ea277617047421018a6cd75cc5dff" ], "author": { "name": "Linus Torvalds", "email": "torvalds@cc.helsinki.fi", "time": "Mon Jan 17 13:18:10 1994 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:27 2007 -0400" }, "message": "ALPHA-pl14q\n" }, { "commit": "5f295b33cf4ea277617047421018a6cd75cc5dff", "tree": "2851c8b70dda15759fac8f1b310402391fb16ada", "parents": [ "9ee059ad6f92a777f874ddf472cd36aa9a901c16" ], "author": { "name": "Linus Torvalds", "email": "torvalds@cc.helsinki.fi", "time": "Thu Jan 13 21:08:25 1994 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:26 2007 -0400" }, "message": "ALPHA-pl14p\n" }, { "commit": "9ee059ad6f92a777f874ddf472cd36aa9a901c16", "tree": "b89bcf7ec57269238ee371f67b6a3e1344c89dba", "parents": [ "a9cc257a8016df80d8bbfdc3ffc6dd2193280c0a" ], "author": { "name": "Linus Torvalds", "email": "torvalds@cc.helsinki.fi", "time": "Sat Jan 08 13:13:10 1994 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:26 2007 -0400" }, "message": "ALPHA-pl14o\n" }, { "commit": "a9cc257a8016df80d8bbfdc3ffc6dd2193280c0a", "tree": "30593a304fc18da6479c71c994e4625b0c27b72d", "parents": [ "eb54995bc4b5a31293b1cd153d1460570f32125d" ], "author": { "name": "Linus Torvalds", "email": "torvalds@cc.helsinki.fi", "time": "Fri Jan 07 14:50:53 1994 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:25 2007 -0400" }, "message": "ALPHA-pl14n\n" }, { "commit": "eb54995bc4b5a31293b1cd153d1460570f32125d", "tree": "0d0b2c837aa3415980b33815e6851be5a27ee13c", "parents": [ "12a5cbbb06981b64fa0f8e3dc7ad3405b3d730a8" ], "author": { "name": "Linus Torvalds", "email": "torvalds@cc.helsinki.fi", "time": "Wed Jan 05 11:21:10 1994 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:25 2007 -0400" }, "message": "ALPHA-pl14m\n" }, { "commit": "12a5cbbb06981b64fa0f8e3dc7ad3405b3d730a8", "tree": "eb1d8930455514f746423da4eff1b95388e2af14", "parents": [ "8c14943430cec122e60f8065d087c095b5b54eca" ], "author": { "name": "Linus Torvalds", "email": "torvalds@cc.helsinki.fi", "time": "Tue Jan 04 12:26:13 1994 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:25 2007 -0400" }, "message": "ALPHA-pl14l\n" }, { "commit": "8c14943430cec122e60f8065d087c095b5b54eca", "tree": "236a6837f1ae3ed632edb61f2f46a020cc9f106c", "parents": [ "5a09e4bfb5fefaf040962a64d74a4ac43b2d20e5" ], "author": { "name": "Linus Torvalds", "email": "torvalds@cc.helsinki.fi", "time": "Mon Jan 03 10:27:45 1994 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:24 2007 -0400" }, "message": "ALPHA-pl14k\n" }, { "commit": "5a09e4bfb5fefaf040962a64d74a4ac43b2d20e5", "tree": "37a56ca6bebdcd0f7b3eac46f685bf18a60b0945", "parents": [ "221e2a9d6d67588221884be20abcc900cd108807" ], "author": { "name": "Linus Torvalds", "email": "torvalds@cc.helsinki.fi", "time": "Fri Dec 31 10:01:44 1993 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:24 2007 -0400" }, "message": "ALPHA-pl14j\n" }, { "commit": "221e2a9d6d67588221884be20abcc900cd108807", "tree": "903d6da9a95c584a2baf4cd7aad92eb72443ed78", "parents": [ "0dde01d03e072edeb56d81e9f027ed4ef5c088eb" ], "author": { "name": "Linus Torvalds", "email": "torvalds@cc.helsinki.fi", "time": "Thu Dec 30 13:17:23 1993 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:24 2007 -0400" }, "message": "ALPHA-pl14i\n" }, { "commit": "0dde01d03e072edeb56d81e9f027ed4ef5c088eb", "tree": "46803ed5c2fb9df5169f05766f3f54fed6bd40f9", "parents": [ "76123c5282eca4effe5fbddf2beb64f30c8a8739" ], "author": { "name": "Linus Torvalds", "email": "torvalds@cc.helsinki.fi", "time": "Mon Dec 27 13:22:44 1993 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:23 2007 -0400" }, "message": "ALPHA-pl14h\n" }, { "commit": "76123c5282eca4effe5fbddf2beb64f30c8a8739", "tree": "54b100dd8bcf7cc81fcf05f649a233452b736f18", "parents": [ "fb11be238769030d3e880b6cad3fc9e5e593495c" ], "author": { "name": "Linus Torvalds", "email": "torvalds@cc.helsinki.fi", "time": "Thu Dec 23 11:51:31 1993 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:23 2007 -0400" }, "message": "ALPHA-pl14g\n" }, { "commit": "fb11be238769030d3e880b6cad3fc9e5e593495c", "tree": "92f5a67c6f98342880c5d5e6a83e56b58a4e004d", "parents": [ "59fa0b1fd08fe53a8a9294460c89dda09ee0255d" ], "author": { "name": "Linus Torvalds", "email": "torvalds@cc.helsinki.fi", "time": "Thu Dec 16 12:27:40 1993 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:23 2007 -0400" }, "message": "ALPHA-pl14f\n" }, { "commit": "59fa0b1fd08fe53a8a9294460c89dda09ee0255d", "tree": "759418321ab216123bd9ddb1dc5494d7325e04c3", "parents": [ "8ee558f91b9940ca35f758c4f240b7b5b3ffb76a" ], "author": { "name": "Linus Torvalds", "email": "torvalds@cc.helsinki.fi", "time": "Tue Dec 14 10:43:29 1993 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:22 2007 -0400" }, "message": "ALPHA-pl14e\n" }, { "commit": "8ee558f91b9940ca35f758c4f240b7b5b3ffb76a", "tree": "0d0a3c6b8b9ab15309a0ac8d2c334d9db7f1fa36", "parents": [ "89fc4813f91b51c803c68c9256d538270035eeae" ], "author": { "name": "Linus Torvalds", "email": "torvalds@cc.helsinki.fi", "time": "Mon Dec 13 11:01:47 1993 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:22 2007 -0400" }, "message": "ALPHA-pl14d\n" }, { "commit": "89fc4813f91b51c803c68c9256d538270035eeae", "tree": "5304bcec26c9e22ce669a66c70c028a87c312193", "parents": [ "3d40b9081a8372ab0152d49d1c43b62264eab0bd" ], "author": { "name": "Linus Torvalds", "email": "torvalds@cc.helsinki.fi", "time": "Fri Dec 10 14:29:01 1993 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:22 2007 -0400" }, "message": "ALPHA-pl14c\n" }, { "commit": "3d40b9081a8372ab0152d49d1c43b62264eab0bd", "tree": "99ab7fb94ecf1619e50de0eed124a016ec3a95e8", "parents": [ "0b99366a1b374cd27a2edf386cd89f0bc1e1946b" ], "author": { "name": "Linus Torvalds", "email": "torvalds@cc.helsinki.fi", "time": "Thu Dec 09 10:56:12 1993 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:21 2007 -0400" }, "message": "ALPHA-pl14b\n" }, { "commit": "0b99366a1b374cd27a2edf386cd89f0bc1e1946b", "tree": "b86e051e69258016440370c9a420d0a2bac6019b", "parents": [ "75bb5836a8a8c0ee44ffd60a51f357b9568f1381" ], "author": { "name": "Linus Torvalds", "email": "torvalds@cc.helsinki.fi", "time": "Fri Dec 03 15:26:49 1993 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:21 2007 -0400" }, "message": "ALPHA-pl14a\n" }, { "commit": "75bb5836a8a8c0ee44ffd60a51f357b9568f1381", "tree": "806df8b9b14f217f49d026d76bd26af4cd05cdfb", "parents": [ "04cb41e1214fd968f1f50d1ecbf30839c41e3efa" ], "author": { "name": "Linus Torvalds", "email": "torvalds@cc.helsinki.fi", "time": "Mon Nov 29 14:21:10 1993 +0200" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:21 2007 -0400" }, "message": "Linux 0.99 patchlevel 14\n\nLinux 0.99 patchlevel 14 is available on nic.funet.fi in the usual place\n(pub/OS/Linux/PEOPLE/Linus). There are no diffs relative to pl13, as\ntoo much has changed (the directory structure changed and the sound\ndriver was added). Diffs relative to the last ALPHA version (13t) are\nin the \"pl13-ALPHA\u0027s\" subdirectory along with the actual ALPHA versions.\n\nThe changes to pl13t are rather minor: most of them are just more printf\nformat fixes to make gcc-2.5.x happy (Chip Salzenberg). Only one very\nminor bugfix which made pl13t not notice the WP bit on a 486.\n\nIt would seem to be a good idea to use gcc-2.5.x to compile the kernel,\nas that seems to fix at least one known bug in earlier gcc versions. I\nhope that pl14 will be even more stable than pl13 has turned out to be,\nand especially the networking code seems to have become much more\ndependable. Thanks Alan \u0026 co.\n\nChanges to the last official release (p13) are too numerous to mention\n(or even to remember), but they include NTP support, updated SCSI and\nnetworking drivers, \u003e16MB swap area handling, added sound support,\nread-only HPFS filesystem, memory management cleanups (especially\ncleaned up mmap() some more). Also, pl14 contains updated ext2fs code,\nalong with minor fixes (especially concerning the time values) in other\nfilesystems, and fixed unnamed/named pipe select() semantics.\n\nThe reorganizations include moving all device drivers to a subdirectory\nof their own (linux/drivers), centralizing the major number handling\n(\u003clinux/major.h\u003e etc... Possibly cleaner and/or easier to keep track of\ndifferent drivers. Finally, the first 4kB of physical memory is no\nlonger cleared on bootup: tytso reports that this feature now enables\nsome portables to use the power-saving features under linux. This could\nalso be useful for the DOS emulator to check where the interrupt\npointers pointed at startup.\n\n\t\t\tLinus\n" }, { "commit": "04cb41e1214fd968f1f50d1ecbf30839c41e3efa", "tree": "1b716d518cf163f369fce42ff70e425401d6bf4c", "parents": [ "3a846f010d35ec39c41d7e455e4b8b775a8c0245" ], "author": { "name": "Linus Torvalds", "email": "torvalds@cs.helsinki.fi", "time": "Mon Oct 25 22:33:27 1993 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:20 2007 -0400" }, "message": "ALPHA-pl13k\n\n[ Committer\u0027s note: This is the only 0.99.13+ release that could be\n found. It is extracted from a private copy apparently from \"aeb\"\n and possibly non pristine. Its date must be approximate as well. ]\n" }, { "commit": "3a846f010d35ec39c41d7e455e4b8b775a8c0245", "tree": "9d96645d784804ffe6f0045d446bfcb8c3cfc5a5", "parents": [ "9b07f0392c1a3b97f8682a967071bcf8210b4934" ], "author": { "name": "Linus Torvalds", "email": "torvalds@cs.helsinki.fi", "time": "Sat Sep 25 19:15:56 1993 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:18 2007 -0400" }, "message": "Linux 0.99.13 released\n\nLinux 0.99pl13 has been released (for almost a week no, actually, sorry\nfor the delayed announcement), and can be found on nic.funet.fi in\npub/OS/Linux/PEOPLE/Linus and mirror sites.\n\nRough changes since pl12 (most of which have been there in various ALPHA\nversions):\n\n - the bad memory management one-liner bug is naturally fixed.\n - compiled with plain C by default instead of C++\n - ELF binary support (Eric Youngdale)\n - Quickport mouse support (and some changes to the PS/2 mouse driver)\n by Johan Myreen and co)\n - core file name change (\"core\" -\u003e \"core.xxxx\" where xxxx is the name\n of the program that dumped code). Idea from ???. Also, core-files\n now correctly truncate any existing core file before being written.\n - some mmap() fixes: better error returns, and handling of non-fixed\n maps for /dev/mem etc.\n - fixes for rename/unlink/rmdir at mount-points.\n - packet mode fixes by Charles Hedrick. Sadly, these are likely to\n break old telnet/rlogin binaries, but it had to be done in order to\n communicate correctly with the rest of the world.\n - FPU emulator patches from Bill Metzenthen. The fprem1 insn should be\n correct now (not that anybody seems to have seen the incorrect\n behaviour..)\n - a few fixes for SCSI (Drew and Eric)\n - signal.c changes to handle multiple segments (for Wine) correctly.\n - updated drivers from Donald Becker: 3c509 and AT1500 drivers, but\n also some other drivers have been edited, and some networking fixes.\n\nNote that I have gotten a lot of patches that simply didn\u0027t make it: I\nthink I have 5-10 reasonable patches that I simply never got around to\ncheck very carefully but that have their own reasons for existence,\nranging from sysv signal handling to quotas. They\u0027ll probably get there\neventually, but didn\u0027t make it into pl13.\n\n\t\tLinus\n" }, { "commit": "9b07f0392c1a3b97f8682a967071bcf8210b4934", "tree": "05f9be038fd6b1ebb19d94f559af868d284418f8", "parents": [ "f1e7ea6f72640d1033eeeb2ff0c4daf9d12256ab" ], "author": { "name": "Linus Torvalds", "email": "torvalds@cc.helsinki.fi", "time": "Wed Aug 18 09:50:58 1993 -0400" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:17 2007 -0400" }, "message": "Very small patch to 0.99pl12\n\nI hate to put out patches this soon after a release, but there is one\npotentially major problem in pl12 which is very simple to fix.. I\u0027m\nincluding patches: both in plain ascii and as a uuencoded gzip file\n(it\u0027s the same patch - the uuencoded one is in case there is any\nnewsserver that messes up whitespace).\n\nThe main patch is just the change from __get_free_page(GFP_BUFFER) into\nget_free_page(GFP_KERNEL), and the two minor patches just add checks\nthat actually enforce the read-only nature of current file mmap\u0027ings so\nthat any program that tries to do a write mapping at least will be told\nthat it won\u0027t work.\n\nI\u0027d suggest anybody compiling pl12 should add at least the file_table.c\npatch: thanks to Alexandre Julliard for noticing this one.\n\n\t\tLinus\n" }, { "commit": "f1e7ea6f72640d1033eeeb2ff0c4daf9d12256ab", "tree": "973041669cc09e5ee3554edbb17b8babad829bf8", "parents": [ "35214e687d47d66b771c59a34632792dc04c7587" ], "author": { "name": "Linus Torvalds", "email": "torvalds@cc.helsinki.fi", "time": "Mon Aug 16 08:50:03 1993 -0400" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:17 2007 -0400" }, "message": "Linux 0.99 patchlevel 12 released\n\nAs promised, the 0.99.12 kernel made it out this weekend: it\u0027s\nessentially the last ALPHA-pl12 with some minor changes that shouldn\u0027t\nbreak anything (famous last words) while they should be a boon\nespecially for NFS users.\n\nNic.funet.fi: pub/OS/Linux/PEOPLE/Linus now contains the 0.99.12 kernel\nas both full source (linux-0.99.12.tar.gz) and as patches against pl11\n(linux-0.99.patch12.gz). It\u0027s usually easier to get the full sources:\nexpecially due to some cosmetic fixes the patches are pretty large.\n\nNote that kbd.tar.gz (at the same place as the kernel) has been updated\nyet again to fix some problems with the Swiss keyboard mappings. Hope\nthat caught the last of these problems. Also note that the manual pages\nin kbd.tar.gz aren\u0027t up-to-date but that the format of the keyboard\nfiles shouldn\u0027t really pose any problems as they are pretty self-\nexplanatory (the man-pages will be fixed eventually, probabably not by\nme).\n\nAlso note that the pl12 kernel is more strict about routing entries, and\nthat there is a bug in the 4.4.1 library which may make adding routes\nextremely difficult (especially if you try to add a C-net route that has\nbeen subnetted from a B-net). Libc-4.4.2 fixes the problem, but if you\ndon\u0027t use subnetting or any other special netmasks, you\u0027ll never see the\nbug anyway.\n\nI\u0027m including some of the README so that people can see what\u0027s new..\n\n\t\tLinus\n\nPS. The network card configs are now in the main \"make config\", but you\nshould check the net/inet/CONFIG file as well anyway. Also, the 3c509\ndriver is probably not functional yet, so don\u0027t get too excited.\n----------\n\n\tLinux kernel release 0.99 patchlevel 12\n\nThese are the release notes for linux version 0.99.12. Read them\ncarefully, as they tell you what\u0027s new, explain how to install the\nkernel, and what to do if something goes wrong.\n\nNOTE! There has been some indication that gcc versions older than 2.4.5\nresult in bad kernels being built: 2.3.3 will fail even to build the\nkernel, and I have at least one report of trouble with a 2.4.3-built\nkernel that went away when the kernel was recompiled with 2.4.5.\n\nCHANGES since 0.99 patchlevel 11 and earlier:\n\n - The memory manager cleanup has continued, and seems to be mostly\n ready, as proven by the ease of adding mmap() over NFS with the new\n routines. So yes, the pl12 kernel will demand-load your binaries\n over NFS, sharing code and clean data, as well as running shared\n libraries over NFS. Memory management by Eric and me, while the NFS\n mmap code was written by Jon Tombs,\n\n - ** IMPORTANT **: The keyboard driver has been enhanced even further,\n and almost everything is completely re-mappable. This means that\n there is a new version of \u0027loadkeys\u0027 and \u0027dumpkeys\u0027 that you must use\n with this kernel or you\u0027ll have problems. The default keyboard is\n still the US mapping, but if you want to create your own mappings\n you\u0027ll have to load them with the new binaries. Get the \u0027kbd.tar.gz\u0027\n archive from the same place you get the kernel.\n\n The new keymappings allow things like function key string changes,\n remapping of the control keys, and freedom to remap any of the normal\n keyboard functions: including special features like rebooting,\n console switching etc. The keyboard remapping code has been done\n mostly by Risto Kankkunen (Risto.Kankku...@Helsinki.FI).\n\n - updated network drivers by Donald Becker\n\n - updated serial drivers - ty...@Athena.mit.edu\n\n - updated 387 emulation (Bill Metzenthen). The updated emulator code\n has more exact trigonometric functions and improved exception\n handling. It now behaves very much like a real 486, with only small\n changes (greater accuracy, slightly different denormal NaN handling\n etc - hard to detect the differences even if you are looking for\n them).\n\n - network timer fixes by Florian La Roche (much cleaned up net/inet/timer.c\n and some bad race-conditions fixed).\n\n - Scsi code updates by Eric Youngdale and others\n\n - Sony CDU-31A CDROM driver by Corey Minyard added to the standard\n kernel distribution.\n\n - The Mitsumi CDROM driver is now part of the standard kernel. Driver\n by Martin Harriss with patches by stu...@cc4.kuleuven.ac.be (yes, he\n probably has a real name, but no, I haven\u0027t found it) and Jon Tombs.\n\n - various other minor patches (preliminary ldt support etc)\n\n[ rest deleted ]\n" }, { "commit": "35214e687d47d66b771c59a34632792dc04c7587", "tree": "7edfa0a6ce69dc8f63534be0d584482201ba9562", "parents": [ "46db5850d3c10ede04f645bd1e8bd7bcbc65acb9" ], "author": { "name": "Linus Torvalds", "email": "torvalds@cc.helsinki.fi", "time": "Mon Jul 19 12:25:58 1993 -0400" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:16 2007 -0400" }, "message": "PATCH: fork.c bug in 0.99.pl11\n\nThere is at least one known problem with 0.99pl11 - it\u0027s very minor and\nwill not lead to any real problems, but it\u0027s also very easy to fix,\nso...\n\nThe problem is a one-liner oversight in kernel/fork.c (thanks to TjL for\nnoticing the symptoms - they aren\u0027t easy to see), which is fixed by the\nfollowing patch:\n\n----- snip snip -----\n--- linux/kernel/fork.c.orig\tMon Jul 19 19:09:45 1993\n+++ linux/kernel/fork.c\tMon Jul 19 16:55:04 1993\n@@ -157,7 +157,7 @@\n \tp-\u003etss.cs \u003d KERNEL_CS;\n \tp-\u003etss.ss \u003d KERNEL_DS;\n \tp-\u003etss.ds \u003d KERNEL_DS;\n-\tp-\u003etss.fs \u003d KERNEL_DS;\n+\tp-\u003etss.fs \u003d USER_DS;\n \tp-\u003etss.gs \u003d KERNEL_DS;\n \tp-\u003etss.ss0 \u003d KERNEL_DS;\n \tp-\u003etss.esp0 \u003d p-\u003ekernel_stack_page + PAGE_SIZE;\n----- snip snip -----\n\nIn fact, it\u0027s probably easiest to \"apply\" this patch by hand: just\nchange the \"p-\u003etss.fs \u003d KERNEL_DS\" in fork.c to \"p-\u003etss.fs \u003d USER_DS\"\nand you should be fine.\n\n\t\tLinus\n" }, { "commit": "46db5850d3c10ede04f645bd1e8bd7bcbc65acb9", "tree": "93499a7b1d8f367c47ed9bb1ed7cb285235ec490", "parents": [ "d3465417f56e99c858e0bc136e63aa67e6ef90bf" ], "author": { "name": "Linus Torvalds", "email": "torvalds@cs.Helsinki.FI", "time": "Sun Jul 18 20:13:34 1993 -0400" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:16 2007 -0400" }, "message": "[ANNOUNCE]: linux 0.99 patchlevel 11\n\nNic.funet.fi now contains the newest linux kernel: this is 0.99\npatchlevel 11. I\u0027ve had a few problems with the \"ls\" and \"dir\" commands\non nic, so if this is true for others as well, you may not be able to\nsee the files, but they are available as:\n\n\tpub/OS/Linux/PEOPLE/Linus:\n\n\t\t- RELEASE-0.99.11 -- release notes\n\t\t- README -- same as RELEASE..\n\t\t- linux-0.99.11.tar.gz -- full sources\n\t\t- linux-0.99.patch11.gz -- patches against pl10\n\nI don\u0027t know if I\u0027ll actually have the energy to update the RELEASE\nfiles every time, but that\u0027s the idea (with README being a copy of the\nlatest one). The README gives info on what changed, how to install it,\nhow to compile the kernel, and what to do with error dumps like \"unable\nto handler kernel paging request\" and the like. Maybe it results in\nbetter bug-reports. Maybe not.\n\nPl11 uses (as you all know by now) C++, so you\u0027ll need to have g++\ninstalled. Also, there seems to be problems compiling the newer kernels\nwith gcc-2.3.3: there have been various reports that even pl10 (which\nisn\u0027t C++) had problems which go away when compiled with gcc-2.4.3 or\nnewer.\n\nAs always, please send any bug-reports etc at least Cc\u0027d to me: I\u0027ll be\nwanting to know if things work or not.\n\n\t\tLinus\n\n\u003cRELEASE-0.99.11\u003e\n\n\tLinux kernel release 0.99 patchlevel 11\n\nThese are the release notes for linux version 0.99.11. Read them\ncarefully, as they explain how to install the kernel, and what to do if\nsomething goes wrong.\n\nCHANGES since 0.99 patchlevel 10 and earlier:\n\n - The keyboard is dynamically changeable (this is true of pl10 as\n well), and you need to get the \"keytables.tar.z\" archive to set the\n keyboard to suit your taske unless you want to live with the default\n US keymaps.\n\n Use the \"loadkeys map/xxx.map\" command to load the keyboard map: you\n can edit the maps to suit yourself if you can\u0027t find a suitable one.\n The syntax of the keyboard maps should be obvious after looking at\n the examples.\n\n - The memory manager has been cleaned up substantially, and mmap()\n works for MAP_PRIVATE. MAP_SHARED is still not supported for\n anything else than /dev/mem, but even so it actually is usable for a\n lot of applications. The shared library routines have been rewritten\n to use mmap() instead of the old hardcoded behaviour.\n\n - The kernel is now compiled with C++ instead of plain C. Very few\n actual C++ features are used, but even so C++ allows for more\n type-checking and type-safe linkage.\n\n - The filesystem routines have been cleaned up for multiple block\n sizes. None of the filesystems use it yet, but people are working on\n it.\n\n - named pipes and normal pipes should hopefully have the right select()\n semantics in the presense/absense of writers.\n\n - QIC-02 tape driver by Hennus Bergman\n\n - selection patches in the default kernel\n\n - fixed a bug in the pty code which led to busy waiting in some\n circumstances instead of sleeping.\n\n - Compressed SLIP support (Charles Hedrick). See net/inet/CONFIG\n\nINTERNAL kernel changes:\n\n - the \u0027clear_bit()\u0027 function was changed to return the previous setting\n of the bit instead of the old \"error-code\". This makes use of the\n bit operations more logical.\n\n - udelay() function for short delays (busy-waiting) added. Used\n currently only by the QIC driver.\n\n - fork() and sheduler changes to make task switches happen only from\n kernel mode to kernel mode. Cleaner and more portable than the old\n code which counted on being able to task-switch directly into user\n mode.\n\n - debugging malloc code.\n\nINSTALLING the kernel:\n\n - if you install by patching, you need a *clean* 0.99.10 source tree,\n which presumably exists in /usr/src/linux. If so, to get the kernel\n patched, just do a\n\n\t\tcd /usr/src\n\t\tpatch -p0 \u003c linux-0.99.patch11\n\n and you should be ok. You may want to remove the backup files (xxx~\n or xxx.orig), and make sure that there are no failed patches (xxx# or\n xxx.rej).\n\n - If you install the full sources, do a\n\n\t\tcd /usr/src\n\t\ttar xvf linux-0.99.11.tar\n\n to get it all put in place.\n\n - make sure your /usr/include/linux and /usr/include/asm directories\n are just symlinks to the kernel sources:\n\n\t\tcd /usr/include\n\t\trm -rf linux\n\t\trm -rf asm\n\t\tln -s /usr/src/linux/include/linux .\n\t\tln -s /usr/src/linux/include/asm .\n\n - make sure you have no stale .o files and dependencies lying around:\n\n\t\tcd /usr/src/linux\n\t\tmake mrproper\n\t\tmake dep\n\n You should now have the sources correctly installed.\n\nCONFIGURING the kernel:\n\n - do a \"make config\" to configure the basic kernel. \"make config\"\n needs bash to work: it will search for bash in $BASH, /bin/bash and\n /bin/sh (in that order), so hopefully one of those is correct.\n\n - edit net/inet/CONFIG to configure the networking parts of the kernel.\n The comments should hopefully clarify it all.\n\n - Check the top Makefile for further site-dependent configuration\n (default SVGA mode etc).\n\nCOMPILING the kernel:\n\n - make sure you have gcc-2.4.3 or newer available with g++. It seems\n older gcc versions can have problems compiling linux 0.99.10 and\n newer versions. If you upgrade, remember to get the new binutils\n package too (for as/ld/nm and company)\n\n - do a \"make zImage\" to create a compressed kernel image. If you want\n to make a bootdisk (without root filesystem or lilo), insert a floppy\n in your A: drive, and do a \"make zdisk\". It is also possible to do\n \"make zlilo\" if you have lilo installed to suit the kernel makefiles,\n but you may want to check your particular lilo setup first.\n\n - keep a backup kernel handy in case something goes wrong.\n\n - reboot with the new kernel.\n\nIF SOMETHING GOES WRONG:\n\n - if you have problems that seem to be due to kernel bugs, please mail\n them to me (Linus.Torvalds@Helsinki.FI), and possibly to any other\n relevant mailing-list or to the newsgroup. The mailing-lists are\n useful especially for SCSI and NETworking problems, as I can\u0027t test\n either of those personally anyway.\n\n - In all bug-reports, *please* tell what kernel you are talking about,\n how to duplicate the problem, and what your setup is (use your common\n sense). If the problem is new, tell me so, and if the problem is\n old, please try to tell me when you first noticed it.\n\n - if the bug results in a message like\n\n\tunable to handle kernel paging request at address C0000010\n\tOops: 0002\n\tEIP: 0010:xxxxxxxx\n\teax: xxxxxxxx ebx: xxxxxxxx ecx: xxxxxxxx edx: xxxxxxxx\n\tesi: xxxxxxxx edi: xxxxxxxx ebp: xxxxxxxx\n\tds: xxxx es: xxxx fs: xxxx gs: xxxx\n\tPid: xx, process nr: xx\n\txx xx xx xx xx xx xx xx xx xx\n\n or similar kernel debugging information on your screen or in your\n system log, please duplicate it *exactly*. The dump may look\n incomprehensible to you, but it does contain information that may\n help debugging the problem. The text above the dump is also\n important: it tells something about why the kernel dumped code (in\n the above example it\u0027s due to a bad kernel pointer)\n\n - in debugging dumps like the above, it helps enourmously if you can\n look up what the EIP value means. The hex value as such doesn\u0027t help\n me or anybody else very much: it will depend on your particular\n kernel setup. What you should do is take the hex value from the EIP\n line (ignore the \"0010:\"), and look it up in the kernel namelist to\n see which kernel function contains the offending address.\n\n To find out the kernel function name, you\u0027ll need to find the system\n binary associated with the kernel that exhibited the symptom. In the\n case of compressed kernels, this will be \u0027linux/tools/zSystem\u0027, while\n uncompressed kernels use the file \u0027tools/system\u0027. To extract the\n namelist and match it against the EIP from the kernel crash, do:\n\n\t\tnm tools/zSystem | sort | less\n\n This will give you a list of kernel addresses sorted in ascending\n order, from which it is simple to find the function that contains the\n offending address. Note that the address given by the kernel\n debugging messages will not necessarily match exactly with the\n function addresses (in fact, that is very unlikely), so you can\u0027t\n just \u0027grep\u0027 the list: the list will, however, give you the starting\n point of each kernel function, so by looking for the function that\n has a starting address lower than the one you are searching for but\n is followed by a function with a higher address you will find the one\n you want. In fact, it may be a good idea to include a bit of\n \"context\" in your problem report, giving a few lines around the\n interesting one.\n\n If you for some reason cannot do the above (you have a pre-compiled\n kernel image or similar), telling me as much about your setup as\n possible will help.\n\n-----\n\nFrom: torvalds@cc.helsinki.fi (Linus Torvalds)\nSubject: ALPHA-pl11 available on nic: C++ support\nDate: Sun, 4 Jul 93 00:03:03 GMT\n\nFor those brave souls that enjoy testing new releases, there is an\nALPHA-release of the 0.99.11 version available on nic.funet.fi in the\nusual place (pub/OS/Linux/PEOPLE/Linus). This has a few changes from\nthe last one, the most notable being that it is compiled using C++, as\nthere was some interest in that on the c.o.l newsgroup. Note that very\nfew C++ features are actually used: the major changes were some minor\nsyntactic editing and the addition of \u0027extern \"C\"\u0027 to functions called\nfrom assembly code. The C++ changes are not the same as those done by\nTristan (although you should thank him for getting it rolling), as I\nwanted to resolve the differences between C and C++ a bit differently.\n\nThe C++ changes shouldn\u0027t actually change the way the kernel works, and\nit\u0027s mainly used currently to do stricter pointer checking. The name\nmangling probably breaks the kmem based \u0027ps\u0027 once again.. Actual code\nchanges for this test-version:\n\n - I added the patches by Charles Hedrick for SLIP: this actually means\n that only CSLIP is available for now, so.. The net code is still not\n ready: Fred is working on it, so this is just an interim version when\n it comes to networking (there are some other minor patches in there as\n well).\n\n - The mm has been cleaned up since pl10, and mmap() actually works for\n most things, while malloc() will return NULL when the kernel thinks\n there isn\u0027t enough memory. Similarly, the buffer cache code should\n now support different block sizes (although this is still in the\n \"early alpha\" stage. Most of the changes by Eric Youngdale, with mm\n cleanups by me.\n\n - A problem with the dynamic inode code (insufficient inode\n invalidation) that could result in fs corruption under some\n circumstances is fixed.\n\nPeople who have written drivers etc should probably check out the\nchanges I did due to the stricter C++ pointer checking.\n\n\t\tLinus\n" }, { "commit": "d3465417f56e99c858e0bc136e63aa67e6ef90bf", "tree": "05c90924879d1385f09d578561f4220c7eb7bad3", "parents": [ "87aa05879df067f26eb0120af913210ff94e32cf" ], "author": { "name": "Linus Torvalds", "email": "torvalds@cc.helsinki.fi", "time": "Tue Jun 08 14:45:33 1993 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:15 2007 -0400" }, "message": "[ANNOUNCE] linux-0.99 patchlevel 10\n\nI\u0027ve finally released an official version of linux-0.99 patchlevel 10:\nthere have been various alpha versions floating around which differ in\ndetails (notably networking code), which shouldn\u0027t be used any more.\nThe new linux version is available only as full source code: the diffs\nwould have been too big to be useful. You can find linux-0.99.10.tar.z\n(along with keytables.tar.z) on nic.funet.fi: pub/OS/Linux/PEOPLE/Linus\nand probably on tsx-11 and other linux archives within a day or two (so\ncheck there first if you are in the states).\n\nLinux-0.99 pl10 has a number of new features and changes in interface.\nThe most notable of these are:\n\n - the networking code is reorganized (generally called \"net-2\",\n although unrelated to the BSD release). The new code implements a\n lot of standard features lacking in net-1, and also changes the user\n interface to be closer to the BSD standards. Notably, the old\n configuration binaries won\u0027t work, so to get the new networking to\n work you\u0027ll have to get the net-2 binaries as well. The networking\n binaries are available on tsx-11.mit.edu (and mirrors) under the\n directory pub/linux/packages/net/net-2 (and the setup syntax has\n changed somewhat..)\n\n The networking code has been mainly organized and rewritten by Fred\n van Kempen, with drivers by Donald Becker.\n\n - serial line setup has been changed: linux 0.99 pl10 does *not* try to\n autodetect serial ports very agressively. If you have other serial\n ports than the standard com1/com2, or nonstandard IRQ etc values,\n this means that it\u0027s less likely to work without any help. The\n solution is not to recompile the kernel - you should get the\n \"setserial\" program available from tsx-11.mit.edu in the directory\n pub/linux/sources/sbin/setserial-2.01.tar.z that allows you to\n dynamically configure your serial ports to suit your setup.\n\n The main organizer behind the serial line changes is tytso (Theodore\n Ts\u0027o).\n\n - Keyboard setup has changed: it is no longer hardcoded at compile\n time, but instead you can use the new \"loadkeys\" program to load in a\n new keyboard map on the fly. The default keyboard map is the normal\n US keyboard (yes, I should have used the Finnish one by default, but\n after thinking of all the problems that would have resulted in I\n forgot about that idea). The loadkeys code can be found in the\n \"keytables.tar.z\" archive, which also contains keymaps for most\n normal keyboard types. To create a custom keyboard table is very\n easy - just take a 5 minute look at the existing map files (they\n resemble the ones used by xmodmap, so if you are familiar with\n those..)\n\n The loadable keymaps were mostly implemented by Risto Kankkunen.\n\nThere are a lot of other internal kernel changes, but they should be\nmostly transparent, and noticeable only indirectly due to new features\nor (hopefully) better/faster/whatever operation. These include:\n\n - the SysV IPC patches are in by default: Krishna Balasubramanian.\n If you need these, you know what it\u0027s about (notably, dosemu 0.49\n wants them).\n - inode handling is updated: inodes and files are now dynamically\n allocated within the kernel, and use a hash table for faster lookup\n (along with a NFU algorithm for the inode cache). Steven Tweedie.\n - Updated FPU emulation: mostly exception handling changes, making the\n emulator handle most exceptions the same way a 486 does. The\n emulator is written by Bill Metzenthen.\n - a few ext2-fs updates by Remy Card and Steven Tweedie.\n - support for the \u0027fsync()\u0027 function (Steven Tweedie)\n - various (minor) SCSI patches to catch some error conditions, add\n support for VLB adaptec controllers without DMA and so on (different\n people).\n - other changes - I forget.\n\nIn addition to patches sent in by others, I\u0027ve naturally made my own\nchanges (often *to* the patches sent in by others :-). Among other\nthings, the pl10 buffer cache code now also tries to share pages with\nexecutables, resulting in better cacheing especially of binaries (giving\nnoticeable improvements in kernel recompilation speed on some machines).\nAlso, I\u0027ve changed a lot of low-level things around to help the iBCS2\nproject: this includes things like internal segment handling and the\nsignal stack (which now looks the same as on SysV i386 unixes). All in\nall, pl10 has a disturbing amount of new code, but will hopefully work\nwell despite (due to?) the number of changes.\n\nThe new networking code in particular will change the network setup a\nlot - it now looks more standard, but if you were used to the old way of\ndoing things.. On the other hand, most people actively using the\nnetworking features have hopefully gotten warnings about this on the NET\nchannel for the last few weeks. Also, the networking code still isn\u0027t\nperfect: Fred is still working on it, but it seems to have reached a\nreasonably stable platform on which it will be easier to build. Look\nout for the new-and-improved networking manual, hopefully out soon(?).\n\nStandard request: please try it all out, give it a real shakedown, and\nsend comments/bug-reports to the appropriate place (I\u0027m always\nappropriate, but you may want to send the report to the mailing lists\nand/or the newsgroup as well). I apologize for the lateness of the\nrelease (forcing hlu to make interim gcc releases that relied on\nnonstandard kernels etc), and the changes are somewhat bigger than I\u0027d\nprefer, so the more testerts that try it out, the faster we can try to\nfix any possible problems. The new kernel has gone through various\nstages of ALPHA-diffs and some late ALPHA-pl10\u0027s, so there shouldn\u0027t be\nany major surprises, but alpha releases tend not to get even close to\nthe coverage a real release gets...\n\n\t\t\tLinus\n" }, { "commit": "87aa05879df067f26eb0120af913210ff94e32cf", "tree": "f8185b866196eadd03fa9ee26d6442b506b5bfb8", "parents": [ "d2c3ec4b1398e55d918df0d1fa718c9940d373f6" ], "author": { "name": "Linus Torvalds", "email": "torvalds@cc.helsinki.fi", "time": "Sat Apr 24 03:04:54 1993 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:15 2007 -0400" }, "message": "Linux 0.99 patchlevel 9\n\nThe latest kernel release is 0.99.9, and can be found on nic.funet.fi:\npub/OS/Linux/PEOPLE/Linus, both as patches relative to pl8 and as full\nsources. The only major new feature is that the ST-0x driver has\nfinally been updated to the scatter-gather code: ST-0x users should with\nluck get about 5 times the performance on disk-operations.. Seagate\ncode written by Drew Eckhardt.\n\n0.99.9 also fixes:\n - the FPU-emulator should now handle all rounding-modes correctly, and\n pass all the paranoia package tests. Patches by Bill Metzenthen.\n - bootup enhancements by Chrisoph Niemann (but the SVGA mode numbers\n have changed, so you may have to edit your lilo configuration file\n and/or the main Makefile to get the mode you normally want)\n - ext2fs updated to the very latest release. Code by Remy Card and\n Stephen Tweedie.\n - various minor patches, some of them cosmetic, some of them fixes to\n smaller bugs.. Thanks to everybody who sent them in (even though not\n all made it)\n\nIt might be a good idea to test it all out,\n\n\t\tLinus\n" }, { "commit": "d2c3ec4b1398e55d918df0d1fa718c9940d373f6", "tree": "d10540f9ee2a6143e28f9706933b284b7d859ce6", "parents": [ "3ccb23223d9cbee8e7946aa23f44a84ceff88d3c" ], "author": { "name": "Linus Torvalds", "email": "torvalds@cc.helsinki.fi", "time": "Fri Apr 09 14:45:07 1993 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:14 2007 -0400" }, "message": "ANNOUNCE: linux 0.99 patchlevel 8 available\n\nYet another kernel release is now available on nic.funet.fi in the usual\nplace (pub/OS/Linux/PEOPLE/Linus for those of you that have already\nforgotten), and will probably show up on the other ftp-sites within a\nday or two. There are two new files:\n\n linux-0.99.8.tar.z - the full gzipped and tarred source-tree of the\n\tlinux kernel.\n\n linux-0.99.patch8.z - unified diffs against the last official release\n\t(0.99pl7).\n\nThere is no SLIP or new networking routines in this kernel despite the\nrumors that have been flying around - the main changes to 0.99.7 are\n(some of them were in 0.99pl7A as well):\n\n - the signal handling code has been extensively reworked, and should be\n POSIX as well as clean.\n - dosfs is upgraded to version 12 (Werner Almesberger)\n - xiafs is upgraded to the latest version (Qi Xia)\n - ext2fs is upgraded to the latest version (Remy Card/Stephen Tweedie)\n - FPU-emulation patches for v86 mode and precision rounding (Bill\n Metzenthen)\n - SCSI patches by various people (Eric Youngdale \u0026 co)\n - XT harddisk support (Pat Mackinlay)\n - new trial code to try to handle 387 lockups on some systems more\n gracefully.\n - keyboard, lp and serial driver fixes\n - various minor changes (mounting root read-only, bootup messages\n cleaned up etc)\n\nAs always, comments/bugs etc are encouraged,\n\n\t\tLinus\n" }, { "commit": "3ccb23223d9cbee8e7946aa23f44a84ceff88d3c", "tree": "35552324421d578052a6bc9267cae9a49aed3008", "parents": [ "eeb7df9e7d9a3a48dc39743a13e0ed1cc0d5cb41" ], "author": { "name": "Linus Torvalds", "email": "torvalds@cc.helsinki.fi", "time": "Mon Mar 22 21:03:45 1993 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:14 2007 -0400" }, "message": "ALPHA-diffs for 0.99 patchlevel 7\n\nI don\u0027t generally announce ALPHA-diffs to quite this large an audience,\nbut I\u0027ll be partying^H^H^H^H^H^H^H^Hunavailable for the rest of the\nweek, and it\u0027s unlikely that I will be able to check mails or the\nnewsgroups until the start of April. As a result, I\u0027m putting up my\nlatest kernel version for ftp as it fixes some things in 0.99.7.\n\nThe ALPHA-diffs can be found on nic.funet.fi: in the directory\npub/OS/Linux/PEOPLE/Linus. If you dislike patching, you can get the\nfull sources in \"linux-0.99.7A.tar.z\", or just get the diff file\n\"ALPHA-diff.z\".\n\nChanges in this release:\n - the new kernel now detects the lock-up condition at startup if you\n have a faulty 386/387 coupling, and will use software floating point\n in that case.\n - the Xia filesystem is updated to the latest version\n - the DOS filesystem is updated to the latest version\n - the XT disk driver is included: I haven\u0027t been able to test it, but\n at least it won\u0027t bother anybody if you don\u0027t configure it in..\n - the latest serial diffs are in\n - minor ultrastor fixes\n - some changes to the keyboard and line printer drivers: I hope the\n keyboard lockups that some people have reported would be gone with\n this release.\n - some fixes to arp.c\n\nI\u0027ll be interested in success/failure reports, although I won\u0027t be able\nto answer them for some time (and I might miss some of the posts on\nc.o.l).\n\n\t\tLinus\n" }, { "commit": "eeb7df9e7d9a3a48dc39743a13e0ed1cc0d5cb41", "tree": "64a9699b70d1d26ec9fce28c108d2cd07bccf379", "parents": [ "56cac397826c13fb61390c9ca3fa5354a7b80ca5" ], "author": { "name": "Linus Torvalds", "email": "torvalds@cc.helsinki.fi", "time": "Sat Mar 13 19:37:57 1993 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:13 2007 -0400" }, "message": "Linux 0.99 patchlevel 7 available..\n\nIt has been two weeks since the last release, so it\u0027s high time you\nshould once more enjoy the pleasures of patching up your kernel to a\nhigher version number if you are into those kinds of perversions. Linux\n0.99pl7 is available as both full source and diffs against pl6 on\nnic.funet.fi: pub/OS/Linux/PEOPLE/Linus, and it will probably show up on\nthe other major sites within days.\n\nAs of pl7, I\u0027m trying out a new format: both the full distribution and\nthe diffs are now compressed with gzip as it is now available at most\nmachines. Also, the diffs are no longer context diffs: they use the\nsmaller unified diff format. At least the stock SunOS \u0027patch\u0027 binary\nseems not to understand them at all, but GNU patch has no problems, and\nunified diffs are a bit smaller (not that it matters much after gzip has\ndone its deed on them).\n\nAs to the changes in pl7: they are many and varied, and hopefully all to\nthe better (-\"Dream on Linus\" -\"Shut up\"). Short list follows, hope I\nhaven\u0027t forgotten anything major.\n\n - ext2fs is in: note that this is version 0.2c and that if you are\n currently using an older version there are some changes. Small\n filesystems (\u003c 256MB) should reportedly be automatically converted,\n bigger filesystems need some assistance. Ext2fs written by Remy Card.\n - xiafs is also in: again, the final version uses a slightly different\n layout to support exact file block counts, so if you use the xiafs,\n you should make sure you have the latest fs-tools. Xiafs written by\n Frank Xia.\n - updated Ultrastor SCSI driver with scatter/gather by Scott Taylor.\n It should be much faster, as well as support the Ultrastor-34F.\n - major changes in the memory manager. Yours truly got carried away,\n and finally cleaned up the mm layer due to pmacdona wanting mmap() on\n /dev/zero. This means that the IPC patches won\u0027t go in, and need\n updating. Krishna?\n - more big changes: I rewrote most of the VFS filename-handling.\n Filenames are copied into kernel space before being used, which\n cleaned things up somewhat, as well as simplifying some race-\n condition handling. As a result, I was also able to easily expand\n the minix fs to cover the \"linux\" fs that some people have been using\n (same layout, but with 30-character names).\n - updated the printer driver: Nigel Gamble. It is now able to use\n interrupts, although the default behaviour is still to poll.\n - serial driver updates by tytso (but no SLIP yet)\n - various minor patches for POSIX compliace: Bruce Evans, Rick Sladkey\n and me.\n - other minor patches all over the place: scsi, tcpip etc.\n\nAll in all, the patches are almost half a megabyte even as unified\ndiffs: getting the full sources might be easier than patching it all up.\n\nAs always, some of the patches are actually tested by me, some aren\u0027t\n(and just because I wrote some of them doesn\u0027t mean I actually *tested*\nthem: I have no idea if mmap() works on /dev/zero, although it should).\nI have neither a printer nor an Ultrastor controller, and I haven\u0027t got\nthe diskspace to test out the new filesystems, so I can only hope they\nwork \"as advertized\". If you have problems, I want to hear about them,\nso keep the reports coming, and try to pinpoint the problem as well as\nyou can (\"when I do *this* it happens every time..\").\n\n\t\tLinus\n" }, { "commit": "56cac397826c13fb61390c9ca3fa5354a7b80ca5", "tree": "bfea3a2a0e1e187ee4a34b2a8ef082488889103d", "parents": [ "1680d965ceb874b34ebbc3788fdb1b0dd594f043" ], "author": { "name": "Linus Torvalds", "email": "torvalds@cc.helsinki.fi", "time": "Mon Feb 22 21:22:01 1993 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:12 2007 -0400" }, "message": "linux 0.99 patchlevel 6\n\nI\u0027m starting soon to run out of patchlevel numbers for 0.99, but I made\na new release anyway (and long-time linux hackers remember my less than\ncompletely logical numbering: when I run out of numbers I\u0027ll start using\nalphabetical characters and other fun characters to indicate new\nversions :-)\n\n0.99pl6 is mainly a syncronization release: it fixes a few bugs and\nchanges the behaviour of \u0027vhangup()\u0027 to be more standard. The vhangup()\nchanges will break some init/login stuff that depended on the earlier\nincorrect behaviour - not everybody may want to use pl6 until you are\nsure your init/login will work happily with it. Better do these things\nbefore 1.0 than to break it later.\n\nPatchlevel 6 also changes the vfs functions for special devices as well\nas adding a \u0027fsync\u0027 field to the inode-operations structure. Thus\next2fs and xfs need updating. Remy and Xia? The special file and fifo\nhandling code is no longer supposed to be in the fs-dependent layer, but\nis handled by the vfs routines, as it\u0027s the same for all \"normal\"\nfilesystems.\n\nOk, here are the actual changes/features of pl6:\n - the kernel can be loaded in gzipped format and de-compressed at\n startup beyond the 1MB mark. Good for bootable rootdisks. Patches\n mainly by Hannu Savolainen.\n - I finally enabled NMI\u0027s everywhere (except at the bootup sequence),\n so if you have memory errors, they will hopefully now result in\n kernel messages (\"NMI received..\")\n - the device registration code for special devices. Special files are\n now registered with a special \"register_[chr|blk]dev()\" function.\n - consolidated fifo/special dev handling\n - vhangup patches. Note that these may break init/login badly, at\n least if you are using poeigl-1.7. Be careful that you don\u0027t get\n totally locked out of your machine.\n - the procfs NULL-dereferencing bugfix (michaelkjohnson)\n - literal next character handling (very losely based on a patch I\n received: I essentially rewrote it with final fixes by jrs).\n - fpu-emu bugfixes by Bill Metzenthen - fixes the \"internal error 112\"\n bug as well as a sign bug with zero.\n - fdomain driver fixes\n - various other minor fixes (wrongly replying to bad ip messages etc)\n\nI\u0027m still not sure about the 387 error detection code: I have had a\ncouple of messages that would suggest that some early clone 387\u0027s have\nproblems with math exceptions in protected mode. With the new (as of\n99pl5) test at startup this can lead to problems at boot-time. Please\nmail me directly if you seem to have problems with this (it should be\nobvious in pl6 due to debugging messages at startup).\n\n\t\tLinus\n" }, { "commit": "1680d965ceb874b34ebbc3788fdb1b0dd594f043", "tree": "4ec1dc6b7651f2e57cff25c70c4b10663d74c345", "parents": [ "b75fb8468bb5353db99f69805408db991ac0ed39" ], "author": { "name": "Linus Torvalds", "email": "torvalds@cc.helsinki.fi", "time": "Tue Feb 09 22:03:01 1993 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:12 2007 -0400" }, "message": "ANNOUNCE: linux 0.99 patchelevel 5 available\n\n\t\"He\u0027s done it yet again - doesn\u0027t he ever rest?\"\n\t\t\t- anonymous linux kernel hacker\n\nOnly complete newbies don\u0027t know what this is all about, but I\u0027d better\ntell you anyway: patchlevel 5 of the 0.99 kernel is now available on\nnic.funet.fi (pub/OS/Linux/PEOPLE/Linus) as both context diffs against\npl4 and complete source code. I\u0027m not even going to speculate on 1.0\nright now.\n\nThe pl5 diffs are about 90kB compressed: the major changes are to the\ntcp/ip code and the serial driver, while there are various minor fixes\nstrewn around the system:\n\n - serial lines/tty changes (tytso \u0026 Fred v Kempen)\n - NFS bugfixes (Rick Sladkey)\n - tcp/ip (Ross Biro)\n - coprocessor handling changes (me)\n - harddisk driver error handling (Mika Liljeberg)\n - various minor patches (me and others)\n\nSerial lines now implement non-blocking opens correctly and support\ndial-out lines (same minor, major\u003d\u003d5). I changed the default startup\nmode to be CLOCAL so that people won\u0027t get confused by the modem line\ncode when not using dial-in.\n\nAnother interesting change is the 387 error-coupling tests at bootup:\nthe code to check if the intel-recommended exception 16 error reporting\nis present is \"non-obvious\". If you have had problems with coprocessor\nerror handling, or have a non-intel coprocessor, I\u0027d suggest you test\nthis out: I\u0027d like to hear about problems/successes.\n\n\t\tLinus\n\nPS. If you tested out the latest ALPHA-diffs (the ones that already\nchanged the kernel version to pl5), the changes to the final pl5 were\nonly cosmetic.\n" }, { "commit": "b75fb8468bb5353db99f69805408db991ac0ed39", "tree": "69196f6614fcca6352ab66b02c1d70caf1b89f36", "parents": [ "184df1a3d0f315a5cf4836d1098d6d9531268e02" ], "author": { "name": "Linus Torvalds", "email": "torvalds@klaava.Helsinki.FI", "time": "Fri Jan 22 12:52:47 1993 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:12 2007 -0400" }, "message": "Re: Freeze up on X\n\nIn article \u003c1993Jan21.181502.23485@miles.com\u003e dennisf@miles.com (Dennis Flaherty) writes:\n\u003e\n\u003eHere\u0027s another clue. Try this: when your system freezes, running X, try\n\u003eMOVING THE MOUSE. It\u0027s weird!! But moving the mouse actually makes the\n\u003esystem run! Stop moving the mouse, and the system freezes again. And\n\u003ethis only happens with 0.99.3, not 0.99.2.\n\nGet pl4, and it should be gone. There was a bug in the handling of\nuninitialized interrupts in pl3, where they could result in either the\nwrong interrupt mask being loaded leading to interrupt lock-out or (in\nsome cases) bit corruption at the user level. The symptoms are exactly\nas you describe: a good interrupt that didn\u0027t happen to be locked out\nwill correct the interrupt mask, and the system goes on (it can be\nmoving the mouse, but it might also be a keyboard event etc).\n\n\t\tLinus\n" }, { "commit": "184df1a3d0f315a5cf4836d1098d6d9531268e02", "tree": "de592f1dbc7d49c811d3efa2c309154160f48721", "parents": [ "533790deaad93ea1eacf12b7831c42ff0bb45a61" ], "author": { "name": "Linus Torvalds", "email": "torvalds@cc.helsinki.fi", "time": "Wed Jan 13 22:05:45 1993 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:11 2007 -0400" }, "message": "ANNOUNCE: linux-0.99 patchlevel 3\n\nStill no 1.0 - I have had a couple of reports of problems, so I\u0027ll make\nyet another 0.99 release. The diffs (against 0.99.2) and complete\nsource can be found at nic.funet.fi: pub/OS/Linux/PEOPLE/Linus as usual,\nand will probably show up at the other sites pretty soon.\n\n0.99.3 contains no real new features, but the diffs are pretty big\nanyway (100kB+ compressed): various things have moved around a bit and\nthere are a lot of minor changes. The changes include (but are not\nlimited to):\n\n - the math emulator code now also understands the unofficial codes (in\n case somebody followed the ML math emulator thread). I\u0027d be\n interested to hear whether ML now works with the emulator.\n - various SCSI driver changes\n - some re-organization of the tty open/close code to remove a few race\n conditions.\n - interrupt handling rewrites (two-level interrupt code cleanups)\n - the serial drivers are tytso\u0027s alpha-drivers: they aren\u0027t quite\n completed, but as they need the interrupt handling patches to get\n ready, this is probably the least traumatic way of doing it.\n - some more minor keyboard driver changes (mostly taking advantage of\n the two-level interrupts)\n\n+ a lot of other minor changes. I once more hope people will try it\nout, and report any problems or successes to me.\n\nKnown problems:\n - there seems to be something weird going on in the ST-0x driver with\n some scsi disks.\n - tcp/ip is reportedly still not quite stable, and I can\u0027t even test it\n out.\n\nNOTE! The DMA functions have changed for the high DMA channels - all DMA\nfunctions now take their arguments as the number of bytes instead of the\nold way of using bytes for ch 0-3 and words for ch 5-7. This might lead\nto problems with the SoundBlaster driver, which may need editing.\n\n\t\tLinus\n" }, { "commit": "533790deaad93ea1eacf12b7831c42ff0bb45a61", "tree": "1b178ed40798bc683f26508a98ea709c3589e00f", "parents": [ "d99dedeec6209a2772ff0525eea13509181b9a2c" ], "author": { "name": "Linus Torvalds", "email": "torvalds@cc.helsinki.fi", "time": "Fri Jan 01 22:38:37 1993 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:11 2007 -0400" }, "message": "linux-0.99 patchlevel 2\n\nYes, as you\u0027ve probably noticed, it\u0027s now 1993 and I still haven\u0027t\nreleased 1.0. Sorry about that, and I have only another patchlevel to\noffer. The new kernel should mainly fix some of the keyboard problems\npeople have experienced, but does contain some other minor fixes.\n\nLinux 0.99.2 is available now at nic.funet.fi: pub/OS/Linux/PEOPLE/Linus\nas both sources and diffs against 0.99.1 the diffs are essentially the\nsame as the second alpha-diffs I released for limited testing, with only\nminor fixes to fs/exec.c and fs/open.c.\n\nPlease try out 0.99.2: the more feedback (hopefully positive) I get on\nit, the faster 1.0 will be out.\n\nChanges from pl1 are mainly:\n - pretty much rewritten low-level keyboard handling IO - this time\n actually trying to do it by the book. It now handles resend requests\n from the keyboard etc.\n - you can run executables from filesystems without bmap support. This\n mainly means NFS and msdos. Note that while it\u0027s possible, it\u0027s\n slower and less memory-efficient than using a \"normal\" linux\n filesystem, and should generally be avoided.\n - /proc filesystem changes: /proc/kmsg can be used to log the kernel\n messages under X11 (instead of using the older system call to do the\n same), and there are changes to the statistics routines (WCHAN).\n\n+ various minor fixes (non-existent devices are handled better, some\nchanges to socket bind behaviour etc).\n\n\t\tLinus\n" }, { "commit": "d99dedeec6209a2772ff0525eea13509181b9a2c", "tree": "e2ffc8c2f9d26ed297ea5e695fb2ffc35123ef24", "parents": [ "ca87352f20470de289c53a58c877df452e8b982c" ], "author": { "name": "Linus Torvalds", "email": "torvalds@cc.helsinki.fi", "time": "Mon Dec 21 23:36:46 1992 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:11 2007 -0400" }, "message": "ANNOUNCE: linux 0.99 patchlevel 1\n\nLinux-0.99.1 is now available on nic.funet.fi, and will probably show up\non the other linux ftp sites within days, unless everybody has gone off\nfor the holidays. The directory is /pub/OS/Linux/PEOPLE/Linus, and it\u0027s\navailable both as a patch against 0.99 and as complete source. The\npatch is pretty clean, although most people probably have changed 0.99\nslightly to get rid of the setup and/or inode.c problems, so if you\nhave, you\u0027ll have to revert or patch by hand.\n\nPatch 1 addresses the following problems:\n - configuration. Hope there are no silly problems left..\n - inode.c: initialization changes (the missing NULL and some other\n minor fixes).\n - some SCSI tape driver patches (Kai M{kisara)\n - tcp/ip patches (Ross Biro, some code by me)\n - keyboard patches (mainly changed initialization - hope the keyboard\n lockups are gone).\n - completed /proc-fs: it should now contain all info needed by \u0027ps\u0027\n (Micheal K Johnson).\n - various minor fixes (the minix-fs link overflow checking etc)\n\nPatch1 also contains support for extended VC switching - this is for the\nupcoming X11 that understands VC\u0027s. One result of this is that console\nredirection now redirects *only* messages actually sent to /dev/console\n(aka /dev/tty0), not just to any foreground VC. Wait for Xfree-1.2 to\nbe able to switch VC\u0027s while under X (yes, including several X-sessions\nactive at the same time..).\n\nI hope there are still people out there that aren\u0027t too busy stuffing\nthemself with turkey to try out a new kernel release. There is just\nover a week left of this year, and I need feedback in order to be able\nto release 1.0.\n\n\t\tLinus\n\nPS. Thanks to everybody who has sent me Christmas/New Year/Birthday\ncards. Some contained money, some didn\u0027t, and I enjoyed them all.\nThanks.\n" }, { "commit": "ca87352f20470de289c53a58c877df452e8b982c", "tree": "c88656d2481bffb6c03617ae22910aac6454fa9e", "parents": [ "3412f7ea7c0414a6191cf5b5d474e8a7f6137fc5" ], "author": { "name": "Linus Torvalds", "email": "torvalds@cc.helsinki.fi", "time": "Sun Dec 13 19:38:12 1992 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:10 2007 -0400" }, "message": "[ANNOUNCE]: linux version 0.99\n\nLinux version 0.99 is now available at nic.funet.fi, in the directory\npub/OS/Linux/PEOPLE/Linus as both full source and patches against\n0.98.6. It will probably show up on the other major sites soon.\n\nNOTE!! The context diffs aren\u0027t very good. The makefiles have changed,\nand due to the new setup they changed radically enough that I couldn\u0027t\nedit away the dependancy changes like I usually do. This means that the\ndiffs are unlikely to patch cleanly, and I\u0027d suggest people get the full\nsource unless you feel comfortable about patching by hand. Also, if you\nget the full source, I\u0027d suggest you remove (or move elsewhere) all of\nthe old kernel to make sure you don\u0027t have any dead files from older\nversions lying around.\n\n0.99 has no major new features: the NFS client code is now in the\nstandard distribution, and the kernel configuration has changed, but\nmost of the rest of the changes are fixes - especially the tcp code\nshould now be pretty stable (knock wood).\n\nChanges:\n\n - NFS is in. As are some stubs for the soud drivers, although it\u0027s only\n stubs right now.\n - various fixes around the place: the serial problems are hopefully\n gone, and there are patches to both TCP/IP and SCSI to make them more\n stable.\n - Minor fixes: the keyboard buglet introduced in 0.98pl6 should be\n gone, and some other bugs are also corrected. The optimized\n read-ahead code in the filesystems (and the raw device read code) was\n too complicated and seemed to have problems with bad blocks, so I\n rewrote it, and it should hopefully work correctly now (this may have\n been the reason \"mkfs -c\" didn\u0027t work in all cases). Thanks for some\n good bug-reports I\u0027ve gotten: I\u0027ve tried to correct all the problems\n I got reports on.\n - The kernel configuration has been re-thought: I decided to take\n advantage of the possibilities offered by GNU make etc. This means\n that you no longer can compile the kernel using any other make, but\n there probably aren\u0027t many (if any) people doing that anyway. This\n way I got rid of the extremely ugly SCSI setup, so it was probably\n worth it.\n\nTo configure the kernel for your setup, do a\n\n\tmake config\n\nand answer the yes/no questions. After that, do a\n\n\tmake dep\n\nto make the dependencies match your setup. After that you should still\ngo edit the top-level Makefile for some of the configuration information\nas before, but the remaining config things are pretty simple. Then you\ncan make the kernel with a simple \"make Image\".\n\nThe new configuration utility (essentially a stupid shell script coupled\nwith some smarts in the Makefiles) tries to minimize compilations: if\nyou disable the SCSI code the scsi drivers won\u0027t even be compiled, much\nless linked in. This should be a win on slower machines.\n\nNOTE!!! Use LILO-0.7 to load the 0.98pl5 and newer kernels: any older\nversion of lilo is liable to result in weird problems.\n\n\t\tLinus\n" }, { "commit": "3412f7ea7c0414a6191cf5b5d474e8a7f6137fc5", "tree": "b61b0be7582787bed7155005e1e27102104b17df", "parents": [ "cbfd22851e4393ec3b1c9975f522fb3e01b204cf" ], "author": { "name": "Linus Torvalds", "email": "torvalds@klaava.Helsinki.FI", "time": "Wed Dec 02 21:24:46 1992 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:10 2007 -0400" }, "message": "ANNOUNCE: linux-0.98 patchlevel 6\n\nYou all know what the above subject-line means by now...\n\nCurrently only the full sources to 0.98pl6 are available (at\nnic.funet.fi in the directory pub/OS/Linux/PEOPLE/Linus), but I\u0027ll\nprobably make context-diffs later today. It depends a bit on the size\nof the diffs: the changes in pl6 aren\u0027t very suitable for diffing\n(renaming of some SCSI files etc).\n\nAnyway, 0.98pl6 is hopefully the last release before 0.99: there are a\nfew known problems left in this release. Most notable is the serial\ncode: it works for most people, but others still have problems with it.\nI hope this will get fixed within a week (tytso is working on it). It\nalso seems as if the PS/2 mouse code has some problems.\n\npl6 contains these fixes:\n\n - all the tcp/ip patches I\u0027ve received (and I fixed one bug that\n gcc-2.3 seems to have found).\n - math-emu patch for the problem that resulted in FPU errors with some\n operations.\n - I fixed gcc-2.3 warnings as well as most of the old warnings. You\n shouldn\u0027t get more than one or two warnings when recompiling the\n whole kernel.\n - /proc filesystem extensions. Based on ideas (and some code) by\n Darren Senn, but mostly written by yours truly. More about that\n later.\n - some tty_io fixes (there was a bug in the /dev/console handling when\n you changed VC\u0027s while using the general console device).\n - re-organization of the keyboard-driver internal data-structures. The\n changes are mostly preliminary: they change the keyboard flags to be\n more easily adaptive to a reprogrammable keyboard driver. No actual\n new features yet.\n - new SCSI drivers: reportedly much faster than the old ones (but not\n all drivers take advantage of it yet..)\n - various other fixes: pty\u0027s etc have minor changes.\n\nI hope to make 0.99 in a week or so, and 1.0 after that has been tested\nsome. I hope people will test out pl6 - 0.99 won\u0027t be much different,\nand if you don\u0027t test pl6, any bugs relating to your particular hardware\nmay not be found in time for 0.99...\n\n\t\tLinus\n\n\u003d\u003d\u003d\u003d\u003d /proc filesystem comments, ignore if not interested \u003d\u003d\u003d\u003d\u003d\n\nIf people want to test out the new /proc filesystem features, please do\n\n # mount -t proc /proc /proc\n\n(or add the line \"/proc /proc proc defaults 0 0\" to your /etc/mtab). You\nobviously need a /proc subdirectory for the above.\n\nAfter mounting the proc fs, your /proc should look something like this:\n\n # ls -l /proc\n ....\n dr-xr-xr-x 4 root root 0 Dec 2 22:55 936\n dr-xr-xr-x 4 root root 0 Dec 2 22:55 983\n -r--r--r-- 1 root root 0 Dec 2 22:55 loadavg\n -r--r--r-- 1 root root 0 Dec 2 22:55 meminfo\n -r--r--r-- 1 root root 0 Dec 2 22:55 uptime\n\nThe numeric entries correspond with the pid\u0027s of the running procedures,\nwhile the alphabetic entries are vaious general info-files (don\u0027t be\nfooled by the zero length).\n\n # cat /proc/meminfo\n\t total: used: free: shared: buffers:\n Mem: 15831040 15601664 229376 3952640 8904704\n Swap: 5521408 0 5521408\n # cat /proc/loadavg\n 0.01 0.02 0.00\n # cat /proc/uptime\n 12981\n\nThe info should be pretty self-evident (uptime is in seconds, and yes, I\nreboot pretty often).\n\nFor each \u003cpid\u003e in the system, the per-process directory looks like this:\n\n # ls -l /proc/50/\n total 3\n -r--r--r-- 1 root root 0 Dec 2 23:00 cmdline\n lrwx------ 1 root root 3 Dec 2 23:00 cwd -\u003e ---\n -r--r--r-- 1 root root 0 Dec 2 23:00 environ\n lrwx------ 1 root root 3 Dec 2 23:00 exe -\u003e ---\n dr-x------ 2 root root 0 Dec 2 23:00 fd\n dr-x------ 2 root root 0 Dec 2 23:00 lib\n crw------- 1 root root 1, 1 Dec 2 23:00 mem\n lrwx------ 1 root root 3 Dec 2 23:00 root -\u003e ---\n -r--r--r-- 1 root root 0 Dec 2 23:00 stat\n\nThe most interesting entries are probably the new ones: cmdline, environ\nand stat. They give the process command-line, environment and some\nstatistics respectively. The file protections will probably have to be\nchanged.\n\n # cat /proc/50/cmdline | tr \u0027\\000\u0027 \u0027 \u0027 \u0026\u0026 echo\n X :0 -pn\n # cat /proc/50/environ | tr \u0027\\000\u0027 \u0027 \u0027 \u0026\u0026 echo\n SHELL\u003d/bin/sh SHLVL\u003d2 BASH\u003d/bin/sh MAIL\u003d/var/spool/mail/root\n HOME\u003d/usr/root PATH\u003d/bin:/usr/bin:/etc:/usr/local/bin:/usr/bin/X11\n LOGNAME\u003droot LESS\u003d-MM TERM\u003dconsole ignoreeof\u003d10 HOSTTYPE\u003di386 DISPLAY\u003d:0\n _\u003d/usr/bin/X11/xinit\n\nThe \u0027tr\u0027 and \u0027echo\u0027 commands are there to make it look a bit nicer: the\ncommand line arguments and environment variables are separated by \u0027\\0\u0027\ncharacters and aren\u0027t terminated by a newline. Note that the /proc\nfilesystem doesn\u0027t try to fetch command line data that is swapped out,\nso when you are swapping, one or both of the above may be empty or\ntruncated. Also, the code I wrote to fetch the above may nor be\ncomplete yet, and doesn\u0027t seem to be able to find all of the cmdline all\nof the time..\n\nFinally there is the (not completely implemented) process status file:\n\n # cat /proc/50/stat\n 50 (X) R 49 50 10 1\n\nThis currently gives the pid, short name (\"X\", max 8 chars, but you get\nit even when the process is swapped out), the state (R for running etc),\nppid (49), pgrp (50), session (10) and controlling terminal (1 -\n/dev/tty1). The status file will have to be extended to tell you more\nabout the process, but that should be easy now that the general code is\nthere..\n" }, { "commit": "cbfd22851e4393ec3b1c9975f522fb3e01b204cf", "tree": "0b5e038045eea74979ecb6858ee42deb085565cd", "parents": [ "1fd95ba642b4a2cdfc947bbc0bf02a05d1e4bef2" ], "author": { "name": "Linus Torvalds", "email": "torvalds@klaava.Helsinki.FI", "time": "Sun Nov 15 22:01:38 1992 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:10 2007 -0400" }, "message": "ANNOUNCE: Linux 0.98 pl5\n\nYou all know what it\u0027s about by now: YAKR (yet another kernel release..)\nlinux 0.98 patchlevel 5 is now available at nic.funet.fi as both full\nsource and as context diffs against 0.98.4. The place to look is (as\nbefore) pub/OS/Linux/PEOPLE/Linus.\n\n0.98.5 mainly fixes the swap-partition bug that was present in pl4 (and\nfor which I did an earlier unofficial emergency patch). The bug\nresulted in incorrect swapping with a partition under some circumstances\n(notably tty events: keypresses could make xterm dump code when swapping\nwas enabled etc).\n\npl5 also has some other changes - nothing major. Setting and querying\ntermios information from a pty master will now set/query the slave info:\nthis seems to be what some programs (telnet) expect. I haven\u0027t seen any\nchanges to any of the programs I use, but I\u0027d like to hear if this\nresults in problems or if it actually does help.\n\nNOTE! READ THIS AND PONDER:\n\npl5 now checks against writing to the text segment. Older binaries\nwhich used the original estdio library (used with the earliest gcc\nversions) are liable to break: not that there should be many of these\nbinaries around. So if you get \"Segmentation fault (core dumped)\" on\nbinaries you know used to work, this is the likely cause.\n\nOne problem spot that I\u0027ve seen even with new binaries is due to a\nlibrary bug in \u0027sigaction()\u0027. If the second argument is NULL (ie the\npointer to the new sigaction structure), sigaction() will incorrectly\ndereference it resulting in a core-dump. The only program so far that\nI\u0027ve seen doing this is \u0027dd\u0027, but there may be others.\n\nOn my system I have found a whopping total of two binaries which didn\u0027t\nlike the text segment protection, so it shouldn\u0027t really be a major\nproblem for anybody. Famous last words.\n\n\t\tLinus\n\nPS. The strace code in pl4 was incorrectly credited in the announcement.\nThe code was written by Branko Lankester, not Ross Biro (who did the\ntcp/ip changes).\n" }, { "commit": "1fd95ba642b4a2cdfc947bbc0bf02a05d1e4bef2", "tree": "86a779933756ecde8bdfd354f8cec4e24589f454", "parents": [ "73795752dade564a88db44bdcfd7c5a946322de1" ], "author": { "name": "Linus Torvalds", "email": "torvalds@klaava.Helsinki.FI", "time": "Mon Nov 09 11:22:01 1992 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:09 2007 -0400" }, "message": "Linux version 0.98-pl4\n\nAs the subject probably tells you, pl4 is out. It\u0027s currently available\nat least on nic.funet.fi: pub/OS/Linux/PEOPLE/Linus/, but will probably\nshow up on the other major sites very soon.\n\npl4 is available both as complete source (linux-0.98.4.tar.Z) and as\npatches against pl3 (linux-0.98.patch4.Z) and against the pre-version\nthat was released for limited testing (pre-patch4.Z or something like\nthat). Things that have changed:\n\n - the inode caching bug (resulting in bad filesystem info when\n mounting/umounting devices) should be gone for good.\n\n - an elusive race-condition in the fs is fixed: this may have been the\n reason some people got fsck errors once in a while. The\n race-condition was pretty hard to find, and depends on a lot of\n things (buffer cache size, speed of the disk and computer speed).\n\n - fpu emulator patches (mainly for the re-entrancy problem) by me and\n W. Metzenthen.\n\n - various wait-queue changes - the kernel uses the waiting mechanism\n more efficiently now.\n\n - the NFS client support code is there: the actual nfs code is still in\n alpha (although reported to be pretty stable) and has to be gotten\n separately.\n\n - NR_OPEN was changed from 32 to 256 (which is what SunOS seems to use,\n so I hope it won\u0027t need any further changes). This has lead to some\n incompatibilities (GNU emacs and the term program seem to need\n recompilation to work correctly), as the \u0027select()\u0027 system call has a\n slightly changed interface due to the new fd_set definition.\n\n - the process kernel stack is now on a separate page (needed due to the\n fact that the task_struct has now grown to almost 3kB due to the\n NR_OPEN changes). This also means \u0027ps\u0027 needs patches.. My patches\n to ps-0.98 are available as \u0027ps-diff.Z\u0027 in the same directory as the\n kernel sources and diffs.\n\n - various other changes: system call tracing by Ross Biro. Changed\n ll_rw_block interface (performance reasons: it will eventually be\n changed to accept several requests at once). Malloc() was changed\n and renamed to kmalloc() due to the new interface. Some tcp/ip\n patches (inode counting correction and some other changes).\n\n0.98.4 should hopefully be pretty stable: the main problem areas are\nprobably still tcp/ip and some of the tty code. I\u0027d appreciate\ncomments, bug-reports etc.\n\n\t\tLinus\n" }, { "commit": "73795752dade564a88db44bdcfd7c5a946322de1", "tree": "7503515d32b5104c3c20fb6710ff69e167980e07", "parents": [ "4c212a02429139af48301686ce7cd24ce1a0b47f" ], "author": { "name": "Linus Torvalds", "email": "torvalds@klaava.Helsinki.FI", "time": "Tue Oct 27 19:49:52 1992 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:09 2007 -0400" }, "message": "ANNOUNCE: linux-0.98.3\n\nOk, I already sent out an announcement last night, but due to the time\n(6AM over here) I wasn\u0027t really in a mood to write a real annoucement.\nHere it is.\n\nlinux-0.98.3 is available by anonymous ftp at least on nic.funet.fi:\npub/OS/Linux/testing/Linus, both as context diffs against 0.98.2 and the\npre-version of 0.98.3 and as complete source. The complete source\npackage was done by directly applying the diffs - this means that the\nMakefile dependancies are probably not 100% up-to-date as I remove those\nfrom the diffs. It shouldn\u0027t be any problem, and you can always do a\n\"make dep ; make clean\" before actually compiling the kernel.\n\n0.98 pl3 fixes several bugs, and should remove all known NULL-pointer\nproblems that made 0.98.2 unusable for most people. In addition to the\nNULL pointer fixes, the following things have changed:\n\n - removed most of the cli-sti pairs in the filesystem code by rewriting\n the locking routines to use a different algorithm, possible due to\n the rewritten wait-queue code that I did back in 0.96c or so.\n Interrupt latency should be better on slow machines, but I don\u0027t know\n if it\u0027s noticeable.\n - Minor 387-emulation fixes by Bill Metzenthen - only noticeable under\n special conditions.\n - Corrected various error-returns in the fs (thanks to Bruce Evans for\n running some error diagnostics). Error messages when opening (and\n renaming etc) files that had a non-directory in the path were wrong,\n and should be ok now (ie giving ENOTDIR instead of EACCESS or ENOENT).\n Some other problems reported by Bruce fixed.\n - Changed the interface for some fs-related functions due to cleaning\n up super-block handling. Most noticeably, iget() and related\n functions no longer specify the inode with a device and inode number,\n but instead with a super-block pointer and inode number. This is\n more logical, and should make unnamed devices (ie internal\n filesystems like nfs and /proc) cleaner. Also note that the calling\n sequence for sb-\u003es_op-\u003eput_inode() also has changed since 0.98. This\n is of interest only if you are writing filesystem drivers..\n - ASK_SVGA was broken in 0.98.2 - it should be ok now.\n\nAlso, various minor fixes as usual. No new features, but I hope 0.98.3\nwill be a lot less bug-prone due to the changes since 0.98.1. Some\nminor tcp/ip corrections (but most of them were in the pre-release), and\nI removed a race-condition in the tty-handling code.\n\nNote that people who use math without a co-processor should certainly\nupgrade to 0.98.3: the new emulator is much better than my original one\nboth in speed and accuracy/exception handling. x11perf is very much\nbearable now even without a 387, and things like ray-tracing etc\nshouldn\u0027t be a problem any more. It\u0027s slower than hardware fp, of\ncourse, but at least it works.\n\nThe new emulator also means there is no reason for a separate soft-float\nlibrary, so I\u0027d assume that will be gone in the next gcc release for\nlinux.\n\nAs usual, a new kernel version probably means you\u0027ll have to recompile\n\u0027ps\u0027 and friends. But at least the same \u0027ps\u0027 sources that worked for\n0.97.6 should still work.\n\n\t\tLinus\n\n-----\n\nFrom: torvalds@klaava.Helsinki.FI (Linus Torvalds)\nSubject: 0.98.3 finally available\nDate: 27 Oct 1992 04:01:01 GMT\n\n0.98.3 is available at nic.funet.fi: pub/OS/Linux/testing/Linus as diffs\nagainst 0.98.2 (linux-0.98.patch3.Z) and against the pre-release that\nwas announced on the mailing-lists (pre-final.diff.Z). I haven\u0027t\nuploaded the complete sources yet, so diffs are all there is: I\u0027ll\nupload the full sources once I\u0027ve gotten some more sleep.\n\n0.98.3 corrects a lot of NULL-pointer bugs, as well as some other\nproblems (notably serial lines). Hope this release will be stable,\n\n\t\tLinus\n\n-----\n\nFrom: torvalds@klaava.Helsinki.FI (Linus Torvalds)\nSubject: 0.98.3 is late ...\nDate: 26 Oct 1992 05:07:26 GMT\n\n ... as you probably all have noticed. I\u0027ve been up all night writing\nthis report for the \"CS Course from Hell\" (*), and haven\u0027t had time to\nplay with linux. I\u0027ll get it done tonight if I can keep awake that\nlong.\n\nIn case anybody still wonders, 0.98.3 mainly fixes the NULL pointer and\nmost of the serial line problems. I also started to remove some of the\nmore unnecessary cli-sti pairs (where races could be avoided by doing it\nsome other way), so it\u0027s possible interrupt latency is better, but I\ndoubt it\u0027s noticeable.\n\n\t\tLinus\n\n(*) Well, they call it \"ATK-suunnittelun laboratorioty|\", but they\naren\u0027t fooling anybody.\n\n-----\n\nFrom: torvalds@klaava.Helsinki.FI (Linus Torvalds)\nSubject: Re: ANNOUNCE: 0.98.pl2 is out\nDate: 18 Oct 1992 22:42:02 GMT\n\nIn article \u003c1992Oct18.144546.28249@klaava.Helsinki.FI\u003e I wrote:\n\u003e [ deleted ]\n\u003e\n\u003e - various minor mm fixes by me: trapping kernel NULL dereferences,\n...\n\nOh, boy. Talk about a feature that proved itself. I have never gotten\nso many bug-reports so fast: it seems there were quite a few\nNULL-pointer problems in the code I hadn\u0027t been able to test. Thus\n0.98.2 isn\u0027t actually very useable unless you have a system very close\nto my setup - I don\u0027t recommend the faint-of-hearted to try it out.\n\nI guess I should be happy the new feature found problems in the kernel\nsources, but I could well have lived with a couple less problems. Main\nproblem areas seem to be:\n\n - the vhangup() system call - used by newer login-binaries, and\n resulting in login getting killed.\n - some scsi driver problems (but they should be fixed by the patch eric\n already sent out)\n - tcp/ip problems - at least three different places where NULL gets\n dereferenced (and one is particularly ugly: it does test for NULL,\n but does so *after* dereferencing it...)\n\nThe vhangup() problem was easy to correct (and excusable: it broke at\nleast partly due to the new dynamic tty-routines). I\u0027ll make some\npreliminary testing-patches available that fix that problem, and\nincludes the scsi fixes.\n\nBut the tcp/ip code is a bit worse: I fixed two of the known problems,\nbut I\u0027d rather use patches from the persons who wrote the code\noriginally, as there are some other problems in there.. And I can\u0027t be\nsure I find them all anyway, as I\u0027ll never see the problems first-hand.\n\nJust wanted to warn everybody about this aspect of pl2 - the\nNULL-pointer finding code is working a bit too well for my taste. I had\nhoped there wouldn\u0027t be quite this many NULL pointer problems, but at\nleast they\u0027ll be found this way.\n\n\t\tLinus\n" }, { "commit": "4c212a02429139af48301686ce7cd24ce1a0b47f", "tree": "db24a9a2636bc5907d08a450303cd8cdeecc3aa9", "parents": [ "aee695cd6a47219958fa2068f4732cbf7e159e7e" ], "author": { "name": "Linus Torvalds", "email": "torvalds@klaava.Helsinki.FI", "time": "Sun Oct 18 14:45:46 1992 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:09 2007 -0400" }, "message": "ANNOUNCE: 0.98.pl2 is out\n\nPatch 2 to 0.98 is out - it\u0027s available as both full source and as\npatches against 0.98.1 at nic.funet.fi: pub/OS/Linux/testing/Linus (and\ntesting is still unreadable, so you have to cd to it blindly).\n\npatch-2 is \u003e150kB compressed, as it contains several big changes. Most\nnotable are:\n\n - the new FPU-emulator by Bill Metzenthen. It\u0027s bigger than the old\n one, but thanks to it, linux fpu emulation is no longer a quick hack,\n but a real emulator: it does all the 387(486) math instructions, and\n does them much faster than the old emulator + the soft library.\n\n The new math-emulator means that a separate soft-float library is no\n longer needed. It also makes even a non-coprocessor system pretty\n useful for limited math-calcs - the complex functions are much faster\n when they no longer have to be calculated using simple functions, and\n even the simpler instructions that my old emulator handled are faster\n using the new one.\n\n The size of the new emulator may mean that people who have little\n RAM, but do have a coprocessor should probably recompile the kernel\n with the emulator disabled.\n\n - various minor mm fixes by me: trapping kernel NULL dereferences,\n cleaning up the page table initializations and the 16MB patches, and\n various other bugfixes. get_free_page(GFP_ATOMIC) should preserve\n the interrupt flag, so malloc() should be safe now - hopefully no\n more of the tcp/ip memory management problems.\n\n The NULL pointer trapping may result in errors like:\n\tUnable to handle kernel paging request at address C0000???\n\tOops: 0000\n\t..... debugging info .....\n\n There were several NULL pointer dereferences in the serial and tty\n drivers, which should now be fixed. I\u0027ve also fixed any other errors\n I\u0027ve seen, but if there are problems in the scsi drivers or similar\n things I cannot test, I\u0027d like to hear about them.\n\n - scsi driver changes by Eric Youngdale. Preliminary support for\n removable media, and some bug-fixes. Due to white-space problems\n with eric\u0027s patches, the scsi patches are a bit bigger than\n necessary, but they should be ok even though I had to put them in\n partly by hand (and being unable to test them...)\n\n - The new tcp/ip patches that were sent to the NET channel not long\n ago. Yes, they are alpha, but so is the whole tcp/ip directory, so I\n put them in even thought they haven\u0027t been extensively tested (and\n they did have a serious problem in the ioctl code, which I fixed).\n\n - psaux mouse patches by Dean Troyer, as well as the mouse.wait \u003d NULL\n patch.\n\nBefore (or after) patching, you should remove the old math-emulator (ie\n\"rm -rf /usr/src/linux/kernel/math\") as it is no longer needed. You\nshould also do a \"make dep\" to update dependencies: as usual, I edited\nout the dependancy-changes. Do a \"make clean\", edit the main (and net)\nMakefiles to suit your system, and compile.\n\nAnd finally: I will no longer be making the bootdisks available -\nthey\u0027ll be made by hlu/jwinstead and will probably be boot+root-disks\nusing lilo, as done on the hlu disks. That may mean that a bootimage\nwon\u0027t be available at once, but most people who want to use the\nabsolutely newest images probably compile them themselves anyway, so\nthat shouldn\u0027t be a problem.\n\n\t\tLinus\n" }, { "commit": "aee695cd6a47219958fa2068f4732cbf7e159e7e", "tree": "392a8a6431c0598d3b22abeda84aaf95d820a249", "parents": [ "1112c60428a179b8d33db59b0b946057b2b50dfb" ], "author": { "name": "Linus Torvalds", "email": "torvalds@klaava.Helsinki.FI", "time": "Mon Oct 05 17:07:03 1992 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:08 2007 -0400" }, "message": "Patch 1 for 0.98 out...\n\nnic.funet.fi: pub/OS/Linux/testing/Linus contains the first patch to\n0.98 (along with a bootimage and the full source for those that don\u0027t\nlike to patch).\n\nPatch1 to 0.98 mainly corrects some driver problems: it contains the\nadded \"inb_p(HD_STATUS)\" for hd.c, as well as a changed mouse driver\nsetup (hope it works - I couldn\u0027t test it..). There are also some SCSI\ndriver patches: the seagate driver uses irqaction() to get irq\u0027s, and\nthe aha1542 driver has the speedup patches.\n\nThe bootimage should be compiled without the auto-SVGA mode, so people\nwho had problems with linux automatically using a SVGA mode should be ok\nin this release.\n\n\t\tLinus\n" }, { "commit": "1112c60428a179b8d33db59b0b946057b2b50dfb", "tree": "804571ee0053e82beca788cfe63d33d7a039c03a", "parents": [ "6969da2918ad2edd45dcf9e65ea2db837179ce9c" ], "author": { "name": "Linus Torvalds", "email": "torvalds@klaava.Helsinki.FI", "time": "Tue Sep 29 21:11:21 1992 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:07 2007 -0400" }, "message": "Finally: 0.98\n\nSorry for being late - I can\u0027t even show any great new features in 0.98,\nbut at least it\u0027s out now, and available at the normal place (ie at\nnic.funet.fi, pub/OS/Linux/testing/Linus). So far there is only a\nfull-source version available, although I\u0027ll probably make it available\nas a patch too tomorrow or so (but the patch won\u0027t contain the tcp/ip\nstuff).\n\n0.98 is essentially the same as 0.97.pl6 - the changes are mostly:\n - tcp/ip (0.8.1) is in. It\u0027s not compiled into the standard bootimage,\n and you\u0027d better be on the tcpip mailing-list to use it, but it\u0027s\n there. I\u0027ve been unable to test it further than just watch it\n compile...\n - extfs patch to correct the problem with big directories with holes.\n - mouse patches (ie improved detection-routines)\n - minor scsi patches (ultrastor driver change)\n - swiss keyboard\n - some serial driver patches\n - the 32mb patches are in, so if you aren\u0027t using a DMA-SCSI driver,\n and have more than 16MB physical memory, you can get it recognized.\n - edited hd.c\n - corrected core-dumping routines\n\nI didn\u0027t get my mm patches working yet, so they\u0027ll have to wait. The\nabove are almost 100% by others - I have edited some of the patches, but\nthere is nothing major new by me. Most of it is minor bug-fixes, and\nthe only thing that might be a bit of a problem are the hd.c changes:\nbut I hope they\u0027ll solve more problems than they cause. Knock wood.\n\nAt nic.funet.fi you can currently find (a) the full sources (b) a\nbootimage (US keyboard, floppy root, no tcp/ip) and (c) the protocols.h\nfile needed for compiling the tcp/ip directory (which should go into\n/usr/include/netinet/). I hope people try it out, and that there are no\nnew problems with this release.\n\n\t\tLinus\n" }, { "commit": "6969da2918ad2edd45dcf9e65ea2db837179ce9c", "tree": "6d2cc521fc9f2bc90f27dce5da11a1d3538581bd", "parents": [ "ca7e955d618e12c34ebc6b644270a1e8f96c046e" ], "author": { "name": "Linus Torvalds", "email": "torvalds@klaava.Helsinki.FI", "time": "Sun Sep 20 17:19:56 1992 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:07 2007 -0400" }, "message": "patchlevel 6\n\nYou all know what the subject line means by now: in case you want to\ntrack my kernel versions, the weekly patch is available at nic.funet.fi,\npub/OS/Linux/testing/Linus.\n\nThis patch does not contain any major bug-fixes: it corrects named pipes\nthat broke with pl5, and has some minor changes in the IO-instructions\nand the hd-driver, but those shouldn\u0027t matter for most of you.\n\nIt does contain all the scsi-patches that I\u0027ve gotten so far, so if the\nbootup sequence died on you in the scsi code, pl6 should correct this.\n\nThe major part of the patch is tytso\u0027s serial line changes, making the\ntty structures dynamic. No more NR_PTY\u0027s - the number of pty\u0027s is now\nbounded only by the minor number setup (max 64 pty\u0027s) or the amount of\nmemory available (opening a pty requires a page of memory for tty\nqueues). Similarly for serial lines.\n\nThe above just means that while pl6 can be useful, the changes to pl5\naren\u0027t big enough to worry about. Most people don\u0027t use named pipes, it\nseems, and the other changes are either cosmetic or hardware-dependent.\nI still hope people upgrade, if only so that I can get new bug-reports.\n\nI had hoped to release 0.98 this weekend, but studies and the scsi/hd\nproblems put an end to that. 0.98 should be out next weekend or so.\nExpect the tcp/ip subdirectory and possibly some mm changes.\n\n\t\tLinus\n" }, { "commit": "ca7e955d618e12c34ebc6b644270a1e8f96c046e", "tree": "0200980bdd9fada625733866ac0892fd92a1f987", "parents": [ "579e466562a7fdf60f9bec70e71da950c4873d32" ], "author": { "name": "Linus Benedict Torvalds", "email": "torvalds@klaava.Helsinki.FI", "time": "Sat Sep 12 18:21:31 1992 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:07 2007 -0400" }, "message": "0.97 patchlevel 5 available\n\nThe subject says it all: the newest patches have been uploaded to\nnic.funet.fi: pub/OS/Linux/testing/Linus. Patchlevel 5 is available both\nas complete source (linux-0.97.5.tar.Z) or as a biggish patch against\npatchlevel 4 (linux-0.97.patch5.Z).\n\nPatch 5 fixes the extended filesystem problems (thanks to Remy Card), as\nwell as including many smaller fixes (some more fs cleanups, the CDROM\npatches and several other minor changes). Pl5 finally removes even the\nlast few header-files that were incompatible with the normal headers, so\nthe \"-nostdinc -I$(KERNELHDRS)\" stuff is gone.\n\nPatch 5 should also fix the problems with iopl() that resulted in the\nX8514-server having problems with 0.97.pl2 and above.\n\nIn case people are wondering, my schedule for 1.0 looks something like\nthis:\n\n - 0.98 out in about a week: this is essentially 0.97.5 + the tcp/ip\n directory, as well as any fixes that may come up. I\u0027ll try to get\n the loadable driver interface into it too.\n\n - 0.99 out after 0.98 has been shaken down: a month or so.\n\n - 1.0 will be the same as 0.99: the only changes will be eventual\n trivial bug-fixes in case 0.99 has some problems. This is just to\n try to get over the \"X.0\" bug syndrome.\n\nThere are a few on-going projects: depending on circumstances these will\nbe implemented sooner or later, so I won\u0027t give any promises. These\ninclude: loadable drivers/fs\u0027s (alpha-patches already availabla), full\nsupport for different block-sizes (some work still required), and a\nextensive rewrite of the mm routines (I\u0027ll want to make a vmm interface\nsimilar to the vfs interface for the filesystem routines).\n\n\t\tLinus\n\n-----\n\nFrom: torvalds@klaava.Helsinki.FI (Linus Benedict Torvalds)\nSubject: Re: 0.97 patchlevel 5 available\nDate: 12 Sep 1992 22:42:41 GMT\n\nIn article \u003c15255@borg.cs.unc.edu\u003e martin@franck.cs.unc.edu (Kevin Martin)\nwrites:\n\u003eIn article \u003c1992Sep12.182131.2168@klaava.Helsinki.FI\u003e\n\u003etorvalds@klaava.Helsinki.FI (Linus Benedict Torvalds) writes:\n\u003e\n\u003e\u003ePatch 5 should also fix the problems with iopl() that resulted in the\n\u003e\u003eX8514-server having problems with 0.97.pl2 and above.\n\u003e\n\u003eI\u0027ve just tested the 8514/A X server with 0.97.pl5 and it works\n\u003ebeautifully. The problems with the X server appearing to hang when\n\u003equitting from TWM with the 0.97.pl2 and more recent kernels are\n\u003efixed. Thanks Linux!\n\nActually, thank Kevin Martin: he was able to give me a good bug-report\nthat made it pretty easy to track it all down (he even had a minimal\nprogram which showed the incorrect behaviour).\n\nOther changes that pl5 has include:\n\n - slightly edited \u003casm/io.h\u003e: easily editable IO delay instructions.\n The default delay-instruction is now a \u0027inb\u0027 from port 0x80: this\n should be a bit safer than the outb. But you can easily change it to\n the \"standard\" two short jumps or whatever.\n - malloc() is cleaned up, and \u0027malloc_grab_pages()\u0027 is gone (Biro)\n - I cleaned up the internal inode structure a bit: i_data[] is no\n longer part of the base inode, but is instead part of the union of\n fs-dependent info. Pipes also have their own cleaner interface.\n - the msmouse patches are in: currently there is no valid test that a\n msmouse actually exists, so linux always says \"mouse detected and\n installed\", but that is nothing to worry about.\n - the msdos-fs one-line performance patch is in.\n\nThe most important fix for ext-fs users should be the fact that pl5\nshould fix the ext-fs bugs: the ext-fs patches are essentially the same\nthat pl3+4 did to the minix-fs. So now the new \u0027strip\u0027 should be safe\non all filesystem types.\n\nNote that the upcoming release of gcc (and thus the upcoming X11 2.0)\nwill require at least 0.97.pl4 in order to be safely used: I\u0027ll probably\nmake a bootimage of 0.97.5 available for those that don\u0027t want to (or\nare unable to) recompile the kernel. I\u0027ll just wait a day or two to see\nif there are some unexpected problems with pl5.\n\n\t\tLinus\n" }, { "commit": "579e466562a7fdf60f9bec70e71da950c4873d32", "tree": "8c99ade602d678289fa153cc54a2c3aefc804dac", "parents": [ "d72651cda1a12e51ae4b03a8fe0e280d89aecc33" ], "author": { "name": "Linus Benedict Torvalds", "email": "torvalds@klaava.Helsinki.FI", "time": "Mon Sep 07 11:40:14 1992 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:07 2007 -0400" }, "message": "0.97 patchlevel 4\n\nLinus \"dances with patches\" Torvalds strikes again: I\u0027ve already made\npatchlevel 4 of 0.97. It may not be a new record, but it\u0027s close :-)\n\nPatch 4 is a very minor patch, but it\u0027s pretty important if you want a\nstable filesystem (and let\u0027s face it: most people seem to prefer a\nfilesystem that stays up a bit longer). While patch3 corrected most of\nthe race-conditions in the minix fs, I overlooked the [f]truncate system\ncalls, and this new patch corrects that.\n\n[f]truncate is a very race-prone function, and as if that wasn\u0027t enough,\nthere was also a pretty bad error in truncate.c that resulted in the\nindirect blocks not being correctly marked dirty when doing partial\ntruncates. The latter problem is probably the reason for most of the\nfilesystem corruptions that have been reported - the race-conditions\nwere a lot harder to fix, but they also happen a lot less often.\n\nNote that the [f]truncate bug isn\u0027t new: it has been in the kernel since\n[f]truncate was first implemented (0.95?). But until now, [f]truncate()\nhasn\u0027t actually been used very much - only the latest versions of the\nbinutils have used ftruncate to strip binaries etc. So the problem\nhasn\u0027t shown up that much.\n\nSo while I consider patch4 to be crucial, you /can/ actually live\nwithout it: I haven\u0027t seen the buffer corruption problem at all (until I\nactually tested for it after getting good bug-reports), so you can\nprovably miss it for a long time. But if you have ever had corruption\nproblems, I\u0027d suggest upgrading to pl4 as soon as possible.\n\nThe corruption problems show up most clearly when using a new \"strip\"\nbinary, although they are theoretically possible with other programs\ntoo. Thanks to \"obz@raster.kodak.com\" and \"jon@robots.ox.ac.uk\" for\ngood bug-reports: thanks to them I was able to pin down the error to\ntruncate.c, and after that it was pretty easy to get rid of it.\n\nAlso note that this patch still hasn\u0027t fixed the extended filesystem: I\nsuspect the same bugs lurk around there. I\u0027ll get it corrected by 0.98\nat the latest.\n\nThe patch is included at the end of this post (it\u0027s very minor - it\ncontains patches mainly against linux/fs/minix/truncate.c) , and I\u0027ll\nalso update nic.funet.fi (pub/OS/Linux/testing/Linus) to have the new\nsources. Sorry for the inconvenience,\n\n\t\tLinus\n" }, { "commit": "d72651cda1a12e51ae4b03a8fe0e280d89aecc33", "tree": "ed862f51f0be682ae27304464939218c2418d1c6", "parents": [ "2288d80abfb91ee08da7989ca1dd2338cc949fd3" ], "author": { "name": "Linus Benedict Torvalds", "email": "torvalds@klaava.Helsinki.FI", "time": "Sat Sep 05 18:46:06 1992 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:07 2007 -0400" }, "message": "0.97 patchlevel 3 available\n\nWell, 0.97.pl3 is available both as full source (linux-0.97.3.tar.Z) and\nas a context diff (linux-0.97.patch3.Z) on nic.funet.fi, in the normal\ndirectory (pob/OS/Linux/testing/Linus). It seems some line is once more\ndown on the way to the US, so I haven\u0027t uploaded it to tsx-11 yet.\nAlso, I haven\u0027t been able to use the pub/Linux/Linus directory on\nsunsite (it dies on me when I try to send the group password), so I\ndon\u0027t know when sunsite will get the new release.\n\nPatch3 is almost 100kB even compressed, as there were quite big changes\nin the mm and minix fs. No major new features: there are two new system\ncalls: swapoff(const char * swapfile) and wait4(), and linux accepts\nseveral swap-files, but the rest of the thing is mostly bug-fixes or\nsimply rewrites.\n\nMajor changes:\n - new swap-page handling: linux no longer uses just one bit to keep\n track of used swap-space, but a counter for each swap-page. This\n allows processes to share swap-pages after a fork(), and should\n result in /major/ performance increases on machines with less memory.\n I\u0027ve seen better performance even with 8MB - I wouldn\u0027t be surprised\n if 4MB machines would re-compile the kernel noticeably faster under\n pl3. I\u0027d be interested to hear numbers.\n - The low 1MB memory that isn\u0027t used directly by the kernel is now\n swappable memory, instead of being hardcoded for buffer cache. The\n patches for this were originally by tytso, and I expanded on it a bit\n more. This might also help better performance on 2-4MB machines.\n Note that this does /not/ mean that you can use 1M machines for\n linux: linux still needs some extended memory.\n - the dosfs has been upgraded to dosfs.8 - patches by almesber.\n - I edited the minix fs pretty heavily to remove a couple of race-\n conditions. The same races still exist in the extended fs, as I\n didn\u0027t have time to edit that yet. The minix-fs took precedence as I\n know that better, and extfs isn\u0027t \"official\" yet anyway.\n\nother changes:\n\n - the mouse-driver now handles both Logitech (minor \u003d 0) and PS/2\n (minor \u003d 1) busmice.\n - there is a proc-fs for access to user memory/files etc.\n - better support for the tcp/ip patches (but see below...)\n - corrected symlink and /dev/[k]mem behaviour\n - Lars Wirzenius\u0027 README (with minimal comments by me) and the GNU\n COPYING notice are now part of the normal kernel setup, and can be\n found in the tar-archive.\n - the floppy ioctl() to get the FD parameters no longer requires root\n priviledges. Thus, the msdos emulator runs even for a normal user.\n\nSome comments on patchlevel 3:\n\n\tmm:\n\nThe swap-page handling resulted in a reduction of swap-file (or\npartition) size to a maximum of 16MB per file. It\u0027s nothing inherent to\nthe code, but it eased some algorithms, so I didn\u0027t bother coding around\nit. After all, 16MB is enough for most people, and if you want more,\nyou can have up to 128 swapfiles of 16MB each. If I get enough\nhate-mail about it, I might just try to find the energy to correct it.\nMaybe.\n\nBigger swapfiles will still work, but linux will take advantage of only\nthe low 16MB. Also, there is no nifty logic to try to optimize the\nusage of the swap-files: pages are simply allocated from one swap-file\nuntil it fills up, and then the next swap-file is used.\n\nThe memory management changes break ps/free once more, but not very\nmuch. Also, I changed the load-average counting, so \u0027w\u0027 also needs\nslight editing. On the other hand, I made \u0027/dev/kmem\u0027 mmap()able, and\n\u0027ps\u0027 and \u0027free\u0027 should be edited to take advantage of that: it should\nresult in much faster operation, as well as possibly using less real\nmemory.\n\n\tfs:\n\nThe fs changes should remove at least two races - the races don\u0027t happen\nvery often, but they were theoretically possible, and might be the\nreason for some fs corruption problems that have been reported. The\nchanges are related to the use of bmap() - the bmap interface doesn\u0027t\nreally lend itself to some things that it was used for. Re-writing\ninternal fs-functions not to use bmap not only should have removed any\nraces, but also actually resulted in cleaner code.\n\nThe proc-fs code isn\u0027t too beautiful, and I\u0027ll probably leave it out\nfrom 0.98 unless I can make it loadable. We\u0027ll see. If anybody wants\nto use it, you can do something like\n\n # mount -t proc /dev/ram /proc\n\nInstead of /dev/ram you can use any block device - it\u0027s not used, and is\nonly a dummy as the proc-fs doesn\u0027t actually use any external device.\n(but note that the device is still marked as mounted, so you cannot\nmount it for anything else).\n\n\tkernel/mm/lib:\n\nThe TCP/IP patches are also essentially in 0.97.pl3 - not the full\nTCP/IP directory, only the patches to the main kernel. NOTE!! I don\u0027t\nlike the \u0027grab_malloc_pages()\u0027 function, so I left that out, and added a\nGFP_ATOMIC priority to get_free_page() that should be used instead. I\nhope this will be used (Ross?), as it\u0027s a lot cleaner.\n\nAlso, I hope the tcp/ip people will clean up malloc() so that it doesn\u0027t\npanic instead of returning NULL etc. Ugly, ugly. This is related to\nthe get_free_page(GFP_ATOMIC) changes, and I\u0027d like to have patches as\nsoon as possible - tcp/ip won\u0027t be part of the standard kernel until\nthat can be cleaned up.\n\n\t\tLinus\n" }, { "commit": "2288d80abfb91ee08da7989ca1dd2338cc949fd3", "tree": "041e1c0d58821e7cd1767d38afafd6e148eab7ec", "parents": [ "fe07d2d9e764decdab478241a2ad5a157392af86" ], "author": { "name": "Linus Benedict Torvalds", "email": "torvalds@klaava.Helsinki.FI", "time": "Sun Aug 23 22:17:41 1992 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:06 2007 -0400" }, "message": "linux-0.97 patchlevel 2\n\nAs promised, 0.97.pl2 is out today (well, over here it\u0027s already\ntomorrow, so I guess I\u0027m 35 minutes late. Naughty, naughty). Right\nnow, the patch (and full source for those that don\u0027t like to patch up\nthe system) is available at \"nic.funet.fi: pub/OS/Linux/testing/Linus\",\nbut I\u0027ll try to put it on some other sites as well if I\u0027m able and\nenergetic enough. Probably tomorrow - together with a binary for those\nthat aren\u0027t willing to comple the kernel on their own.\n\n0.97.2 has mostly my mm/fs patches, along with some relatively minor\ndiffs by others (including file locking by Doug Evans). User-level\nchanges are minor: but the mm has changed a lot, and the vfs routines\nhave been changed to keep track of the error-messages a bit better.\nAlso, the vfs-interface to \"follow_link()\" changed slightly: people who\nare making filesystems should look at the changes (but they are\nrelatively minor, and shouldn\u0027t result in any problems - both the\nextended fs and minix fs needed just a simple change in their respective\nsymlink.c files).\n\nThe mm changes /might/ lower performance slightly, as the paging TLB\u0027s\nare now flushed at every task-switch due to the new system, but I doubt\nit\u0027s noticeable. The other performance changes (dynamic buffers etc) in\n0.97(.pl1) should overshadow that particular problem.\n\nI hope this release means that these kinds of low-level rewrites aren\u0027t\nneeded for a while: the last couple of releases have changed some very\nfundamental things. Nothing seems to have suffered too badly, but I\u0027d\nbe happier if it all got tested more thoroughly. Anyway, discounting\nthe ps/free etc suite of programs, everything I have tried has worked\nflawlessly despite the big kernel changes.\n\nI\u0027m still worried about the reports about messed-up buffers, but have\nbeen unable to reproduce the problem, and nobody has so far\ndisillusioned me about my guess that it\u0027s a problem with the SCSI code\n(which at least gives me an excuse for not doing anything about it :-).\nOther problems include at least one report of spontaneous re-booting,\nwhich is totally inexplicable, so I\u0027m blaming hardware once more until I\ncan get better data on the thing.\n\nAs to patches sent by others: 0.97.2 contains very little of that kind\nof code. I\u0027ve been too busy either working, or implementing my own\nchanges that I have simply ignored them for the most part. Remind me\n(or resend them relative to the new kernel) if you have a patch that is\nstill needed.\n\nThere is one new system call: \u0027vm86(struct vm86_struct * info)\u0027. It\u0027s\nnot ready for general use yet - it works, but will probably need some\ntweaking before being practical. But supporting a virtual 86 mode was\nso easy after the mm rewrite that I felt it was worth implementing: the\nvm86 code is less than 50 lines of C right now.\n\n\t\tLinus\n\nPS. The bright spot of the week goes to \"The Oxford Beer Trolls\" - all\nUK inhabitants should probably be locked into some (big) mental\ninstitution and TOBT should probably have a wing of their own, but\nthanks to them linux can now call itself \"beerware\" :-)\n" }, { "commit": "fe07d2d9e764decdab478241a2ad5a157392af86", "tree": "d404d08d9546e77f296b0c1d678f6dcb68b3d829", "parents": [ "36c51b40b2a02f608b7214efd71cb1e0bbd17788" ], "author": { "name": "Linus Benedict Torvalds", "email": "torvalds@klaava.Helsinki.FI", "time": "Thu Aug 06 10:47:32 1992 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:06 2007 -0400" }, "message": "0.97 patch1 available on nic\n\nIt seems all connections to Finland are down (or at least some server is\nacting up), so patch1 is available only on nic.funet.fi:\n\n\tpub/OS/Linux/testing/Linus/linux-0.97.patch1.Z\n\nI\u0027ll upload it to the other sites when I get a working ftp-connection.\n\nPatch 1 is essentially a performance-release, but it also contains some\nother patches: Ross Biro\u0027s tcp-ip stubs are there (but not the tcpip\nsubdirectory: alpha-testers should know where to find that), as are the\next-fs superblock cleanups. The first header-file patch by hlu is also\nin there.\n\nThe resulting patch is pretty big - it\u0027s also not as cleaned up as I\u0027d\nlike it to be. The swapping/buffer-block handling heuristics are\nbetter, but could still do with some tuning. Also, the idle task in\nthis version doesn\u0027t do very much: it will be expanded to do some more\npage-table calculations.\n\nI will be unable to hack on linux for a couple of weeks (I\u0027ll still\nanswer mails, read the newsgroup and fix bugs, but no heavy-duty\nhacking) due to some \"circumstances beyond my control\". That probably\nmeans that this patch is the last one for a while (three weeks) unless\nsome bad bugs show up.\n\n\t\tLinus\n" }, { "commit": "36c51b40b2a02f608b7214efd71cb1e0bbd17788", "tree": "d98963643d94ad6edce0b75cc62ac162dca870ad", "parents": [ "67406b341328a6503eaa3de23e553be903ed5891" ], "author": { "name": "Linus Benedict Torvalds", "email": "torvalds@klaava.Helsinki.FI", "time": "Sat Aug 01 15:08:38 1992 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:06 2007 -0400" }, "message": "Linux v. 0.97 is out\n\nLinux version 0.97 is available as both a complete source-tree\n(linux-0.97.tar.Z) and a bootimage (bootimage-0.97.Z) on the normal\nftp-sites. It\u0027s in incoming on tsx-11.mit.edu, so it will take a day or\ntwo to actually show up, but it\u0027s available right now on\n\n\tnic.funet.fi:\n\t\tpub/OS/Linux/testing/Linus/\n\tbanjo.concert.net:\n\t\tpub/Linux/Linus/\n\nThe nic.funet.fi-directory is under \u0027testing\u0027 not so much because this\nwould be a testing-release, but because the directory-setup is in\ntesting :-). I think \u0027testing\u0027 is unreadable, so you have to cd to the\ndirectory blindly.\n\nThere is also a kernel-compilation README (written by Lars Wirzenius),\nas well as a COPYING (which is just a pointer to the GNU copyleft). The\nlatter not because anything has changed, but because I got a few mails\npointing out that the copyright of linux wasn\u0027t too clear. That also\nresulted in changing the \u0027(C)\u0027s in the source to \u0027Copyright\u0027.\n\nChanges in 0.97:\n\n - The VESA-support was removed. I\u0027d be happy to put it back once it\n works on all hardware. Instead of the VESA-code, I finally put in\n the automatic SVGA setup patches. See the top-level Makefile.\n\n - The IRQ code has solidified, and should work on all machines. Not\n all of the SCSI drivers use it yet, so I expect patches for that..\n\n - Serial interrupts are handled slightly differently, and performance\n should be up. I\u0027ve sent out a few alpha-releases, and testing seems\n to indicate that\u0027s actually true this time. Reactions have ranged\n from \"nice\" to \"wonderful\" :-)\n\n - The buffer-cache and memory management code has been edited quite a\n bit. ps/free etc programs that reads kernel memory directly no\n longer work, and even a recompilation won\u0027t be enough. They actually\n need editing before they work.\n\n The buffer-cache now grows and shrinks dynamically depending on how\n much free memory there is. Shift+PrintScreen will give some memory\n statistics. (Ctrl+PrSc gives task-info, ALT+PrSc gives current\n register values).\n\n The mm code changes removed some race-conditions in the VM code, and\n I also tried to make the Out-of-swapspace error less severe (better\n thrashing-detection etc).\n\n - The super-block code has been cleaned up. Especially the extended fs\n needs to be edited a bit to take advantage of the new setup, and I\n expect Remy Card will have a patch out eventually.\n\n - include-files have been moved around some more: there are still some\n names that clash with the standard headers, but not many.\n\n - Unswappable processes implemented: by default only \u0027init\u0027 is\n unswappable. This is a bit safer in low-memory conditions, as at\n least init won\u0027t die due to low memory. I also made killing init\n impossible: if init doesn\u0027t recognize a signal, it simply won\u0027t get\n it. Some other changes (\"while (1) fork();\" won\u0027t kill the machine\n for non-root users etc)\n\n - The new SCSI drivers are in. These make the kernel noticeably\n bigger, but you can leave them out if you don\u0027t want them.\n\n - The floppy- and hd-drivers print out more debugging-info in case of\n errors: this might be irritating if you have hardware that works, but\n often gives soft-errors. On the other hand, some old debugging-info\n was removed - notably for user-level protection errors etc.\n\n - Various minor fixes. I haven\u0027t made cdiffs (and I haven\u0027t gotten any\n requests for them, so I probably never will), but they would be\n pretty big.\n\nThings that I didn\u0027t have time for:\n\n - I wanted to rewrite the tty drivers to be more \"streams-like\" (ie not\n an actual streams-implementation, but some of the ideas from\n streams). I never got around to it: there was simply too much else\n to do.\n\n - I got a lot of patches, and some went in, others didn\u0027t. If you\n think your patch was important, please re-send it relative to the new\n version.\n\nI\u0027d like comments on the new system: performance / clarity of code etc.\n0.97 should correct all known bugs (at least the ones I know about), but\nI guess that\u0027s just wishful thinking.\n\nNote that the dynamic buffer-code also handles differently-sized\nbuffers, but that the rest of the system (block device drivers,\nfilesystem code etc) cannot yet take advantage of this - there is still\nsome coding needed.\n\n\t\tLinus\n" }, { "commit": "67406b341328a6503eaa3de23e553be903ed5891", "tree": "96af6c83e4c742aa19c72da458615ee6db6e2c57", "parents": [ "1c368f1aedaa10a03fcec9d19eec9f9ba60015d2" ], "author": { "name": "Linus Benedict Torvalds", "email": "torvalds@klaava.Helsinki.FI", "time": "Sat Jul 18 20:54:48 1992 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:06 2007 -0400" }, "message": "0.96c second patch\n\nThe subject pretty much says it all: I\u0027ve sent out the \"weekly patch\"\nand I\u0027d be very interested in comments. As with patch1, there are some\nvery fundamental changes in the kernel, and they might have some\nproblems. I\u0027d want as many as possible to test out linux-0.96c.pl2, as\nthat has always been the best way to test out the changes. Everything\nworks on my machine, but that doesn\u0027t guarantee it will work on other\nsetups...\n\nThe MAJOR change in 0.96c.pl2 is the totally rewritten sleep/wakeup\ncode. That, together with the IRQ code introduced in pl1 and slightly\nedited in pl2, means that two very fundamental things in the linux\nkernel have changed in the last two weeks. The code is cleaner, easier\nto add devices to, and hopefully faster, but it\u0027s still a bit risky to\nchange this kind of very low-level behaviour.\n\nSelect() is now implemented using the vfs jump tables, and thanks to the\nbetter sleep/wakup interface, select() performance should be noticeably\nbetter. At least xload seems to give lower load-averages, and I hope\nka9q will work better with the new kernel. Note that things like the\ntty code doesn\u0027t yet take full advantage of the new features the\nrewritten sleep offers, but I wanted to get a good testing-release out\nbefore actually tweaking all the routines to use the new interface.\n\nThe IRQ routines have changed slightly, and all known bugs are fixed.\n\nWhile I\u0027m most interested to hear comments about the IRQ and\nselect/sleep/wakup code, there are a few other changes in pl2:\n\n - Swiss keyboard support.\n\n - Screen blanking now only reacts to key-presses and kernel messages:\nnormal tty output doesn\u0027t make the screen unblank.\n\n - DOS-fs version 5 is in. It wouldn\u0027t hurt to try it out. It\u0027s\nsomewhat alpha still, but it seems to work. mtools should be a thing of\nthe past once the dosfs is a bit more tested.\n\n - core-file magic number, and a minor bug in ptrace is fixed\n\n - a bus-mouse is supported. I\u0027d like to hear if it still works after I\ndid the select() patches \"blind\" (I can\u0027t test it on my machine).\n\n - iopl changing is possible (but requires root priviledges): this\nallows access to all IO ports, as well as the interrupt flag. Don\u0027t use\nit unless /absolutely/ necessary: a bug in your program will most likely\ncrash the machine if you are running with IO priviledges. It\u0027s needed\nfor some X VGA drivers.\n\nAs a result of all the changes, the diff is pretty big. Apply and build\nit with something like:\n\n cd /usr/src\n zcat linux-0.96c.patch2.Z | patch -p0\n cd linux\n make dep\n make clean\n make Image\n\nassuming you have the 0.96c.pl1 kernel in /usr/src/linux. I\u0027ve had some\nreports that my patches won\u0027t always go in cleanly: I know for a fact\nthat patch1 patches cleanly (I rebuilt 0.96c.pl1 by downloading it all\nfrom banjo), so the error is in your end.\n\nPossible problems:\n\n - The VESA code in setup.S has some problems. I haven\u0027t even looked\ninto it yet, so if it won\u0027t work for you, please either (a) use the\nunpatched setup.S from 0.96c, or (b) try to find the problem and tell\nme. (b) is preferable, of course. I\u0027d like to have VESA support, but\nif the bug isn\u0027t found, I\u0027ll have to use the non-VESA version for 0.97.\n\n - The IRQ code in 0.96c.pl1 could overrun the stack if linux got\nun-ending interrupt requests, resulting in a re-boot. With pl2, this\nshouldn\u0027t happen: linux should print out something like \"Recursive\ninterrupt on IRQx. Shutting down\" and simply disable the problematic\nIRQ line. If you see this message, I\u0027d be very interested to hear about\nit (which IRQ, what devices you have, etc).\n\n - And any new or old bugs I haven\u0027t found yet.\n\nI have one report that 0.96c.pl1 has problems with the inode table, and\npanics on bootup with a \"no more inodes in mem\" report. Can anybody\nconfirm this sighting? I haven\u0027t found the reason for it, and haven\u0027t\nseen it myself. I\u0027m hoping it\u0027s an installation problem, but if anybody\nelse sees the same behaviour, I\u0027m SOL.\n\n\t\tLinus\n" }, { "commit": "1c368f1aedaa10a03fcec9d19eec9f9ba60015d2", "tree": "e07b7f31e1a321407fa03f1b8c7eca3330df57d6", "parents": [ "f79e5351608302264fbd909f99f5420b01079582" ], "author": { "name": "Linus Benedict Torvalds", "email": "torvalds@klaava.Helsinki.FI", "time": "Sat Jul 11 20:40:11 1992 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:05 2007 -0400" }, "message": "Patch1 to 0.96c out\n\nI have sent off my first patch to 0.96c: as usual it\u0027s available\ndirectly from banjo.concert.net while nic.funet.fi and tsx-11.mit.edu\nmight take a day or two to put it into the right directories.\n\nlinux-0.96c.patch1 contains more changes than I originally envisioned: I\nchanged the IRQ routines and the serial code to be easier and cleaner\n(and hopefully more efficient) and I thought that would be it. I was\nwrong.\n\nI got several patches (and one bug-report) again, and while I haven\u0027t\nhad time to check them all, some of them are in. Fixes:\n\n - Remy Cards correction to the out-of-space problem with the extended\nfs is here. Most people using the ext-fs might already have applied\nthis patch, in which case you might have problems patching.\n\n - my ftruncate() fix is here. Again, if you already did the trivial\npatch by hand, you\u0027ll get errors when patching.\n\n - almesber\u0027s implementation of read-only filesystems is here (after\nediting by yours truly). The mount() system call now accepts a flags\ninteger as well as a pointer to some arbitraty data in user space for\nsome special mount() calls. The general flags allow (a) read-only\nmounting, (b) disabling of suid executables (c) disabling of device\nspecial files and (d) total disabling of executables on a per-filesystem\nbasis. The filesystem specific mount() info isn\u0027t currently used by any\nfs, but can be used to specify additional information that depends on a\nspecial fs type (a password or similar would be possible..)\n\n - the rename() system call had a bug in that it allowed moving over a\ndirectory: I think the code to handle this was lost in the vfs editing,\nand although the GNU mv utility checked it, a malicious (or just\nunsuspecting) program can destroy the fs using this. Thanks for the\nbug-report: it was very easy to add once I saw the problem.\n\n - support for vesa-standard svga cards in setup.S. I\u0027m unable to test\nthis, but my svga card still works after the patch, so I left it in in\nthe hope that it doesn\u0027t break for anybody else.\n\n - various minor editing by me, or minor patches sent in by others.\n\nThe full cdiff is almost 50kB compressed, so this is a bigger-than-usual\npatch. Hope there are no problems. People who are using the new SCSI\ndrivers might have problems with my changes to the SCSI irq-setup\nchanges, so be careful (actually using the original sources might be a\ngood idea, and then upgrading again). I hope to get the new SCSI\ndrivers into the kernel soon (definitely in time for 0.98).\n\nI\u0027d be interested to hear comments on serial line performance, bugs,\nfeatures, etc. As usual, I\u0027m hoping this release won\u0027t contain any new\nbugs while fixing all the old ones, but I guess that\u0027s likely to happen\nright after the first winter olympics in Hell.\n\n\t\tLinus\n\nPS. Maybe nobody noticed, but I actually never made the patches to\n0.96c available. As nobody has complained, I assume everybody just got\nthe complete kernel sources and used that. Unless somebody actually\nasks for them, I don\u0027t think there is any point in making them any more.\n" }, { "commit": "f79e5351608302264fbd909f99f5420b01079582", "tree": "dc6f792d46650571f991d1419648ebf213501ebe", "parents": [ "aa6801a43050b29edd52bdf39a69d8479d1582e9" ], "author": { "name": "Linus Benedict Torvalds", "email": "torvalds@klaava.Helsinki.FI", "time": "Sat Jul 04 23:25:40 1992 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:05 2007 -0400" }, "message": "0.96c out\n\nThe latest kernel version is 0.96c: the binary and sources can be found\non banjo.concert.net: pub/Linux/Linus as usual. I haven\u0027t made the\ncdiffs yet, and I\u0027ll upload those tomorrow (at the same time I\u0027ll put it\non the other sites as well).\n\n0.96c is actually what I called patch3 earlier this week, but as the new\nfeatures were pretty big and the cdiff\u0027s are probably going to be bigger\nthan the normal patches, I decided I might as well make it a totally new\nminor release and make a bootimage and complete source available.\n\n0.96c contains:\n - bugfixes (tty, console driver, pty\u0027s, sockets)\n - fifo\u0027s (names pipes - Paul Hargrove \u0026 editing by me)\n - the alpha extended filesystem (Remy Card)\n - st_blocks implemented (ie du, ls give reasonable if not exact values\n for disk-space used)\n - Makefile cleanups and warnings at compile-time removed\n\nNote that while the extended filesystem code is there, and this kernel\nsuccessfully mounts and uses the new filesystem (with long filenames and\n\u003e64MB partitions), it\u0027s still under testing: I haven\u0027t made the mkefs\nprogram available, and the extended filesystem features shouldn\u0027t be\nused for other than testing right now.\n\nSome of the changes are just cleanups: most of the warnings when\ncompiling the new kernel should be gone (not counting the scsi code\nwhich is still the old non-cleaned-up version), and the make\u0027ing of the\nkernel is more logical now.\n\nThe bugfixes include the corrected console.c driver, the socket\ncorrections (without which X sometimes locks up), some pty semantics\ncorrections (although I\u0027m still not certain it\u0027s correct) and some\nediting in the general tty driver (including fixing the bug introduced\nin 0.96b.pl2 that caused a reboot with uninitialized tty devices).\n\nWhile the extended filesystem support isn\u0027t \"official\" yet, I can\nhappily report that my limited testing hasn\u0027t found any problems with\nlong filenames etc. It still needs a fsck program, but 1.0 looks like a\nreal possibility soon.\n\n\t\tLinus\n" }, { "commit": "aa6801a43050b29edd52bdf39a69d8479d1582e9", "tree": "cac285521fce8b7a8f188526aefa0bef0bbf1c89", "parents": [ "a5760f09ab1472f9d89669ce9e16702d617cab76" ], "author": { "name": "Linus Benedict Torvalds", "email": "torvalds@klaava.Helsinki.FI", "time": "Fri Jun 26 15:42:35 1992 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:05 2007 -0400" }, "message": "Serial lines patch: 0.96b.patch2\n\nAs promised, here is the second patch for 0.96b which hopefully clears\nup the problems with some mice by implementing most of the serial line\nflags like 5-8 bit characters and parity. It mainly corrects only\nserial problems, but there are a couple of other patches in it too: the\nfsqrt emulation patch is here, so if you already did it, you\u0027ll get a\nbad patch for that file (which you can ignore). This patch also changes\nall instances of signal-setting to use the \"send_sig()\" subroutine which\nshould allow gdb to debug all signals.\n\nApart from the serial lines, I also cleaned up the general tty-handling\nroutines slightly and removed at least one race-condition in the tty\ncode. I don\u0027t know if it\u0027s noticeable, though.\n\nYou\u0027ll need patch1 (available from all the normal sites) in order to\napply this one. As usual, I\u0027d like to hear if this patch does help\npeople, or if there are new problems. This patch will also be available\non the normal ftp sites, but as it was pretty minor, I decided I might\nas well include it in the post (uuencoded and compressed).\n\n(I also corrected the all-time favourite bug: linux now reports the\nright version number once more..)\n\n\t\tLinus\n\n[ Note to people that have sent me patches: I haven\u0027t had time to do\nthem. In some cases (the IBM char-set \u0026 BBS patch) other changes made\nthem unpatchable, in other cases I did the patch in a different way, and\nyet other patches I was just too lazy to apply. As usual, re-sending\nthe patch relative to the newest version is a good idea, although it\nstill doesn\u0027t guarantee I\u0027ll use it ]\n" }, { "commit": "a5760f09ab1472f9d89669ce9e16702d617cab76", "tree": "e85b6becf3cf1c0fbf0596d7138ecaf34c952f41", "parents": [ "e38481072b4a12a490335fbe1917ffc7bb9ce32e" ], "author": { "name": "Linus Benedict Torvalds", "email": "torvalds@klaava.Helsinki.FI", "time": "Tue Jun 23 17:28:24 1992 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:05 2007 -0400" }, "message": "Re: Too much uneaten serial causes system hang?\n\nIn article \u003carumble.709312764@extro.ucc.su.OZ.AU\u003e arumble@extro.ucc.su.OZ.AU\n(Anthony Rumble) writes:\n\u003e\n\u003eYES! I have noticed this VERY exact thing also!\n\nOh, well: it\u0027s a bug in the serial drivers that I have already fixed,\nbut I haven\u0027t done the c-diffs yet. I have rewritten big parts of the\nserial line code to be more easily configured for different IRQ numbers,\nand I noticed the bug while doing that. I\u0027ll make patch1 for 0.96b\navailable later today or tomorrow.\n\npatch1 will be mostly just the serial driver code: it allows changing\nthe irq\u0027s (and port addresses) of serial devices on the fly (with an\nioctl call), so people that have ser4 on irq5 etc shouldn\u0027t have to\nrecompile the kernel. It also returns EBUSY if you try to open a serial\nline that shares the irq-line with another line etc.\n\nAnother change in patch1 will the the handling of ctrl-alt-del: it will\nsend a SIGINT to the init process if the reset-function is disabled.\nThis makes it ideal for a controlled shutdown, but it does need a\n/bin/init that knows about this.\n\n\t\tLinus\n\nPS. It seems both the DOS-fs and the extended fs will be out for\nalpha-testing next week, so I assume 0.97 will have them both if things\nwork out ok.\n" }, { "commit": "e38481072b4a12a490335fbe1917ffc7bb9ce32e", "tree": "f1b96876efa74c63cbc97887368df7d9470900d5", "parents": [ "104ec363bda00e91a1701e18e691d39a666a0180" ], "author": { "name": "Linus Benedict Torvalds", "email": "torvalds@klaava.Helsinki.FI", "time": "Sun Jun 21 23:12:23 1992 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:05 2007 -0400" }, "message": "Linux v 0.96b available\n\nI have uploaded my newest kernel sources and a US-keyboard image to both\ntsx-11.mit.edu and nic.funet.fi. I\u0027ll put it on banjo too as soon as\nbanjo wakes up (that\u0027s when I\u0027ll do the patches too: I\u0027ll need access to\nbanjo to get a 0.96a.pl4 kernel to make patches against..). As usual,\nit will probably be a day or two until the files have been moved from\nthe incoming directory to their proper places.\n\n0.96b is not a new major release: it\u0027s pretty close to 0.96a with all my\npatches (1-4). However, as there has been 4 patches already, I decided\nit would be time for a full kernel release along with a bootimage, so\nthat people who don\u0027t feel confident with patching can use the new\nfeatures.\n\nIf you already have 0.96a patchlevel 4, 0.96b will offer you these new\nfeatures:\n\n - the math-emulation now handles fsqrt, as gcc-2.2.2 generates that\n inline. I haven\u0027t tested the kernel code at all: I tested the\n algorithm in user space, but I\u0027m lazy, so I never turned off my 387\n to do real testing. I hope it works.\n - better vt100 terminal emulation thanks to Mika Liljeberg.\n - I removed a possible race-condition in the buffer-cache code.\n - minor fixes\n\nThe vt100 emulation should now be complete enough for almost everything\n(including vt100 test suites): as a result the setterm utility had to be\nchanged (as the old setterm codes aren\u0027t compatible with the full vt100\ncodes). setterm-0.96b.tar.Z contains the new setterm.\n\nThe soon-to-be-released gcc-2.2.2 will need the 0.96b kernel: (a) due to\nthe fsqrt emulation and (b) it uses the new stat() system call. So\nupgrading is a good idea. (If you have a co-processor, (a) isn\u0027t used,\nbut (b) still stands)\n\nIf you have an unpatched 0.96a, the differences to 0.96b are roughly\n(not counting the above-mentioned new things):\n\n - corrected the disk-buffer-list bug with read/write-errors\n - fixed read-ahead warning messages at end of disk\n - better support for text-mode restoration after running MGR and X\n - full core-dumping, attach/detach etc debugging features\n - 16550A support\n - less low 1MB memory used for kernel structures\n - various minor fixes\n\nNote that the fact that new versions (pl4 and above) use more memory in\nthe 1M+ area means that linux will report less free memory (it\u0027s used\nfor buffer-cache instead). This could concievably be a problem on 2MB\nmachines. The standard kernel comes with only 4 pty\u0027s though, and if\nyou use the standard 80x25 text modes instead of svga modes, the VC\nbuffers will be smaller. Please contact me if there are problems even\nwith this minimal setup.\n\n0.96b does /not/ contain: the new scsi drivers, new filesystems or some\nother patches I have gotten (ibm character set mode, loop-devices etc).\nIf you have sent me any other patch, you might want to remind me about\nit.\n\n\t\tLinus\n" }, { "commit": "104ec363bda00e91a1701e18e691d39a666a0180", "tree": "d159cdb49511ca4bb29678a4630c1a85fa061602", "parents": [ "2d7bd0389f1494040efcddd5a1b651f362bbf2c9" ], "author": { "name": "Linus Benedict Torvalds", "email": "torvalds@klaava.Helsinki.FI", "time": "Thu Jun 18 14:50:12 1992 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:04 2007 -0400" }, "message": "32-bit inodes: patch4\n\nOk, patch4 implements 32-bit inode numbers (and thus the new\nstat/lstat/fstat system calls), as well as correcting the bad\nrs-performance on some machines that showed up in patch3. It\u0027s\ncurrently only on banjo, but I\u0027ll copy it around eventually.\n\nAgain, you don\u0027t miss much if you don\u0027t use this patch: it\u0027s mainly for\n(a) the serial problems and (b) for hlu etc that want to test out the\n32-bit interface. It does some other magical tricks as well (uses less\nmemory in the low 1M region by moving the screen and tty buffer to high\nmemory), if anybody is interested.\n\n\t\tLinus\n" }, { "commit": "2d7bd0389f1494040efcddd5a1b651f362bbf2c9", "tree": "beffbf6d95378739fe8ea8dfac77cd4d14a13f3a", "parents": [ "13f296db85e43116f5e1de944b6997dd98093640" ], "author": { "name": "Linus Benedict Torvalds", "email": "torvalds@klaava.Helsinki.FI", "time": "Mon Jun 15 20:04:54 1992 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:04 2007 -0400" }, "message": "patch3....\n\nOk, I already announced it on the kernel mailing-list, but I might as\nwell go all the way. I put out patch3 to 0.96a yesterday, and it\u0027s\navailable on banjo in pub/Linux/Linus, and I\u0027ll upload it to the other\nnormal ftp-sites tonight.\n\nNOTE! Patch3 is (like patch2) more of a kernel-hacker patch: it\u0027s just\nin case you want to keep up with my kernel. It has some problems with\nsome serial lines, and if you experience them, I\u0027d like to know what\ntype of chip you are running (and what linux reports on bootup). If you\ndon\u0027t think patching the kernel is fun, you might as well forget this\nand wait for a real release (next month?).\n\nPatch 3 contains:\n\n - support for attaching and detaching processes under gdb (but you need\n a gdb that knows about this).\n - 16550A support\n - full core-dumping (again, you need a gdb that supports it)\n - sockets have no problems with non-root binding etc\n - /dev/zero implemented (mknod /dev/zero c 1 5)\n\nNone of the patches are very big (the whole patch is 17kB compressed,\nmost of it attach/detach code), but they are all pretty useful.\n\nThe 16550A support means that with the appropriate chip you now should\nbe able to use the serial ports at much higher speeds, but as mentioned,\nit seems to break on some machines.\n\nThe detaching isn\u0027t perfect yet (I noticed only after making the diffs\nthat I had forgotten to do some cleanups), but it\u0027s not generally a\nproblem (the code just forgets to give the process back to it\u0027s rightful\nfather).\n\nThe patch is relative to the pl2 kernel, so you have to use the earlier\npatches first. This time, I\u0027ve added the lib/itimer.c code.\n\n16550A support was written by tdavis, the correct format of the\ncore-dumps was written by eric (who also wrote the attach/detach code I\nused as an example when implementing it), /dev/zero was written by\nalmesber. Nice to see good patches: I just did the socket-thing and\nrewrote the attaching to suit me.\n\n\t\tLinus\n" }, { "commit": "13f296db85e43116f5e1de944b6997dd98093640", "tree": "7658905de163908272cc1291b5ad5c858356c7ac", "parents": [ "46d255794351e97e7b06958d2859b6ec7c373f73" ], "author": { "name": "Linus Benedict Torvalds", "email": "torvalds@klaava.Helsinki.FI", "time": "Thu Jun 04 22:56:21 1992 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:04 2007 -0400" }, "message": "Second patch to 0.96a\n\nI have just sent off the second patch to 0.96a: it should be on the\nnormal ftp-sites (nic, tsx-11 and banjo), although the only site which I\ncan make it directly readable on is banjo, so on the other sites it will\ntake the site-managers to make the patch available.\n\nPatch 2 implements:\n\n- itimers (by Darren Senn), which are now also used to implement the\n alarm() system call.\n\n- ultrastor scsi driver patches (by gentzel)\n\n- [f]statfs() system call is implemented (so df can be made fs-\n independent). Also some other minor fs-changes for the upcoming new\n filesystem. Patches by Remy Card.\n\n- preliminary core-file dumping code (linux creates a core-file, but\n it\u0027s not in the correct format yet [*]).\n\n- minor changes/bugfixes.\n\nWhile patching in patch1 is a good idea for anybody, patch 2 isn\u0027t\nreally vital. I\u0027ve made it available just so kernel hackers can keep up\nwith the kernel I have right now if they wish. Patch 2 is relative to\npatch 1: you have to patch that in first.\n\n[*] The current core-file is very simple, and the kernel code is there\njust so that some enterprising character can expand it. A core-file\nlooks like this right now:\n\noffset data\n0x0000 \"core-dump: regs\u003d\\n\"\n0x0040 struct pt_regs (see \u003csys/ptrace.c\u003e)\n0x0400 \"floating-point regs:\\n\"\n0x0440 struct i387 (see \u003clinux/sched.h\u003e)\n0x0800 the first 1kB of user-space\n\nNot very practical, but it /might/ help if the X-server dies of a\nsegmentation fault or similar (you can use pt_regs.eip to see where it\nhappened). The kernel code is very easy to change to accomodate for the\nreal core-file format, I just didn\u0027t know what it should be.\n\n\t\tLinus\n" }, { "commit": "46d255794351e97e7b06958d2859b6ec7c373f73", "tree": "e05f50c73ea148da4276f0343999a3038f913ddb", "parents": [ "8404560378fc29692399a65c34d39d80699f7288" ], "author": { "name": "Linus Benedict Torvalds", "email": "torvalds@klaava.Helsinki.FI", "time": "Sun May 24 11:46:43 1992 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:04 2007 -0400" }, "message": "linux-0.96a.patch1\n\nHere is the patch to 0.96a that corrects the harddisk error bug, dup2()\nand X11 text-mode restoration. Thanks to Rick Sladkey for finding the\ndup2() bug.\n\nThis patch has also been sent to the ftp-archives.\n\n\t\tLinus\n" }, { "commit": "8404560378fc29692399a65c34d39d80699f7288", "tree": "e3ef2741645be87422fb3cfcef160aa137c02dac", "parents": [ "30cd071cd0121a1f0a965ce3a533131d55e9a1b4" ], "author": { "name": "Linus Benedict Torvalds", "email": "torvalds@klaava.Helsinki.FI", "time": "Fri May 22 21:01:59 1992 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:04 2007 -0400" }, "message": "0.96a available\n\nOh, well, no use in delaying it any more, so I sent out my latest\nrelease to nic.funet.fi, tsx-11.mit.edu and banjo.concert.net. They\nshould show up in the next couple of days (they are already visible on\nbanjo: /pub/Linux/Linus). I hope all the bugs got fixed, but I did\nsomething potentially stupid:\n\nI had expected that lankaster wouldn\u0027t get his hd-speedup patches ready\nfor 0.96a, and I was resigned to the same hd-performance as with all\nolder releases. But when I saw them on the newsgroup today I thought\nI\u0027d try them out just in case, as I could always use my backup-version\nif they backfired...\n\nThe point here is that the patch ended up in 0.96a after some minor\nedits by me (one benign bug and some editing due to changed files). So\nnow hd-performance is much better on most operations. I just hope it\ndoesn\u0027t result in yet another release just to fix new bugs (but I doubt\nit: the patches were really minor and clean - no ugly hacks needed I\u0027m\nhappy to say). Branko already posted benchmarks, and I can only confirm\nthat it\u0027s indeed snappier, especially on writes. syncing is no longer a\npain even after heavy writing.\n\nOther than the hd-performance, there are no new features I haven\u0027t\nalready mentioned in other posts. Bug-fixes, rewrites in C, better\ndebugging. I haven\u0027t made the cdiffs yet, so right now the new release\nis only available as complete source and as a binary, but I\u0027ll try to\nget patches done tonight. Possibly tomorrow. The patches will be\nagainst the original 0.96, and shouldn\u0027t be too big.\n\n\t\tLinus\n\nPS. No need for more 16550 info: even if somebody doesn\u0027t implement it\nfor the next release, I think I can get it going. Doesn\u0027t seem too\nhard. I\u0027ll also start to look into core-files. Eventually. Promise.\n\nPPS. Ja 0.96a on taas saatavilla klaavasta /usr/tmp/linux\u0027ssa\nyliopistolla kirjoilla oleville.\n\n-----\n\nFrom: torvalds@klaava.Helsinki.FI (Linus Benedict Torvalds)\nSubject: 0.96a out today/tomorrow\nDate: 21 May 1992 12:19:55 GMT\n\nThe subject say it all: I haven\u0027t heard any bad results (except somewhat\nworse performance on the serial lines) from the testimage testers, so\nI\u0027ll make 0.96a available tonight (Finnish time: EST) or tomorrow.\ntsx-11, nic and banjo will be the sites I\u0027ll upload it to.\n\n0.96a doesn\u0027t have any major new features: I just tried to correct known\nbugs and get a stable version out the door. I hope 0.96 won\u0027t need any\nmore interim releases, but we\u0027ll see.\n\n0.96a has gone back to non-interruptible hd-interrupts, and as long as\nI\u0027m not sure about the reason for the need of this, it\u0027s going to stay\nthat way. I hope this new version will run on all drives that have been\nsupported in the past, and the only known remaining problem is the Kalok\ndrives which have spurious interrupts. That\u0027s definitely a hardware\nproblem, and Branko Lankaster has patches that fake away these..\n\n0.96a should also correct most of the serial line problems, and the\nkeyboard driver is now in C - I expect somebody will change it to allow\nrun-time keyboard changing. Additionally, there are various minor fixes\nin various places:\n\n- the kill fix by tytso that allows any process to wake up a stopped\n process as long as it\u0027s in the same session.\n\n- chown fix by card and me. Users can chgrp files they own to any group\n they belong to etc - _POSIX_CHOWN_RESTRICTED should now be implemented\n correctly. Also, chown() on a symlink now actually chown\u0027s the symlink\n and not the file it points to (but GNU fileutils don\u0027t seem to want to\n chown symlinks anyway - or maybe I have an older version)\n\n- SIGSEGV fix by lankaster (?) which allows for better debugging after\n protection faults. Most such errors are now trappable in gdb\n (although some circumstances can still result in the process exiting\n at once: out-of-memory errors etc)\n\nI have also changed the memory manager to check for illegal memory\nreferences (ie using the area between brk and stack), so debugging bad\npointer references should be much easier. This fix will hopefully also\ntrap the uncompress problem with bad files, but I haven\u0027t tried it out.\nPrograms can still call brk() with any value, so it doesn\u0027t mean that\nyou can\u0027t use up all available memory, but at least this should fix\n/bad/ memory references. My limited tests with gdb seems to indicate it\nall works nicely.\n\nCore files are still not generated (I haven\u0027t got the slightest idea\nwhat format they should have), so debugging isn\u0027t perfect yet, but it\nshould certainly be much easier.\n\n\t\tLinus\n\n-----\n\n[ Committer\u0027s note: this commit also folds the lost 0.96 release. ]\n\nFrom: torvalds@klaava.Helsinki.FI (Linus Benedict Torvalds)\nSubject: 0.96 uploaded\nDate: 12 May 1992 16:51:12 GMT\n\nWell, the title says it all: I\u0027ve sent off 0.96 to banjo.concert.net,\nwhere it can be found in pub/Linux/Linus along with a new bootimage and\na program for formatting floppies under linux (which works only under\n0.96).\n\nI\u0027ve also sent it to tsx-11.mit.edu and I\u0027ll send it to nic as soon as\nthe lines clear up.\n\nGeneral warning about 0.96:\n\n- the scsi code is in the kernel, but I haven\u0027t personally tested it,\n so who knows... The SCSI code also results in a 4-5 second pause at\n bootup with the current bootimage, while it searches for an adapter:\n if you find this disturbing, you have to recompile the kernel with the\n appropriate changes to config.h.\n\n- The harddisk timings have changed: the testimage got a generally good\n review, but it hasn\u0027t been tested very much. The changes seem to help\n at least some \"HD times out\" problems, but there might be new bugs..\n\n- The serial code was totally rewritten this weekend, and I haven\u0027t\n tested it out under any heavier load. I found one bug as late as\n today, and there might be others lurking around.\n\n- There have been generally pretty heavy rewrites: it\u0027s binary\n compatible with the old kernels, but the changes might not all be\n correct. Oh, well.\n\nThat said, I hope 0.96 will be an improvement on earlier versions, and\nmost of the old bugs corrected. If the new version still has some\nproblem - please mail me with a new bugreport. Otherwise I\u0027ll just\nassume the problem went away: I\u0027m afraid don\u0027t have time to go through\nold mail searching for any bugs that might still be in there.\n\nPartial list of features:\n\n- automatic floppy detection. Please add the following devices:\n\n\tmknod /dev/fd0 b 2 0\n\tmknod /dev/fd1 b 2 1\n\n which act as A and B floppies respectively, finding out automatically\n what kind of disk there is.\n\n The floppy driver now also contains a timeout, so an empty diskdrive\n no longer results in a floppy driver hang.\n\n- serial lines now support dropping DTR on closing, and sending SIGHUP\n to the process group that is logged in on a serial line. It\u0027s also a\n lot easier to change the interrupts etc of the lines.\n\n- unix sockets supported for X, as well as mmap() on /dev/mem etc.\n\n- pty\u0027s corrected. Hopefully no more hangs under X due to pty trouble.\n\n- better IO-performance when there are computationally intensive jobs in\n the background or on another VC. Partly due to new scheduler\n mechanism, partly due to read-ahead on normal files.\n\n- cleaned up vfs layer.\n\n- no more mismatched children\n\n- minor corrections all over the place.\n\nThe new release is in fact different enough that there is no use trying\nto make context diffs: files have disappeared, others are new, others\nhave simply changed a lot. Even compared to pre-0.96 there has been\nquite a lot of changes.\n\n\t\tLinus\n\n-----\n\nFrom: torvalds@klaava.Helsinki.FI (Linus Benedict Torvalds)\nSubject: 0.96 out next week\nDate: 4 May 1992 07:38:03 GMT\n\nOk, the subject says most of it: I\u0027ll send out 0.96 sometimes next week\n(ie 92.05.11-17), and this is just an announcement to confirm that.\n\n0.96 has a lot of changes (even relative to pre-0.96), and it\u0027s entirely\npossible that making it available as cdiffs isn\u0027t feasible. It contains\na lot of new files, as well as some re-organizations in the old ones.\n\nMain new things are:\n\n- The SCSI distribution is now in the standard package. I (obviously)\n haven\u0027t been able to test my patchings, so there might be problems in\n this first release. I had to do some changes \"blind\" to the cdiffs,\n but most of them were pretty trivial.\n\n- X11r5 as ported by obz is supported. It\u0027s still in beta-testing (join\n the X11-channel on the original mailing-list), but as I\u0027m writing this\n from an xterm under linux, it works pretty well. Changes to pre-0.96\n are just the socket-code by obz, and some small tweaking by me.\n\n- Hopefully better interrupt latency - I\u0027ve changed select() not to use\n cli-sti, and most IRQ\u0027s to enable interrupts, and instead disable just\n their own interrupt-line. The interrupt latency has been noticeable at\n higher serial speeds, and I hope 0.96 will be better in this respect.\n Again, I only have 2400bps, so I\u0027ve never seen the problems, and\n cannot guarantee the new version will help. (btw, I hope the problems\n with select are gone now)\n\n- Reorganisation of the vfs routines and minix filesystem driver. These\n shouldn\u0027t bother anyone but people that have implemented their own\n filesystems (I know of just 2 to date), but I hope the current\n vfs-interface will prove to be relatively stable. The new vfs\n interface has made some things much cleaner, and the promised cleanup\n of special devices has happened.\n\n- ps/uptime patches + added readahead, so having computationally\n intensive background processes isn\u0027t as noticeable any more when doing\n IO.\n\nAdditionally, there /might/ be a new floppy-driver that supports\nformatting and autodetecting floppies, but I haven\u0027t had time to check\ninto it yet.\n\nThere are probably any number of minor changes: I\u0027ve lost track. People\nhave sent me some diffs, and some of them went in, depending on how\nenergetic I was that day. I\u0027ve tried to correct all the bugs I\u0027ve\ngotten reports on, and hopefully 0.96 will work with just about\neverything (gdb etc).\n\nThings I wanted, but didn\u0027t have time for:\n\n- The config patches aren\u0027t there. Sorry everybody. That means still no\n wd8003 driver etc.\n\n- Any number of minor patches (quota, auto-SVGA etc)\n\nGenerally, 0.96 is cleaning some things up, but on the other hand the\nnew features can have their share of problems. We\u0027ll see. Anyway, most\nthings seem to work, and I hope there won\u0027t be the same type of problems\nas with the first 0.95 release.\n\n\t\tLinus\n" }, { "commit": "30cd071cd0121a1f0a965ce3a533131d55e9a1b4", "tree": "781ba7a0989b4cba28b08cc03f2757221360482f", "parents": [ "b5de4a78ce363e94ad92d82c6ef2d6909bdd00dd" ], "author": { "name": "Linus Benedict Torvalds", "email": "torvalds@klaava.Helsinki.FI", "time": "Tue Apr 21 23:15:10 1992 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:03 2007 -0400" }, "message": "pre-0.96 (was Re: gdb still isn\u0027t working)\n\nIn article \u003c1992Apr20.085143.23027@klaava.Helsinki.FI\u003e\ntorvalds@klaava.Helsinki.FI (Linus Benedict Torvalds) writes:\n\u003e [ trace not working in gdb ]\n\u003e\n\u003eMy personal version handles this correctly (as well as doing some other\n\u003ethings in a cleaner manner), but I\u0027m not quite ready for a new release\n\u003eyet. I could make YAAR (yet another alpha-release) or just mail\n\u003einterested parties the fixes needed - mail me if you\u0027re interested, and\n\u003edepending on the number of messages I get I\u0027ll make it a new release.\n\nOk, the response seems to make a new pre-release appropriate: I have\nuploaded \"pre-0.96.tar.Z\" to tsx-11 and nic.\n\nHere is what the pre-release contains:\n\n- truncate/ftruncate/fchmod/fchown system calls\n\n\tnote that there aren\u0027t any library functions for these, so they\n\taren\u0027t very useful yet...\n\n\t[f]truncate needed a change in the logic of the internal\n\ttruncate VFS call - anybody that has any nonstandard filesystem\n\tprobably needs to look it up.\n\n- io-bitmap syscalls giving root-processes access to selected io ports\n from user space. There is a \"ioperm()\" system call that lets the\n process select which ports it wants to enable/disable (all ports\n disabled as default) as well as a (standard sysv?) ioctl interface\n that X uses.\n\n\tagain, no library stubs, but it allows things like reading and\n\tsetting the cmos clock without using /dev/port, as well as\n\tcontrol over the VGA registers...\n\n- mmap for /dev/mem\n\n\tmore things needed for X...\n\n- the signal-handling fixes needed for gdb\n\n\tThese aren\u0027t yet complete: serial lines still send signals under\n\tinterrupts that can result in problems (ie ptrace doesn\u0027t\n\tcorrectly get them), but that\u0027s pretty unlikely (and will be\n\tfixed in the final 0.96). Breakpoints should work etc..\n\n- multiple shared libraries\n\n\tUp to 6 simultaneous shared libraries/process: the patches were\n\toriginally by pmacdona, but they were heavily changed by me, and\n\tI think they work in a more natural manner now. One user-level\n\tchange is that the libraries are now checked for read and\n\texecute permissions for safety-reasons.\n\n- cleaned up special files.\n\n\tread/write/ioctl no longer has special-case code: it is all\n\thandled with tables to functions. This will mean that the SCSI\n\tpatches won\u0027t patch in quite cleanly into 0.96: you\u0027ll need to\n\tadd the code that sets up the functions.\n\n\tAgain: device drivers and vfs-filesystem hackers need to look\n\tinto the changes, although they are pretty logical (earlier\n\tversions just didn\u0027t implement all the vfs-routines)\n\n\tNote that the vfs-code for select is still not used: select is\n\thardcoded for the devices it supports right now.\n\n- ptrace() has a new interface\n\n\tas gdb for versions \u003c 0.95c don\u0027t work on the new version, and\n\tgdb won\u0027t work very well at all on 0.95c[+], there was no reason\n\tnot to break ptrace. Thus 0.96 has a new calling convention for\n\tptrace, and the old ptrace library function no longer works.\n\tI\u0027m including the new ptrace library function at the end of this\n\tpost.\n\n- mount() takes 4 arguments, and checks that only the super-user can\n mount/umount things.\n\n\tHappily this shouldn\u0027t break any old binaries.\n\n- some general cleanups\n\nI\u0027ve made the pre-release available only as pure source code: no diffs,\nno binary. The reason is that most people that needed this release want\nit for the gdb-fixes: and they should have no problem recompiling the\nkernel. Others just have to wait for the real 0.96.\n\nChanges that are NOT in this pre-release, but which I hope to have in\nthe real 0.96:\n\n\t- more include-file cleanups - I\u0027m still working on these\n\n\t- the wd8003 driver and hopefully some other parts of biro\u0027s\n\t config.\n\n\t- select() using the vfs-tables.\n\nAnd possibly bugfixes that people find in this pre-release...\n\n\t\tLinus\n\n\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d library ptrace.c (wants gcc\u003d2.1) \u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\n#define __LIBRARY__\n#include \u003ctime.h\u003e\n#include \u003cunistd.h\u003e\n\nint ptrace(int request, int pid, int addr, int data)\n{\n\tlong ret;\n\tlong res;\n\n\tif (request \u003e 0 \u0026\u0026 request \u003c 4)\n\t\t(long *)data \u003d \u0026ret;\n\t__asm__ volatile (\"int $0x80\"\n\t\t:\"\u003da\" (res)\n\t\t:\"0\" (__NR_ptrace),\"b\" (request), \"c\" (pid),\n\t\t \"d\" (addr), \"S\" (data)\n\t\t: \"si\",\"bx\",\"cx\",\"dx\");\n\tif (res \u003e\u003d 0) {\n\t\tif (request \u003e 0 \u0026\u0026 request \u003c 4) {\n\t\t\terrno \u003d 0;\n\t\t\treturn (ret);\n\t\t}\n\t\treturn (int) res;\n\t}\n\terrno \u003d -res;\n\treturn -1;\n}\n" }, { "commit": "b5de4a78ce363e94ad92d82c6ef2d6909bdd00dd", "tree": "ac9843150c2941e31a19baa5774900e5654e196d", "parents": [ "4fd4c0232da4a9f7e163df707699cd74c1dc6fa8" ], "author": { "name": "Linus Benedict Torvalds", "email": "torvalds@klaava.Helsinki.FI", "time": "Wed Apr 08 09:03:21 1992 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:02 2007 -0400" }, "message": "linux-0.95c known bugs\n\nWell, only one known bug so far, but a couple of problems. I thought I\u0027d\nmention them before anyone else does, and we\u0027ll call them \"features\" :^)\n\nThe BUG:\n\n when using the readdir() system call, linux incorrectly doesn\u0027t let go\n of the last buffer used for reading: this results in the buffer being\n marked as used (and if you use readdir() heavily, the counter will\n eventually wrap around, which might result in incorrect marking as\n \"not used\"). This bug happily isn\u0027t easy to find: no current binary\n uses the readdir() system call unless you have gotten your hands on\n the new VFS gcc-2.1 release. Thanks to Remy Card for finding this\n one.\n\n Not too bad a bug though: the fix is very easy. Add a \u0027brelse(bh);\u0027\n in the minix_readdir() function, like this:\n\n\tif (i) {\n\t\tput_fs_long(de-\u003einode,\u0026dirent-\u003ed_ino);\n\t\tput_fs_byte(0,i+dirent-\u003ed_name);\n\t\tput_fs_word(i,\u0026dirent-\u003ed_reclen);\n\t\treturn i;\n\n should really be :\n\n\tif (i) {\n\t\tput_fs_long(de-\u003einode,\u0026dirent-\u003ed_ino);\n\t\tput_fs_byte(0,i+dirent-\u003ed_name);\n\t\tput_fs_word(i,\u0026dirent-\u003ed_reclen);\n+ brelse (bh);\n\t\treturn i;\n\n That is, just add the brelse before the early return. Sorry for the\n lack of real cdiffs - I\u0027m not at home right now, and the above was\n taken directly from the bug-report mail.\n\nThe problems:\n\n I\u0027ve had one report that the floppy-driver in versions 0.95x breaks\n when accessing drive nr 2. It doesn\u0027t on my machine, but I\u0027d\n appreciate it if people would test it out, and mail me about any\n problems. So far, only one report, but that\u0027s one too many.\n\n 0.95c doesn\u0027t correctly keep track of the \u0027rss\u0027 field in the\n task-structure. A fix was already posted (and nothing breaks even if\n you don\u0027t apply the fix - ps just gives slightly incorrect output)\n\n And the expected troubles with the change of \u0027a.out.h\u0027 - the old gdb\n doesn\u0027t recognize new executables etc.\n\nAs soon as I get my own sources cleaned up, I\u0027ll send out a new binary\nfor the 0.95c+ kernel to the ftp-sites. I\u0027ve gotten a few mails from\npeople unable to recompile everything - either because of lack of\ndiskspace or some other problem. Tomorrow I\u0027ll put the new kernel image\non nic.funet.fi and tsx-11 - it\u0027s basically the 0.95c kernel + the above\nbugfix + the lp-patches (somewhat edited by me).\n\n\t\tLinus\n\n\u003cREADME-0.95c+\u003e\n\nThis is release 0.95c+ of the linux kernel - it contains some\nenhancements and bugfixes to the 0.95a kernel, as well as some minor\nfixes relative to the last alpha-patch (0.95c). The release is\navailable as\n\n- binary\t\t(bootimage-0.95c+.Z)\n- full source\t\t(linux-0.95c+.tar.Z)\n- patches rel. to 0.95c\t(diff-0.95c.c+.Z)\n- patches rel. to 0.95a\t(diff-0.95a.c+.Z)\n\nNOTE TO PATCHERS!! Before patching, do this:\n - make an empty include-file linux/include/checkpoint.h\n - rename linux/kernel/math/math_emulate.c as just emulate.c\nThat is, from the linux source directory do:\n\n\t$ \u003e include/checkpoint.h\n\t$ mv kernel/math/math_emulate.c kernel/math/emulate.c\n\nAlso note that patching from the 0.95a version is probably not worth it\nas it\u0027s easier to get the complete new sources.\n\nAlthough I\u0027m making binaries and the full source available, this isn\u0027t\nreally a major release: there is no new rootdisk, and this is more \"my\ncurrent kernel\" and not really tested (I put in the last changes 5\nminutes before packing all this up).\n\nThe reason I\u0027m making this available is that with the advent of gcc-2.1\nand the VFS-library the old kernel doesn\u0027t really do everything the new\nlibraries want: the readdir system call is needed to get things working.\nThe default compiler after this release is considered to be gcc-2.0 or\nhigher (although 1.40 still works - you don\u0027t /have/ to change). People\nwho are unable or unwilling to patch a new kernel shouldn\u0027t be unable to\nrun the new binaries.\n\nThis kernel should be totally backwards compatible, so no binaries\nshould break. I resisted adding the changed mount() system call into\nthis release: the next major release will have a third parameter for\nmount() - the filesystem type name (ie mount /dev/xxx /mnt minix).\n\nFixes relative to 0.95c:\n\n- corrected two minor bugs in readdir() (thanks to R Card)\n\n- lp-patches are in. I\u0027ve edited them a bit, and will probably do some\n more editing in the future, but they seem to work fine.\n\n- 8-bit ISO latin output to the console (ie part of Johan Myreens\n general latin-1 patches: the keyboard patches aren\u0027t there)\n\n- other minor bug-fixes (thanks to HH Bergman for noticing the\n timer-table bug)\n\nThings I haven\u0027t had time to look into yet:\n\n- select still has some problems\n- reports that VC-output sometimes isdiscarded (never seen it myself)\n- probably other things I\u0027ve simply forgot...\n\n\t\t\tLinus\n" }, { "commit": "4fd4c0232da4a9f7e163df707699cd74c1dc6fa8", "tree": "87ac70daf7404f03dcd0957ce9e255f232feb82c", "parents": [ "edb046236801c4975482840240127aeefdd0f7d7" ], "author": { "name": "Linus Benedict Torvalds", "email": "torvalds@klaava.Helsinki.FI", "time": "Sat Apr 04 14:42:10 1992 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:02 2007 -0400" }, "message": "Second 0.95a alpha-patch\n\nThis is the promised patch to 0.95a, which hopefully corrects some of\nthe problems encountered. This is /not/ an offical new release: it\u0027s\njust a set of patches to get the same kernel I am currently running.\n\nBugfixes:\n\n- extended partitions should finally work correctly (this release also\n contains code for the hd-ioctl call, needed for fdisk). Code mostly\n by hedrick.\n\n- I corrected my original ptrace-fix (writing a long word to another\n process\u0027 data space could fail with my original patches)\n\n- 387-emulation bug with the instructions \"fcom[p] %st(x)\" which\n resulted in bad results on non-387 machines with newer versions of\n gcc. The emulation is still ugly, but it seems to work.\n\n- the cooked mode deletion/linekill bugs should be fixed.\n\n- various error-returns were wrong: I correted some of them (thanks to\n bruce evans who pointed them out). The bad error-values resulted in\n incorrect or spurious error-messages from \u0027rm\u0027 etc.\n\n- various minor fixes (including some in the hd-driver: this might help\n persons with unexpected-interrupt and/or timeout errors)\n\nAdditionally this version contains VFS-code from entropy, and a\nreaddir() system call needed for the VFS. The latter was inspired by\npatches sent by Remy Card, who did it with a getdirents sys-call. My\nversion is slightly simpler, but is probably slower. Things might yet\nchange.\n\nThe installation has also changed slightly: the keyboard type and\nmath-emulation are specified in the main Makefile. That one also\ncontains the -fcombine-regs flag needed for 1.40. The other makefiles\nshould no longer need editing.\n\nI\u0027ve also incorporated the ps095 kernel patches: to get the actual\nuser-level stuff you still have to get the ps-distribution. Printer\nports /still/ aren\u0027t in there, but this time I positively /promise/ to\nput it in next week. Really.\n\nPeople who have been patching their kernel might have problems getting\nthis patch to work: it was made against a clean 0.95a kernel. I\u0027ll\nconsider a patched-up kernel version 0.95c - and I\u0027d appreciate if\nfuture patches to me would be sent against this version. I\u0027ll still\naccept older patches, or course.\n\n\t\tLinus\n" }, { "commit": "edb046236801c4975482840240127aeefdd0f7d7", "tree": "f204d177cb6e85af83b46c2ee9653677a40be9cb", "parents": [ "741d52384b75d810b8a04d7f16733b9a6ff809a5" ], "author": { "name": "Linus Benedict Torvalds", "email": "torvalds@klaava.Helsinki.FI", "time": "Wed Mar 18 00:32:46 1992 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:02 2007 -0400" }, "message": "0.95a uploaded - soon available\n\nI just uploaded the new kernel source and image to nic.funet.fi into the\nincoming-directory under linux. As usual, they won\u0027t show up for a day\nor two: wait for arl to announce that it\u0027s available.\n\nThe new version called 0.95a has no major new features, and is just a\nbug-fix release for 0.95 - it hopefully fixes all known bugs, but that\u0027s\nwhat I thought about plain 0.95 :(. There are even less installation\ndocuments than before: (a) they haven\u0027t changed since 0.95 and (b) the\nrootdisk is no longer made by me (which means that 0.95a will probably\nbe the first release that actually has all the needed device special\nfiles there.)\n\nEveryone\u0027s favourite bug is also fixed: 0.95a actually returns the\ncorrect version number from uname(). (Boy did I get bug reports on this\none :^)\n\nI\u0027m including the release-note at the end of the post.\n\n\t\tLinus\n\nPS: Whoever made the GNU-emacs binary available for linux - the\ninability to use the \"meta-X shell\" command is a bit disturbing. I\nlooked into it a bit, and my guess is that the linux config file doesn\u0027t\ndefine HAVE_SETSID. Could somebody with more diskspace than I have try\nto recompile gemacs with HAVE_SETSID defined, and tell me if this fixes\nthe problem?\n\n\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\n\n\t\tRELEASE NOTES FOR LINUX v0.95a\n\t\tLinus Torvalds, March 17, 1992\n\nThis is file mostly contains info on changed features of Linux, and\nusing old versions as a help-reference might be a good idea.\n\n\t\tCOPYRIGHT\n\nLinux-0.95a is NOT public domain software, but is copyrighted by me. The\ncopyright conditions are the same as those imposed by the GNU copyleft:\nget a copy of the GNU copyleft at any major ftp-site (if it carries\nlinux, it probably carries a lot of GNU software anyway, and they all\ncontain the copyright).\n\nThe copyleft is pretty detailed, but it mostly just means that you may\nfreely copy linux for your own use, and redistribute all/parts of it, as\nlong as you make source available (not necessarily in the same\ndistribution, but you make it clear how people can get it for nothing\nmore than copying costs). Any changes you make that you distribute will\nalso automatically fall under the GNU copyleft.\n\nNOTE! The linux unistd library-functions (the low-level interface to\nlinux: system calls etc) are excempt from the copyright - you may use\nthem as you wish, and using those in your binary files won\u0027t mean that\nyour files are automatically under the GNU copyleft. This concerns\n/only/ the unistd-library and those (few) other library functions I have\nwritten: most of the rest of the library has it\u0027s own copyrights (or is\npublic domain). See the library sources for details of those.\n\n\t\tNEW FEATURES OF 0.95a\n\n0.95a is mainly a bug-fix release: it didn\u0027t even get it\u0027s own version\nnumber. Plain 0.95 fixed a lot of bugs in 0.12, but also introduced\ntotally new bugs: 0.95a tries to correct these. The bugs corrected\n(knock wood) are:\n\n- floppy and harddisk drivers should now once more work with most\n hardware: I\u0027d be interested in reports of \"unexpected HD interrupt\"\n and \"reset-floppy called\" with the new kernel.\n\n- A rather serious tty-bug corrected: this one messed up the screen\n under 0.95, and switched characters over the serial lines. Under\n extreme circumstances it could even crash the machine.\n\n- ptrace had a bug: hopefully it works now.\n\n- The extended partitions didn\u0027t work under 0.95, although most of the\n code was there. Please somebody tell me it works under 0.95a.\n\n- the 0.95 fdisk was broken: a new one with the new root-floppy should\n clear up the confusion.\n\n- select() and the sleep-wakeup code had fundamental (but relatively\n benign) problems under 0.95 (and all earlier versions). The sleeping\n code is totally redesigned, and select should work better even under\n load.\n\nOne actual new feature, not just a bug-fix:\n\n- ser3-4 support is there, although I\u0027ve been unable to test it (as I\n haven\u0027t got more than ser2). NOTE! Due to AT hardware limitations,\n ser1 cannot be active at the same time as ser3, and likewise ser2 and\n ser4 are mutually exclusive. The interrupt-handlers should have no\n problems with shared interrupts, but the actual hardware probably has,\n so the kernel disables interrupts from one serial line when the other\n one is opened.\n\n- faster default keyrepeat rate: this is going to need some getting used\n to, but is extremely practical especially with bigger screen sizes.\n\n- VGA cards that aren\u0027t recognized at bootup are put into the 80x50\n character mode if \u003center\u003e was pressed when asking about SVGA modes.\n\n\t\tNEW FEATURES OF 0.95\n\n\tInit/login\n\nYeah, thanks to poe (Peter Orbaeck (sp?)), linux now boots up like a\nreal unix with a login-prompt. Login as root (no passwd), and change\nyour /etc/passwd to your hearts delight (and add other logins in\n/etc/inittab etc).\n\n\tVirtual consoles on any (?) hardware.\n\nYou can select one of several consoles by pressing the left alt-key and\na function key at the same time. Linux should report the number of\nvirtual consoles available upon bootup. /dev/tty0 is now \"the current\"\nscreen, /dev/tty1 is the main console, and /dev/tty2-8 can exist\ndepending on your text-mode or card.\n\nThe virtual consoles also have some new screen-handling commands: they\nconfirm even better to vt200 control codes than 0.11. Special graphic\ncharacters etc: you can well use them as terminals to VMS (although\nthat\u0027s a shameful waste of resources), and the PF1-4 keys work somewhat\nin the application-key mode.\n\n\tExtended vt200 emulation\n\n0.95 contains code to handle a vt200 application keymap mode: the cursor\nkeys send slightly different codes when in application mode, and the\nnumeric keyboard tries to emulate the vt200 application keys. This\nprobably isn\u0027t complete yet.\n\n\tSymbolic links.\n\n0.95 now allows symlinks to point to other symlinks etc (the maximum\ndepth is a rather arbitrary 5 links). 0.12 didn\u0027t like more than one\nlevel of indirection.\n\n\tVirtual memory.\n\nVM under 0.95 should be better than under 0.12: no more lockups (as far\nas I have seen), and you can now swap to the filesystem as well as to a\nspecial partition. There are two programs to handle this: mkswap to set\nup a swap-file/partition and swapon to start up swapping.\n\nmkswap needs either a partition or a file that already exists to make a\nswap-area. To make a swap-file, do this:\n\n\t# dd bs\u003d1024 count\u003dNN if\u003d/dev/hda of\u003dswapfile\n\t# mkswap swapfile NN\n\nThe first command just makes a file that is NN blocks long (initializing\nit from /dev/hda, but that could be anything). The second command then\nwrites the necessary setup-info into the file. To start swapping, write\n\n\t# swapon swapfile\n\nNOTE! \u0027dd\u0027 isn\u0027t on the rootdisk: you have to install some things onto\nthe harddisk before you can get up and running.\n\nNOTE2! When linux runs totally out of virtual memory, things slow down\ndramatically. It tries to keep on running as long as it can, but at\nleast it shouldn\u0027t lock up any more. ^C should work, although you might\nhave to wait a while for it..\n\n\tFaster floppies\n\nOk, you don\u0027t notice this much when booting up from a floppy: bash has\ngrown, so it takes longer to load, and the optimizations work mostly\nwith sequential accesses. When you start un-taring floppies to get the\nprograms onto your harddisk, you\u0027ll notice that it\u0027s much faster now.\nThat should be about the only use for floppies under a unix: nobody in\ntheir right mind uses floppies as filesystems.\n\n\tBetter FS-independence\n\nHopefully you\u0027ll never even notice this, but the filesystem has been\npartly rewritten to make it less minix-fs-specific. I haven\u0027t\nimplemented all the VFS-patches I got, so it\u0027s still not ready, but it\u0027s\ngetting there, slowly.\n\n\tAnd that\u0027s it, I think.\n\nHappy hacking.\n\n\t\t\tLinus (torvalds@kruuna.helsinki.fi)\n" }, { "commit": "741d52384b75d810b8a04d7f16733b9a6ff809a5", "tree": "0c1c295237dfd711541e0a70bebfc4a7823d6da1", "parents": [ "26d73589f54a0ffa39ca6b16cfe4a46a63f6b781" ], "author": { "name": "Linus Benedict Torvalds", "email": "torvalds@klaava.Helsinki.FI", "time": "Sun Mar 08 14:43:12 1992 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:02 2007 -0400" }, "message": "Linux-0.95\n\nAll right: it\u0027s three days late, but finally 0.95 has been sent to\nnic.funet.fi. As usual, it will probably take a few days to find it\u0027s\nway into a readable directory, so don\u0027t start ftp\u0027ing right now. I\u0027m\nsure arl will inform people when it\u0027s available.\n\nI\u0027m also pretty sure there will be problems setting things up again:\nsome things have changed, and the docs are up to their usual wonderful\nstandard. For people that have used 0.12, there shouldn\u0027t be too many\nsurprises, although the new harddisk names/numbers can be confusing.\nMany bugs have been corrected, but there are probably new ones that have\ntaken their place.\n\nOne bad thing with the new setup (which will confuse new users) is that\nthe rootdisk finally got too small for everything, and compress+tar\naren\u0027t on the disk any more. Talk about confusing, but if you have 0.12\ninstalled on your system, everything should be mostly a case of \"boot\nfrom floppy, drop in the new things, and reboot with the harddisk\".\n\nRe: patches. There were 5 major patches (VC\u0027s, VFS, swapon, ptrace and\nfaster floppies) that were installed, and of these only ptrace and VC\u0027s\ngot installed without any major changes (pmacdona has had three releases\nto learn my coding style, and it seems to have paid off :). The other\npatches are so heavily edited as to be totally unrecognizeable, and I\nexpect my changes weren\u0027t always for the better: but I\u0027d rather be\nover-conservative than use a patch even if it\u0027s great. I expect they\nwill still have to be edited, and hope none of the authors mind me\nchanging their code heavily (sorry also to all those that have used the\nVFS routines: they\u0027ll have to wait for the next release before getting\nall the functionality in the alpha-VFS patches.)\n\nRe: recompiling the kernel. I\u0027ve used most of the last week to make the\nkernel compileable by gcc-2, as that is what I use now. The bootimage\nI\u0027ve made available is compiled totally with 2.0, and the source-files\nare set up for that compiler. For people without gcc-2, the kernel can\nstill be compiled with 1.40, but you\u0027ll have to add the flag\n-fcombine-regs to some makefiles (without it gcc-1.40 runs out of\nregisters, and aborts).\n\nI\u0027m including the file RELNOTES-0.95, which is just the old 0.12\nRELNOTES edited for a new release. I\u0027m lazy.\n\n\t\tLinus\n\n\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\n\n\t\tRELEASE NOTES FOR LINUX v0.95\n\t\tLinus Torvalds, March 7, 1992\n\nThis is file mostly contains info on changed features of Linux, and\nusing old versions as a help-reference might be a good idea.\n\n\t\tCOPYRIGHT\n\nLinux-0.95 is NOT public domain software, but is copyrighted by me. The\ncopyright conditions are the same as those imposed by the GNU copyleft:\nget a copy of the GNU copyleft at any major ftp-site (if it carries\nlinux, it probably carries a lot of GNU software anyway, and they all\ncontain the copyright).\n\nThe copyleft is pretty detailed, but it mostly just means that you may\nfreely copy linux for your own use, and redistribute all/parts of it, as\nlong as you make source available (not necessarily in the same\ndistribution, but you make it clear how people can get it for nothing\nmore than copying costs). Any changes you make that you distribute will\nalso automatically fall under the GNU copyleft.\n\nNOTE! The linux unistd library-functions (the low-level interface to\nlinux: system calls etc) are excempt from the copyright - you may use\nthem as you wish, and using those in your binary files won\u0027t mean that\nyour files are automatically under the GNU copyleft. This concerns\n/only/ the unistd-library and those (few) other library functions I have\nwritten: most of the rest of the library has it\u0027s own copyrights (or is\npublic domain). See the library sources for details of those.\n\n\t\tINSTALLATION\n\nThis is a SHORT install-note. The installation is very similar to 0.11\nand 0.12, so you should read INSTALL-0.11 too. There are a couple of\nprograms you will need to install linux: something that writes disk\nimages (rawrite.exe or NU or...) and something that can create harddisk\npartitions (fdisk under xenix or older versions of dos, edpart.exe or\nsomething like that).\n\nNOTE! Repartitioning your harddisk will destroy all data on it (well,\nnot exactly, but if you know enough to get back the data you probably\ndidn\u0027t need this warning). So be careful.\n\nREAD THIS THROUGH, THEN READ INSTALL-0.11, AND IF YOU ARE SURE YOU KNOW\nWHAT YOU ARE DOING, CONTINUE. OTHERWISE, PANIC. OR WRITE ME FOR\nEXPLANATIONS. OR DO ANYTHING BUT INSTALL LINUX - IT\u0027S VERY SIMPLE, BUT\nIF YOU DON\u0027T KNOW WHAT YOU ARE DOING YOU\u0027LL PROBABLY BE SORRY. I\u0027D\nRATHER ANSWER A FEW UNNECESSARY MAILS THAN GET MAIL SAYING \"YOU KILLED\nMY HARDDISK, BASTARD. I\u0027M GOING TO FIND YOU, AND YOU\u0027LL BE SORRY WHEN I\nDO\".\n\nMinumum files needed:\n\n\tRELNOTES-0.95 (this file)\n\tINSTALL-0.11 (+ any other docs you might find: the FAQ etc)\n\tbootimage-0.96.Z\n\trootimage-0.95.Z\n\trootimage-0.12.Z (for tar+compress)\n\trawrite.exe\n\tsome disk partitioner\n\n1) back up everything you have on your harddisk - linux-0.95 is still in\n beta and might do weird things. The only thing I guarantee is that\n it has worked fine on /my/ machine - for all I know it might eat your\n harddisk and spit it out in small pieces on any other hardware.\n\n2) Test out the linux boot-disk with the root file system. If it\n doesn\u0027t work, check the hardware requirements, and mail me if you\n still think it should work. I might not be able to help you, but\n your bug-report would still be appreciated.\n\n Linux-0.95 now has an init/login: there should be 4 logins started on\n the first 4 virtual consoles. Log in as root (no password), and test\n it out. Change to the other logins by pressing left-alt + FN[1-4].\n Note that booting up with a floppy as root is S..L..O..W.. - the\n floppy driver has been optimized for sequential access (backups etc),\n and trashes somewhat with demand-loading.\n\n Test that linux can read your harddisk at least partly: run the fdisk\n program on the root-disk, and see if it barfs. If it tells you about\n any partitions at all, linux can successfully read at least part of\n your harddisk.\n\n NOTE! Harddisk device names and numbers have changed between versions\n 0.12 and 0.95: the new numbering system was needed for the extended\n partitions, and a new naming scheme was in order so that people\n wouldn\u0027t cunfuse the old devices with the new ones.\n\n The new harddisk device names are: /dev/hd followed by an \u0027a\u0027 for the\n first drive, or a \u0027b\u0027 for the second one. After that comes the\n partition number, 1-4 for the primary partitions, 5- for possible\n extended partitions. No number means the complete disk. Like this:\n\n\t/dev/hda the whole first harddisk (old: /dev/hd0)\n\t/dev/hdb3 partition nr 3 on the second disk (old: /dev/hd8)\n\n3) Make sure that you have a free /primary/ partition. There can be 4\n primary partitions per drive: newer DOS fdisks seem to be able to\n create only 2 (one primary and one extended). In that case use some\n other partitioning software: edpart.exe etc. Linux fdisk currently\n only tells you the partition info - it doesn\u0027t write to the disk.\n\n Remember to check how big your partition was, as that can be used to\n tell which device Linux thinks it is.\n\n NOTE! Linux-0.95 /might/ recognize extended partitions: but the code\n for this is utterly untested, as I don\u0027t have any of those. Do NOT\n use the extended partitions unless you can verify that they are\n indeed correctly set up - if my routines are wrong, writing to the\n extended partitions might just overwrite some other partition\n instead. Not nice.\n\n4) Boot up linux again, fdisk to make sure you now have the new\n partition, and use mkfs to make a filesystem on one of the partitions\n fdisk reports. Write \"mkfs -c /dev/hdX nnn\" where X is the device\n number reported by linux fdisk, and nnn is the size - also reported\n by fdisk. nnn is the size in /blocks/, ie kilobytes. You should be\n able to use the size info to determine which partition is represented\n by which device name.\n\n5) Mount the new disk partition: \"mount /dev/hdX /mnt\". Copy over the\n root filesystem to the harddisk, eg like this:\n\n\t# for i in bin dev etc usr tmp\n\t# do\n\t# cp +recursive /$i /mnt\n\t# done\n\n You caanot use just \"cp +recursive / /mnt\", as that will result in a\n loop.\n\n6) Sync the filesystem after you have played around enough, and reboot.\n\n\t# sync\n\t# lo\n\n\t(none) login: sync\n\t\u003cwait for it to sync\u003e\n\tctrl-alt-del\n\n THIS IS IMPORTANT! NEVER EVER FORGET TO SYNC BEFORE KILLING THE MACHINE.\n\n7) Change the bootdisk to understand which partition it should use as a\n root filesystem. See INSTALL-0.11: it\u0027s still the word at offset\n 508 into the image. You should be up and running.\n\n8) When you\u0027ve successfully started up with your harddisk as root, you\n can mount the older rootimage (rootimage-0.12) from a floppy, and\n copy over any files you find there that weren\u0027t on the newer\n root-image.\n\n Mounting a floppy is easy: make the directory /floppy, and write:\n\n\t# mount /dev/PS0 /floppy (if you have a 3.5\" drive)\n\n or\n\n\t# mount /dev/at0 /floppy (for 5.25\" floppies)\n\n After that the files can be copied to your harddisk, eg:\n\n\t# cp /floppy/usr/bin/compress /usr/bin\n\t# ln -s /usr/bin/compress /usr/bin/compress\n\t# cp /floppy/usr/bin/tar.Z /usr/bin\n\t# uncompress /usr/bin/tar.Z\n\nThat\u0027s it. Now go back and read the INSTALL-0.11, until you are sure you\nknow what you are doing.\n\n\t\tNew features of 0.95, in order of appearance\n\t\t\t(ie in the order you see them)\n\n\tInit/login\n\nYeah, thanks to poe (Peter Orbaeck (sp?)), linux now boots up like a\nreal unix with a login-prompt. Login as root (no passwd), and change\nyour /etc/passwd to your hearts delight (and add other logins in\n/etc/inittab etc).\n\n\tBash is even bigger\n\nIt\u0027s really a bummer to boot up from floppies: bash takes a long time to\nload. Bash is also now so big that I couldn\u0027t fit compress and tar onto\nthe root-floppy: You\u0027ll probably want the old rootimage-0.12 just in\norder to get tar+compress onto your harddisk. If anybody has pointers\nto a simple shell that is freely distributable, it might be a good idea\nto use that for the root-diskette.\n\nEspecially with a small buffer-cache, things aren\u0027t fun. Don\u0027t worry:\nlinux runs much better on a harddisk.\n\n\tVirtual consoles on any (?) hardware.\n\nYou can select one of several consoles by pressing the left alt-key and\na function key at the same time. Linux should report the number of\nvirtual consoles available upon bootup. /dev/tty0 is now \"the current\"\nscreen, /dev/tty1 is the main console, and /dev/tty2-8 can exist\ndepending on your text-mode or card.\n\nThe virtual consoles also have some new screen-handling commands: they\nconfirm better to vt200 control codes. Special graphic characters, and\nthe PF1-4 keys work somewhat in the application-key mode.\n\n\tSymbolic links.\n\n0.95 now allows symlinks to point to other symlinks etc (the maximum\ndepth is a rather arbitrary 5 links). 0.12 didn\u0027t like more than one\nlevel of indirection.\n\n\tVirtual memory.\n\nVM under 0.95 should be better than under 0.12: no more lockups (as far\nas I have seen), and you can now swap to the filesystem as well as to a\nspecial partition. There are two programs to handle this: mkswap to set\nup a swap-file/partition and swapon to start up swapping.\n\nmkswap needs either a partition or a file that already exists to make a\nswap-area. To make a swap-file, do this:\n\n\t# dd bs\u003d1024 count\u003dNN if\u003d/dev/hda of\u003dswapfile\n\t# mkswap swapfile NN\n\nThe first command just makes a file that is NN blocks long (initializing\nit from /dev/hda, but that could be anything). The second command then\nwrites the necessary setup-info into the file. To start swapping, write\n\n\t# swapon swapfile\n\nNOTE! \u0027dd\u0027 isn\u0027t on the rootdisk: you have to install some things onto\nthe harddisk before you can get up and running.\n\nNOTE2! When linux runs totally out of virtual memory, things slow down\ndramatically. It tries to keep on running as long as it can, but at\nleast it shouldn\u0027t lock up any more. ^C should work, although you might\nhave to wait a while for it..\n\n\tFaster floppies\n\nOk, you don\u0027t notice this much when booting up from a floppy: bash has\ngrown, so it takes longer to load, and the optimizations work mostly\nwith sequential accesses. When you start un-taring floppies to get the\nprograms onto your harddisk, you\u0027ll notice that it\u0027s much faster now.\nThat should be about the only use for floppies under a unix: nobody in\ntheir right mind uses floppies as filesystems.\n\n\tBetter FS-independence\n\nHopefully you\u0027ll never even notice this, but the filesystem has been\npartly rewritten to make it less minix-fs-specific. I haven\u0027t\nimplemented all the VFS-patches I got, so it\u0027s still not ready, but it\u0027s\ngetting there, slowly.\n\n\tAnd that\u0027s it, I think.\n\nHappy hacking.\n\n\t\t\tLinus (torvalds@kruuna.helsinki.fi)\n" }, { "commit": "26d73589f54a0ffa39ca6b16cfe4a46a63f6b781", "tree": "cf5abfad6cf200ada585f1fb37b894f5759fe8ce", "parents": [ "7aadf253807264ad5561479e2a8b9de9602f966d" ], "author": { "name": "Linus Benedict Torvalds", "email": "torvalds@cc.helsinki.fi", "time": "Wed Jan 15 02:13:40 1992 +0200" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:02 2007 -0400" }, "message": "linux-0.12 is available\n\nOk, the subject says it all. The kernel source and the disk-images are\navailable on nic.funet.fi, and I\u0027m assuming they will migrate to other\nplaces in a couple of days. There is a RELNOTES-0.12, but installation\nis similar to 0.11, so the installinfo in relnotes is pretty minimal.\n\nNote that even users of 0.11 should boot up from floppy first and copy\nall the binaries to their proper places on the harddisk partition: fsck,\nls etc have changed with symlinks, and bash (/bin/sh) due to job\ncontrol.\n\nAlso note that I\u0027ve decided to change the copyright to have the same set\nof rules as the GNU copyleft - I got some mail asking about it, and I\nagree. The old copyright still holds for now - I want to make sure none\nof the \"co-authors\" would mind, but I\u0027m pretty sure they won\u0027t. It won\u0027t\nactually change the copyright very much: it\u0027s the handling-costs extra\netc.\n\nI\u0027ll send some additional files (the complete fileutils, the new library\nand libc.a etc) tomorrow.\n\n\t\tLinus\n\n\u003cRELNOTES-0.12\u003e\n\n\t\tRELEASE NOTES FOR LINUX v0.12\n\nThis is file mostly contains info on changed features of Linux, and\nusing old versions as a help-reference might be a good idea.\n\n\t\tCOPYRIGHT\n\nThe Linux copyright will change: I\u0027ve had a couple of requests to make\nit compatible with the GNU copyleft, removing the \"you may not\ndistribute it for money\" condition. I agree. I propose that the\ncopyright be changed so that it confirms to GNU - pending approval of\nthe persons who have helped write code. I assume this is going to be no\nproblem for anybody: If you have grievances (\"I wrote that code assuming\nthe copyright would stay the same\") mail me. Otherwise The GNU copyleft\ntakes effect as of the first of February. If you do not know the gist\nof the GNU copyright - read it.\n\n\t\tINSTALLATION\n\nThis is a SHORT install-note. The installation is very similar to 0.11,\nso read that (INSTALL-0.11) too. There are a couple of programs you will\nneed to install linux: something that writes disk images (rawrite.exe or\nNU or...) and something that can create harddisk partitions (fdisk under\nxenix or older versions of dos, edpart.exe or something like that).\n\nNOTE! Repartitioning your harddisk will destroy all data on it (well,\nnot exactly, but if you know enough to get back the data you probably\ndidn\u0027t need this warning). So be careful.\n\nREAD THIS THROUGH, THEN READ INSTALL-0.11, AND IF YOU ARE SURE YOU KNOW\nWHAT YOU ARE DOING, CONTINUE. OTHERWISE, PANIC. OR WRITE ME FOR\nEXPLANATIONS. OR DO ANYTHING BUT INSTALL LINUX - IT\u0027S VERY SIMPLE, BUT\nIF YOU DON\u0027T KNOW WHAT YOU ARE DOING YOU\u0027LL PROBABLY BE SORRY. I\u0027D\nRATHER ANSWER A FEW UNNECESSARY MAILS THAN GET MAIL SAYING \"YOU KILLED\nMY HARDDISK, BASTARD. I\u0027M GOING TO FIND YOU, AND YOU\u0027LL BE SORRY WHEN I\nDO\".\n\n1) back up everything you have on your harddisk - linux-0.12 is still in\n beta and might do weird things. The only thing I guarantee is that\n it has worked fine on /my/ machine - for all I know it might eat your\n harddisk and spit it out in small pieces on any other hardware.\n\n2) Test out the linux boot-disk with the root file system. If it\n doesn\u0027t work, check the hardware requirements, and mail me if you\n still think it should work. I might not be able to help you, but\n your bug-report would still be appreciated.\n\n Test that linux can read your harddisk at least partly: run the fdisk\n program on the root-disk, and see if it barfs. If it tells you about\n any partitions at all, linux can successfully read at least part of\n your harddisk.\n\n3) Make sure that you have a free /primary/ partition. There can be 4\n primary partitions per drive: newer DOS fdisks seem to be able to\n create only 2 (one primary and one extended). In that case use some\n other partitioning software: edpart.exe etc. Linux fdisk currently\n only tells you the partition info - it doesn\u0027t write to the disk.\n\n Remember to check how big your partition was, as that can be used to\n tell which device Linux thinks it is.\n\n4) Boot up linux again, fdisk to make sure you now have the new\n partition, and use mkfs to make a filesystem on one of the partitions\n fdisk reports. Write \"mkfs -c /dev/hdX nnn\" where X is the device\n number reported by linux fdisk, and nnn is the size - also reported\n by fdisk. nnn is the size in /blocks/, ie kilobytes. You should be\n able to use the size info to determine which partition is represented\n by which device name.\n\n5) Mount the new disk partition: \"mount /dev/hdX /user\". Copy over the\n root filesystem to the harddisk, eg like this:\n\n\t# for i in bin dev etc usr tmp\n\t# do\n\t# cp +recursive /$i /user\n\t# done\n\n You caanot use just \"cp +recursive / /user\", as that will result in a\n loop.\n\n6) Sync the filesystem after you have played around enough, and reboot.\n\n\t# sync\n\t\u003cwait for it to sync\u003e\n\tctrl-alt-del\n\n The folklore says you should do this three times before rebooting:\n once should be enough, but I admit I do it three times anyway :) THIS\n IS IMPORTANT! NEVER EVER FORGET TO SYNC BEFORE KILLING THE MACHINE.\n\n7) Change the bootdisk to understand which partition it should use as a\n root filesystem. See INSTALL-0.11: it\u0027s still the word at offset\n 508 into the image. You should be up and running.\n\nThat\u0027s it. Go back and read the INSTALL-0.11\n\n\t\tNew features of 0.12, in order of appearance\n\t\t\t(ie in the order you see them)\n\n\tLinux now prints cute dots when loading\n\nWoW. Run, don\u0027t walk, to see this :). Seriously, it should hopefully now\nload even on machines that never got off the ground before, but\notherwise the loading hasn\u0027t changed. Implemented by drew.\n\n\tSuper-VGA detection for extended alphamun modes\n\nI cannot guarantee it, I didn\u0027t write it, but it works great on a ET400\nSVGA card. I\u0027m addicted to the new look with 100x40 character editing,\ninstead of a cramped 80x25. This only works on VGA-cards that support\nhigher text-resolutions, and which are correctly identified. Implemented\nby d88-man.\n\n\tJob Control.\n\nOk, everybody used to typing ^Z after they started a long command, and\nforgot to put it in the background - now it works on linux too. Bash\nknows the usualy job-control commands: bg, fg, jobs \u0026 kill. I hope\nthere will be no nasty surprises. Job control was implemented by\ntytso@athena.mit.edu.\n\n\tVirtual consoles on EGA/VGA screens.\n\nYou can select one of several consoles by pressing the left alt-key and\na function key at the same time. Linux should report the number of\nvirtual consoles available upon bootup. /dev/tty0 is now \"the current\"\nscreen, /dev/tty1 is the main console, and /dev/tty2-8 can exist\ndepending on your text-mode or card.\n\nNOTE! Scrolling is noticeably much slower with virtual consoles on a\nEGA/VGA. The reason is that no longer does linux use all the screen\nmemory as a long buffer, but crams in several consoles in it. I think\nit\u0027s worth it.\n\nThe virtual consoles also have some new screen-handling commands: they\nconfirm even better to vt200 control codes than 0.11. Special graphic\ncharacters etc: you can well use them as terminals to VMS (although\nthat\u0027s a shameful waste of resources).\n\n\tpty\u0027s\n\nOk. I have to admit that I didn\u0027t get the hangup-code working correctly,\nbut that should be easy to add. The general things are there.\n\n\tselect\n\nI\u0027ve never used it, so I cannot say how well it works. My minor testing\nseems to indicate that it works ok. vc\u0027s, pty\u0027s and select were\nimplemented by pmacdona, although I hacked it heavily.\n\n\t387-emulation.\n\nIt\u0027s not complete, but it works well enough to run those gcc2.0 compiled\nprograms I tested (few). None of the \"heavy\" math-functions are\nimplemented yet.\n\n\tSymbolic links.\n\nTry out a few \"ln -s xx yy\", and ls -l. Note that I think tar should be\nrecompiled to know anout them, and probably some other programs too. The\n0.12 rootimage-disk has most of the recompiled fileutilities.\n\n\tVirtual memory.\n\nIn addition to the \"mkfs\" program, there is now a \"mkswap\" program on\nthe root disk. The syntax is identical: \"mkswap -c /dev/hdX nnn\", and\nagain: this writes over the partition, so be careful. Swapping can then\nbe enabled by changing the word at offset 506 in the bootimage to the\ndesired device. Use the same program as for setting the root file\nsystem (but change the 508 offset to 506 of course).\n\nNOTE! This has been tested by Robert Blum, who has a 2M machine, and it\nallows you to run gcc without much memory. HOWEVER, I had to stop using\nit, as my diskspace was eaten up by the beta-gcc-2.0, so I\u0027d like to\nhear that it still works: I\u0027ve been totally unable to make a\nswap-partition for even rudimentary testing since about christmastime.\nThus the new changes could possibly just have backfired on the VM, but I\ndoubt it.\n\n\tAnd that\u0027s it, I think.\n\nHappy hacking.\n\n\t\t\tLinus\n" }, { "commit": "7aadf253807264ad5561479e2a8b9de9602f966d", "tree": "d4303c76a830525f6aa6152569b3e3ee83b8c47d", "parents": [ "be068f1a017608faa9b4a0652686426df2e87689" ], "author": { "name": "Linus Benedict Torvalds", "email": "torvalds@cc.helsinki.fi", "time": "Sun Dec 08 17:51:24 1991 +0200" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:01 2007 -0400" }, "message": "linux-0.11 available\n\nThe subject says it all: I have uploaded linux version 0.11 to\nnic.funet.fi and tupac-amaru. They won\u0027t show up for a while on nic,\nbut are already visible in pub/msdos/replace/incoming at amaru. rtx-11\nis so slow to use from here that I didn\u0027t upload them there: I guess\nthey\u0027ll show up within a couple of days (tytso?).\n\nThe files uploaded were:\n\tbootimage.Z - US keyboard compressed bootimage\n\trootimage.Z - 1200kB compressed root image\n\tlinux-0.11.tar.Z - sources\n\tas86.tar.Z - linux binaries for bruce evans\u0027\n\t\t\t\t 16-bit assembler and loader.\n\tINSTALL-0.11 - updated install-info\n\nThis version has a lot of corrections, and is stable at least on my\nmachine. I /hope/ every known bug is fixed, but no promises (and all\nunknown bugs are still there, probably with reinforcements ;-). Those\nwho like to use caps lock as a ctrl-key, you\u0027ll have to re-patch the\nkernel. It\u0027s not implemented in the standard version.\n\n\t\tLinus\n\nPS. I\u0027ll be a bit busy with the #\"$/% physics-course I\u0027m taking, so I\nmight not be as active on the net this coming week as I would like to.\n\nPPS. corsini: the problem you have with the recompiled uemacs sounds\nlike it isn\u0027t resetting the ISIG bit in c_lflags when moving to raw\nmode. Search for c_lflags in the source: it should effectively be set to\nzero (I think) - no signals, no canonical mode etc.\n\n\u003cINSTALL-0.11\u003e\n\n\t\tUsing Linux v0.11\n\t\tLinus Torvalds 08.12.91\n\nNOTE: Users of 0.10, please check the \"changed\" list before using 0.11.\n\n\t\tBooting linux\n\nLinux-0.11 can easily be booted by getting the 2 files bootimage-0.11.Z\nand rootimage-0.11.Z from the linux archive, uncompressing them and\nwriting them out to disks of the same size (ie 2 1.44M floppies or 2\n1.2M floppies). Writing the disks is done with the \"rawrite.exe\" program\nfrom dos, or with \"dd\" from unix. Linux is then booted simply by\ninserting the bootdiskette in drive A, and rebooting the machine. If\neverything goes well, linux will ask you to insert the root-disk after\nloading the system. Hopefully linux will then correctly load the shell\nexecutable, and leave you as root on the new system (prompt \u0027# \u0027).\n\n\t\tUsing it.\n\nYou can get a complete list of available commands by pressing \u003ctab\u003e\ntwice: the root-disk contains mostly setup-programs needed to install\nthe system on a harddisk. You can test them a bit, reading directories\netc.\n\nIn order to install linux on the harddisk, first check out your harddisk\nby executing the command \"fdisk\" - it should show you all the partitions\navailable. If you have only 1 AT-harddisk, you should get a\nerrormessage, just ignore it. At my system fdisk reports the following:\n\n/dev/hd1: 20476 blocks minix\n/dev/hd2: 19975 blocks minix\n/dev/hd3: 1020 blocks minix\n/dev/hd4: 170 blocks active 16-bit DOS (\u003e\u003d32M)\n/dev/hd6: 41641 blocks active minix\n\nThe partition type given (12-bit DOS, minix etc) doesn{t really mean\nanything, unless it\u0027s a \"extended partition\", in which case you\nshouldn\u0027t use that partition for anything: linux doesn\u0027t yet understand\nthem. When later using \"mkfs\" to make a linux file system, it won\u0027t\nchange the output of fdisk, so fdisk may well report \"DOS\", while in\nfact you have made it a linux partition.\n\nIf fdisk doesn\u0027t print out anything but errors, linux is unable to read\nyour harddisk, and you are f**ked. Play around with the floppy version,\nbut you won\u0027t be able to do anything real.\n\n\t\tMaking a filesystem\n\nIn order to really use linux, you will have to make a filesystem on your\nharddisk. This starts by deciding which partition you can use. Look\nagain at what fdisk reports, and try to figure out which of the\npartitions you are using for DOS, OS/2 etc. /dev/hdX where X\u003d{1,2,3,4}\nalways refers to the first harddisk, X\u003d{6,7,8,9} always refers to the\nsecond disk. /dev/hd0 and /dev/hd5 are special: they are all of the\ndrive, and mkfs will refuse to use them for a filesystem.\n\nWhen you are certain you know which device points to which partition,\nyou make a filesystem on the partition of your choice by writing:\n\n\tmkfs -c /dev/hdX blocks\n\nwhere \"-c\" means that you want mkfs to check for errors, \"dev/hdX\" is\nthe free partition you intend to use for linux, and \"blocks\" is the\nnumber of blocks fdisk reports for that particular partition. NOTE! mkfs\nwill overwrite the partition you selected, so be doubly (or triply) sure\nthat you don\u0027t mind that.\n\nNote that when using the \"-c\" flag, mkfs will read through the entire\npartition: this can take some time. If there are read errors, mkfs will\nmark the particular block as bad, and continue: linux will also print a\nlittle message \"harddisk I/O error\". After running mkfs these messages\nshould never occur again: if they do, your data may be corrupted.\n\n\t\tMounting the filesystem\n\nAfter mkfs has exited, it\u0027s time to mount the file-system, and do the\nnecessary things to make it a root file system. Mount the new filesystem\non /user by writing:\n\n\tcd /\n\tmount /dev/hdX /user\n\nIf you get errors for this, mkfs failed, and there is probably something\nseriously wrong.\n\nAfter mounting the device, you want to move all the files on the current\nfloppy-root to the new fs. This can most easily be done by writing:\n\n\tcd /user\n\tfor i in bin dev etc usr tmp floppy\n\tdo\n\tcp +recursive +verbose /$i $i\n\tdone\n\tsync\n\nwhich will also tell you what it is doing (/bin/sh -\u003e bin/sh etc).\n\nAfter that, you should have a new filesystem that contains the bare\nnecessities to start hacking linux. Play around some more, and exit\nlinux by writing \"logout or exit\". This should result in\n\n\tchild 4 died with error code 0000\n\t#\n\nDo a couple of syncs (3 is a magic number), and reboot the machine.\nALWAYS remember to sync before rebooting: terrible things happen if you\ndon\u0027t.\n\n\t\tUsing the harddisk as root\n\nOnce you have happily made a new root, you will want to boot up with it.\nThis is done by changing a word at offset 508 in the boot-image. The\nword (in 386-order, ie low byte first) tells the system which device to\nuse as root: it is initially 0, which means that we want to use a floppy\nof the same type as the boot-disk (and this is the reason that you may\nnot use a 360kB boot-disk even though the system fits on one: it has to\nbe the same type as the root-diskette).\n\nIn order to use the harddisk as root, this value has to be changed to\npoint to the correct device. Harddisks have a major number of 3 under\nlinux, and the minor nr is the same as the number X in /dev/hdX. The\ncomplete device number is then calculated with\n\n\tDEV_NO \u003d (major\u003c\u003c8)+minor\n\nor alternatively major*256+minor. Thus /dev/hd1 is (3\u003c\u003c8)+1 \u003d 0x301,\n/dev/hd6 \u003d 0x0306 etc. Assuming the partition you made into the new root\nwas /dev/hd2, you will have to write 0x0302 into the boot-image. That\nis, you should change the 508th byte in the image to 0x02, and the 509th\nbyte to 0x03. There is a sample program for this in some of the older\nINSTALL-notes, if you don\u0027t understand what it\u0027s all about.\n\n\t\tOk, I got the root on hd, what now?\n\nAs you have probably noticed, you cannot get very far with the binaries\nfound on the original root-diskette. So the first thing you want to do\nis to import some new binaries. To do this you need to tell linux what\nkind of floppies you have, as that\u0027s the easiest way to import things.\n\nAs with harddisk, floppies have device numbers, but this time major \u003d 2\ninstead of 3. The minor number is not as easy: it\u0027s a composite that\ntells which drive (A, B, C or D) and what type of drive (360kB, 1.2M,\n1.44M etc). The formula is \u0027minor \u003d type*4+nr\u0027, where nr is 0-3 for A-D,\nand type is 2 for 1.2M disks, and 7 for 1.44M disks. There are other\ntypes, but these should suffice for now.\n\nThus if you have a 1.2M A-drive, and want to call it \"floppy0\", you have\nto tell linux so. This is done with the \"mknod\" command. mknod takes 4\nparamters: the unix name of the device, a \"b\" or a \"c\" depending on\nwhether it\u0027s a Block of Character device, and the major and minor\nnumbers. Thus to make \"floppy0\" a 1.2M A-drive, you write:\n\n\tmknod /dev/floppy0 b 2 8\n\nb is for Block-device, the 2 is for floppy, and the 8 is 4*2+0, where\nthe 2 is 1.2M-drive and the 0 is drive A. Likewise to make a \"floppy1\"\ndevice that is a 1.44M drive in B, you write:\n\n\tmknod /dev/floppy1 b 2 29\n\nwhere 29 \u003d 4*7 + 1. There are a couple of standard names, for users\nthat are used to minix (major, minor in parentheses): /dev/PS0 is a\n1.44M in A (2,28), /dev/PS1 a 1.44M in B (2,29), /dev/at0 is a 1.2M in A\n(2,8), /dev/at1 is a 1.2M in B (2,9). Use mknod to make those that fit\nyour computer.\n\nAfter you have made these special block devices, you can now read a\nfloppy under linux. The easiest way to import things into linux is by\nwriting a tar-file to a floppy with rawrite.exe, and then using:\n\n\ttar xvf /dev/floppy0\n\nto untar it under linux. This way you can get the gcc binaries etc\navailable from the linux-carrying sites.\n\n\t\tChanges from 0.10:\n\n- /bin/update is no longer automatically executed upon bootup: instead\nthe file /etc/rc is evaluated by the shell. This file can then start the\nupdate process, mount andy needed filesystems, possibly fsck\u0027ing them\nfirst. A minimal /etc/rc looks like this:\n\n\t/bin/update \u0026\n\t\u003e /etc/mtab\n\techo \" Ok.\"\n\n- init() restarts the shell every time it is exited: logout from the\nlogin shell results in a \"child xxx died with error code yyy\", a sync\nand then a new shell as root.\n\n- floppies work a lot better than in 0.10. Even using two floppies at\nthe same time seems to work out ok. Reading big chunks at a time is also\nfaster then in 0.10 (I think).\n\n- harddisk errors are handled better. Use the \"-c\" option in mkfs to map\nout all errors.\n\n- linux accepts most video-cards: harcules, MDA, CGA etc seem to work.\n\n- ^G beeps on the console, so command completion under bash etc will\nnotify of errors.\n\n- sticky directories, corrected handling of uid/gid bits, and better\nhandling of protections when not root. Most of these won\u0027t be noticeable\nuntil we get a init/login.\n" }, { "commit": "be068f1a017608faa9b4a0652686426df2e87689", "tree": "fff959a22d75c4a82192c99f1b4caa26a236eafc", "parents": [ "8331b82c1ac6ebe96db75d5853b4329b1502d5d3" ], "author": { "name": "Linus Benedict Torvalds", "email": "torvalds@klaava.Helsinki.FI", "time": "Mon Nov 04 00:00:00 1991 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:01 2007 -0400" }, "message": "Linux-0.10\n\n[ Committer\u0027s note: official announcement and exact release date unknown.\n Some posts are included below to narrow the possible release date.\n There is a possibility that this version might not be pristine.\n Also this commit folds the 0.02 version which is apparently lost. ]\n\n\u003cINSTALL-0.10\u003e\n\nWarning: I have personally not done this the hard way, so I don\u0027t know\nwhat problems could surface. In general, this version is still meant\nfor people with minix: they are more used to the system, and can do some\nthings that DOS-based persons cannot. If you have only DOS, expect some\ntroubles. As the version number suggests, this is still not the final\nproduct.\n\nThis is a \"fast hack\", meant as a minimal guide to what you must do.\nI\u0027ll expand this as soon as people tell me what they have problems with\netc etc. If somebody who has successfully installed the system wants to\nwrite something better, I\u0027d be delighted. This guide stinks to high\nheaven.\n\n\t\tInstalling Linux-0.10 on your system\n\nThere are 5 major steps in installing linux on your system:\n\n1 - BACK UP ANY IMPORTANT DATA. Linux accesses your hardware directly,\nand if your hardware differs from mine, you could be in for a nasty\nsurprise. Doublecheck that your hardware is compatible: AT style\nharddisk, VGA controller. (If somebody has EGA, please tell me if the\nscreen driver should happen to work)\n\n2 - Make a file-system on your harddisk. This is easy if you have\nminix, but if you haven\u0027t got minix, you\u0027ll have to get the minix\ndemo-disk from somewhere (plains.nodak.edu is one place), and use that.\nThere should be a manual accompanying the demo-disk, and you had better\nread that carefully. Although this version of linux will boot up\nwithout minix, a knowledge of minix would help. Especially if you have\nnever done any unix work, you\u0027ll be very confused.\n\nMaking a filesystem means getting a empty partition (with DOS fdisk or\nsimilar), and using the \u0027mkfs /dev/hdX nnn\u0027 command to write out a empty\nfile-system.\n\n3 - copy the diskimages to two floppies. Again, under minix (or any\nunix), this is easy, as you can just do a simple \u0027dd\u0027 to a floppy, but\nfrom within MS-DOS this might be a bit trickier. \u0027debug\u0027 should be able\nto write diskettes directly, or you could get the sources to \"raw-write\"\nfrom the same place as you got the minix demo disk, and modify them to\nwrite out any disk image (or do they do that already?).\n\nNOTE! The floppies MUST be of the same type: even though the boot-image\nwill fit nicely on a 360kB floppy, you have to write it to the same type\nof floppy as the root-image. That means a 1.2M or 1.44M floppy. The\nreason is that the floppy-type is determined at boot-time from the\nboot-floppy. Thus the same binary works on both 3.5\" and 5.25\" drives.\n\n4 - boot up from floppy. This should be obvious. Having a floppy as\nroot-device isn\u0027t very fast (especially on a machine with less than 6MB\ntotal ram -\u003e small buffer cache), but it works (I hope). Test the\nprograms on the root-floppy (cat mkdir etc).\n\n5 - Mount the harddisk partition (I do it on /user: ie\n\u0027mount /dev/hdX /user\u0027), and copy the file system over to the new\npartition. The following is a example of how to do this:\n\n$ cd /user\n$ mkdir usr\n$ for i in bin etc usr/bin usr/root mtools\n\u003e do\n\u003e mkdir $i\n\u003e cp `ls -A /$i` $i\n\u003e done\n$ mkdir dev\n$ cd dev\n$ for i in 0 1 2 3 4 5 6 7 8 9\n\u003e do\n\u003e mknod \u0027hd\u0027$i b 3 $i\n\u003e done\n$ mknod tty c 5 0\n$ mknod tty0 c 4 0\n$ mknod tty1 c 4 1\n$ mknod tty2 c 4 2\n\nYou should now have a filesystem you could boot from. Play around a bit,\ntry to get aquainted with the new system. Log out when you\u0027ve had\nenough.\n\n6 - Changing the boot-diskette use your new harddisk partition as root.\nThe root device to be used for linux is encoded in a word at offset 508\nin the boot image. Normally this is 0, meaning that the root is to be\nthe same type of floppy as was used in the boot process. This can be\nchanged to whatever you like.\n\nUse a short program like the one at the end to change the word (I assume\neverybody has access to some kind of C compiler, be it under dos or\nunix). You can then write out the new bootdisk, and boot from it, now\nusing the harddisk as root (much faster). Once you have successfully\ndone that you might want to install additional programs (gcc etc) by\nreading them from a dos-floppy with \u0027mcopy\u0027.\n\n\t\tLinus (torvalds@kruuna.helsinki.fi)\n\n------\texample program: use \u0027a.out \u003c oldboot \u003e newboot\u0027 ----\n#include \u003cunistd.h\u003e\nchar tmp[512];\n\nvoid main(void)\n{\n\tint i;\n\n\tif (512 !\u003d read(0,tmp,512))\n\t\texit(1);\n\tif (0xAA55 !\u003d *((unsigned short *)(tmp+510)))\n\t\texit(2);\n\t*((unsigned short *)(tmp+508)) \u003d NEW_DEV;\n\tif (512 !\u003d write(1,tmp,512))\n\t\texit(3);\n\twhile ((i\u003dread(0,tmp,512)) \u003e 0)\n\t\tif (i !\u003d write(1,tmp,i))\n\t\t\texit(4);\n\texit(0);\n}\n-------\n\n\t\tDevices:\n\nHarddisks:\n0x301 - /dev/hd1 - first partition on first drive\n...\n0x304 - /dev/hd2 - fourth partition on first drive\n\n0x306 - /dev/hd1 - first partition on second drive\n...\n0x309 - /dev/hd2 - fourth partition on second drive\n\n0x300 - /dev/hd0 - the whole first drive. BE CAREFUL\n0x305 - /dev/hd5 - the whole second drive. BE CAREFUL\n\nFloppies:\n0x208 - 1.2M in A\n0x209 - 1.2M in B\n0x21C - 1.44M in A\n0x21D - 1.44M in B\n\n-----\n\nFrom: torvalds@klaava.Helsinki.FI (Linus Benedict Torvalds)\nNewsgroups: comp.os.minix\nSubject: Free minix-like kernel sources for 386-AT\nKeywords: 386, preliminary version\nMessage-ID: \u003c1991Oct5.054106.4647@klaava.Helsinki.FI\u003e\nDate: 5 Oct 91 05:41:06 GMT\nOrganization: University of Helsinki\n\nDo you pine for the nice days of minix-1.1, when men were men and wrote\ntheir own device drivers? Are you without a nice project and just dying\nto cut your teeth on a OS you can try to modify for your needs? Are you\nfinding it frustrating when everything works on minix? No more all-\nnighters to get a nifty program working? Then this post might be just\nfor you :-)\n\nAs I mentioned a month(?) ago, I\u0027m working on a free version of a\nminix-lookalike for AT-386 computers. It has finally reached the stage\nwhere it\u0027s even usable (though may not be depending on what you want),\nand I am willing to put out the sources for wider distribution. It is\njust version 0.02 (+1 (very small) patch already), but I\u0027ve successfully\nrun bash/gcc/gnu-make/gnu-sed/compress etc under it.\n\nSources for this pet project of mine can be found at nic.funet.fi\n(128.214.6.100) in the directory /pub/OS/Linux. The directory also\ncontains some README-file and a couple of binaries to work under linux\n(bash, update and gcc, what more can you ask for :-). Full kernel\nsource is provided, as no minix code has been used. Library sources are\nonly partially free, so that cannot be distributed currently. The\nsystem is able to compile \"as-is\" and has been known to work. Heh.\nSources to the binaries (bash and gcc) can be found at the same place in\n/pub/gnu.\n\nALERT! WARNING! NOTE! These sources still need minix-386 to be compiled\n(and gcc-1.40, possibly 1.37.1, haven\u0027t tested), and you need minix to\nset it up if you want to run it, so it is not yet a standalone system\nfor those of you without minix. I\u0027m working on it. You also need to be\nsomething of a hacker to set it up (?), so for those hoping for an\nalternative to minix-386, please ignore me. It is currently meant for\nhackers interested in operating systems and 386\u0027s with access to minix.\n\nThe system needs an AT-compatible harddisk (IDE is fine) and EGA/VGA. If\nyou are still interested, please ftp the README/RELNOTES, and/or mail me\nfor additional info.\n\nI can (well, almost) hear you asking yourselves \"why?\". Hurd will be\nout in a year (or two, or next month, who knows), and I\u0027ve already got\nminix. This is a program for hackers by a hacker. I\u0027ve enjouyed doing\nit, and somebody might enjoy looking at it and even modifying it for\ntheir own needs. It is still small enough to understand, use and\nmodify, and I\u0027m looking forward to any comments you might have.\n\nI\u0027m also interested in hearing from anybody who has written any of the\nutilities/library functions for minix. If your efforts are freely\ndistributable (under copyright or even public domain), I\u0027d like to hear\nfrom you, so I can add them to the system. I\u0027m using Earl Chews estdio\nright now (thanks for a nice and working system Earl), and similar works\nwill be very wellcome. Your (C)\u0027s will of course be left intact. Drop me\na line if you are willing to let me use your code.\n\n\t\tLinus\n\n-----\n\nFrom: torvalds@klaava.Helsinki.FI (Linus Benedict Torvalds)\nNewsgroups: comp.os.minix\nSubject: Re: [comp.os.minix] Free minix-like kernel sources for 386-AT\nSummary: Still only ftp\nMessage-ID: \u003c1991Oct31.101252.13981@klaava.Helsinki.FI\u003e\nDate: 31 Oct 91 10:12:52 GMT\nReferences: \u003c3255@cluster.cs.su.oz.au\u003e \u003cAA7ig3fuR5@inzer.demos.su\u003e\nOrganization: University of Helsinki\n\nIn article \u003cAA7ig3fuR5@inzer.demos.su\u003e moroz@inzer.demos.su writes:\n\u003echrisa@extro.ucc.su.OZ.AU (C. G. Albone) writes:\n\u003e\u003eHello all..\n\u003e\u003e\tI missed the original posting, so could someone please tell me\n\u003e\u003ehow I may obtain the sources.\n\u003e And me too !!!!!!!!!!!!!!\n\u003e Also, are them available via e-mail (mail server or smth), as I\u0027m\n\u003ein the former Soviet Union and can\u0027t do any FTP.\n\nOk, as I\u0027ve gotten quite a few questions, I guess I\u0027d better follow up\nagain.\n\nLinux is currently ONLY available via ftp from nic.funet.fi, directory\n/pub/OS/Linux. As the sources change rather rapidly (next release due\nout this weekend after I have tested some more), it is also currently\nimpractical to make them available from other places. There is a\nmail-server possibility fron nic, but I think it\u0027s still in testing (you\ncould try mailing \"mailserver@nic.funet.fi\" with \"help\" in the body,\nbut I don\u0027t know if it will work).\n\nLinux is a full kernel that has so far worked on a number (5-10?) of\nat-386 (and one 486 as far as I know). It supports GNU cc (gcc), bash\nand some other free stuff. It is currently more of a hackers kernel (and\nminix-386 is needed, but that will change with this weeks release), and\nthe current version number is 0.03 (next is 0.10 I think).\n\nGood things about linux:\n\n- it\u0027s free, full source, and I try to correct bugs you find.\n- it\u0027s a bit faster than minix, I think.\n- uses paging for memory management (not to disk yet)\n- multithreaded fs (but then you can get patches to minix that do\nsimilar stuff)\n- mostly full termios and vt100-console.\n- most things easy to port (easier than to minix).\n\nBad points:\n\n- ONLY 386/486\n- early versions: there might be lots of bugs, and you might need to\nport/hack things to work.\n- minix is recommended even for the upcoming version that doesn\u0027t\nabsolutely need it.\n- currently only VGA (EGA?) support, limited keyboard drivers (US and\nFinnish) etc\n\nYou can mail me for more info. \"finger torvalds@kruuna.helsinki.fi\"\nmight tell you something too.\n\n\t\tLinus (torvalds@kruuna.helsinki.fi)\n\n-----\n\nDate: Thu, 7 Nov 91 14:19:23 -0500\nTo: linux-activists-mtg@tsx-11.MIT.EDU\nFrom: Robert Lund \u003crml@bighorn.uswest.com\u003e\nReply-To: tytso@Athena.MIT.EDU\n\nHello,\n\nI have installed Linux 0.10 using the minix demo to build the root\nfile system on a hd partition. I couldn\u0027t figure out how to use the\nmtools commands to get gccbin, utils, etc. from DOS to linux but I did\ndiscover an alternate approach that might prove useful to others.\n\nI happen to have a 1.44 drive as my A drive so ubder linux I did\n\n\tmknod /dev/PS0 B 2 28\n\nNext, I formated a 1.44 floppy under DOS.\n\nThen, I uncompressed the tar files that I wanted to get from DOS to linux\n(I actually uncompressed under UNIX but I assume that a 16 bit uncompress\nutility under DOS would work).\n\nThen, I used the rawrite command available with the minix demo to dump a tar\nfile, e.g. gccbin.tar, to the 1.44 floppy in the A drive (back under DOS\nagain)\n\nNext, I booted Linux and did\n\n\ttar -xvf /dev/PS0\n\nand lo and behold, it worked; tar read the raw device and\nsuccessfully extracted the files.\n\nHope this helps someone.\n\nBob Lund\n" }, { "commit": "8331b82c1ac6ebe96db75d5853b4329b1502d5d3", "tree": "7ae3e06d404805b6e7867894ac3a8355c559fc1a", "parents": [], "author": { "name": "Linus Benedict Torvalds", "email": "torvalds@klaava.Helsinki.FI", "time": "Tue Sep 17 17:29:55 1991 +0000" }, "committer": { "name": "Nicolas Pitre", "email": "nico@cam.org", "time": "Sun Aug 19 14:19:01 2007 -0400" }, "message": "linux release 0.01\n\n\t\t1. Short intro\n\nThis is a free minix-like kernel for i386(+) based AT-machines. Full\nsource is included, and this source has been used to produce a running\nkernel on two different machines. Currently there are no kernel\nbinaries for public viewing, as they have to be recompiled for different\nmachines. You need to compile it with gcc (I use 1.40, don\u0027t know if\n1.37.1 will handle all __asm__-directives), after having changed the\nrelevant configuration file(s).\n\nAs the version number (0.01) suggests this is not a mature product.\nCurrently only a subset of AT-hardware is supported (hard-disk, screen,\nkeyboard and serial lines), and some of the system calls are not yet\nfully implemented (notably mount/umount aren\u0027t even implemented). See\ncomments or readme\u0027s in the code.\n\nThis version is also meant mostly for reading - ie if you are interested\nin how the system looks like currently. It will compile and produce a\nworking kernel, and though I will help in any way I can to get it\nworking on your machine (mail me), it isn\u0027t really supported. Changes\nare frequent, and the first \"production\" version will probably differ\nwildly from this pre-alpha-release.\n\nHardware needed for running linux:\n\t- 386 AT\n\t- VGA/EGA screen\n\t- AT-type harddisk controller (IDE is fine)\n\t- Finnish keyboard (oh, you can use a US keyboard, but not\n\t without some practise :-)\n\nThe Finnish keyboard is hard-wired, and as I don\u0027t have a US one I\ncannot change it without major problems. See kernel/keyboard.s for\ndetails. If anybody is willing to make an even partial port, I\u0027d be\ngrateful. Shouldn\u0027t be too hard, as it\u0027s tabledriven (it\u0027s assembler\nthough, so ...)\n\nAlthough linux is a complete kernel, and uses no code from minix or\nother sources, almost none of the support routines have yet been coded.\nThus you currently need minix to bootstrap the system. It might be\npossible to use the free minix demo-disk to make a filesystem and run\nlinux without having minix, but I don\u0027t know...\n\n\t\t2. Copyrights etc\n\nThis kernel is (C) 1991 Linus Torvalds, but all or part of it may be\nredistributed provided you do the following:\n\n\t- Full source must be available (and free), if not with the\n\t distribution then at least on asking for it.\n\n\t- Copyright notices must be intact. (In fact, if you distribute\n\t only parts of it you may have to add copyrights, as there aren\u0027t\n\t (C)\u0027s in all files.) Small partial excerpts may be copied\n\t without bothering with copyrights.\n\n\t- You may not distibute this for a fee, not even \"handling\"\n\t costs.\n\nMail me at \"torvalds@kruuna.helsinki.fi\" if you have any questions.\n\nSadly, a kernel by itself gets you nowhere. To get a working system you\nneed a shell, compilers, a library etc. These are separate parts and may\nbe under a stricter (or even looser) copyright. Most of the tools used\nwith linux are GNU software and are under the GNU copyleft. These tools\naren\u0027t in the distribution - ask me (or GNU) for more info.\n\n\t\t3. Short technical overview of the kernel.\n\nThe linux kernel has been made under minix, and it was my original idea\nto make it binary compatible with minix. That was dropped, as the\ndifferences got bigger, but the system still resembles minix a great\ndeal. Some of the key points are:\n\n\t- Efficient use of the possibilities offered by the 386 chip.\n\t Minix was written on a 8088, and later ported to other\n\t machines - linux takes full advantage of the 386 (which is\n\t nice if you /have/ a 386, but makes porting very difficult)\n\n\t- No message passing, this is a more traditional approach to\n\t unix. System calls are just that - calls. This might or might\n\t not be faster, but it does mean we can dispense with some of\n\t the problems with messages (message queues etc). Of course, we\n\t also miss the nice features :-p.\n\n\t- Multithreaded FS - a direct consequence of not using messages.\n\t This makes the filesystem a bit (a lot) more complicated, but\n\t much nicer. Coupled with a better scheduler, this means that\n\t you can actually run several processes concurrently without\n\t the performance hit induced by minix.\n\n\t- Minimal task switching. This too is a consequence of not using\n\t messages. We task switch only when we really want to switch\n\t tasks - unlike minix which task-switches whatever you do. This\n\t means we can more easily implement 387 support (indeed this is\n\t already mostly implemented)\n\n\t- Interrupts aren\u0027t hidden. Some people (among them Tanenbaum)\n\t think interrupts are ugly and should be hidden. Not so IMHO.\n\t Due to practical reasons interrupts must be mainly handled by\n\t machine code, which is a pity, but they are a part of the code\n\t like everything else. Especially device drivers are mostly\n\t interrupt routines - see kernel/hd.c etc.\n\n\t- There is no distinction between kernel/fs/mm, and they are all\n\t linked into the same heap of code. This has it\u0027s good sides as\n\t well as bad. The code isn\u0027t as modular as the minix code, but\n\t on the other hand some things are simpler. The different parts\n\t of the kernel are under different sub-directories in the\n\t source tree, but when running everything happens in the same\n\t data/code space.\n\nThe guiding line when implementing linux was: get it working fast. I\nwanted the kernel simple, yet powerful enough to run most unix software.\nThe file system I couldn\u0027t do much about - it needed to be minix\ncompatible for practical reasons, and the minix filesystem was simple\nenough as it was. The kernel and mm could be simplified, though:\n\n\t- Just one data structure for tasks. \"Real\" unices have task\n\t information in several places, I wanted everything in one\n\t place.\n\n\t- A very simple memory management algorithm, using both the\n\t paging and segmentation capabilities of the i386. Currently\n\t MM is just two files - memory.c and page.s, just a couple of\n\t hundreds of lines of code.\n\nThese decisions seem to have worked out well - bugs were easy to spot,\nand things work.\n\n\t\t4. The \"kernel proper\"\n\nAll the routines handling tasks are in the subdirectory \"kernel\". These\ninclude things like \u0027fork\u0027 and \u0027exit\u0027 as well as scheduling and minor\nsystem calls like \u0027getpid\u0027 etc. Here are also the handlers for most\nexceptions and traps (not page faults, they are in mm), and all\nlow-level device drivers (get_hd_block, tty_write etc). Currently all\nfaults lead to a exit with error code 11 (Segmentation fault), and the\nsystem seems to be relatively stable (\"crashme\" hasn\u0027t - yet).\n\n\t\t5. Memory management\n\nThis is the simplest of all parts, and should need only little changes.\nIt contains entry-points for some things that the rest of the kernel\nneeds, but mostly copes on it\u0027s own, handling page faults as they\nhappen. Indeed, the rest of the kernel usually doesn\u0027t actively allocate\npages, and just writes into user space, letting mm handle any possible\n\u0027page-not-present\u0027 errors.\n\nMemory is dealt with in two completely different ways - by paging and\nsegmentation. First the 386 VM-space (4GB) is divided into a number of\nsegments (currently 64 segments of 64Mb each), the first of which is the\nkernel memory segment, with the complete physical memory identity-mapped\ninto it. All kernel functions live within this area.\n\nTasks are then given one segment each, to use as they wish. The paging\nmechanism sees to filling the segment with the appropriate pages,\nkeeping track of any duplicate copies (created at a \u0027fork\u0027), and making\ncopies on any write. The rest of the system doesn\u0027t need to know about\nall this.\n\n\t\t6. The file system\n\nAs already mentioned, the linux FS is the same as in minix. This makes\ncrosscompiling from minix easy, and means you can mount a linux\npartition from minix (or the other way around as soon as I implement\nmount :-). This is only on the logical level though - the actual\nroutines are very different.\n\n\tNOTE! Minix-1.6.16 seems to have a new FS, with minor\n\tmodifications to the 1.5.10 I\u0027ve been using. Linux\n\twon\u0027t understand the new system.\n\nThe main difference is in the fact that minix has a single-threaded\nfile-system and linux hasn\u0027t. Implementing a single-threaded FS is much\neasier as you don\u0027t need to worry about other processes allocating\nbuffer blocks etc while you do something else. It also means that you\nlose some of the multiprocessing so important to unix.\n\nThere are a number of problems (deadlocks/raceconditions) that the linux\nkernel needed to address due to multi-threading. One way to inhibit\nrace-conditions is to lock everything you need, but as this can lead to\nunnecessary blocking I decided never to lock any data structures (unless\nactually reading or writing to a physical device). This has the nice\nproperty that dead-locks cannot happen.\n\nSadly it has the not so nice property that race-conditions can happen\nalmost everywhere. These are handled by double-checking allocations etc\n(see fs/buffer.c and fs/inode.c). Not letting the kernel schedule a\ntask while it is in supervisor mode (standard unix practise), means that\nall kernel/fs/mm actions are atomic (not counting interrupts, and we are\ncareful when writing those) if you don\u0027t call \u0027sleep\u0027, so that is one of\nthe things we can count on.\n\n\t\t7. Apologies :-)\n\nThis isn\u0027t yet the \"mother of all operating systems\", and anyone who\nhoped for that will have to wait for the first real release (1.0), and\neven then you might not want to change from minix. This is a source\nrelease for those that are interested in seeing what linux looks like,\nand it\u0027s not really supported yet. Anyone with questions or suggestions\n(even bug-reports if you decide to get it working on your system) is\nencouraged to mail me.\n\n\t\t8. Getting it working\n\nMost hardware dependancies will have to be compiled into the system, and\nthere a number of defines in the file \"include/linux/config.h\" that you\nhave to change to get a personalized kernel. Also you must uncomment\nthe right \"equ\" in the file boot/boot.s, telling the bootup-routine what\nkind of device your A-floppy is. After that a simple \"make\" should make\nthe file \"Image\", which you can copy to a floppy (cp Image /dev/PS0 is\nwhat I use with a 1.44Mb floppy). That\u0027s it.\n\nWithout any programs to run, though, the kernel cannot do anything. You\nshould find binaries for \u0027update\u0027 and \u0027bash\u0027 at the same place you found\nthis, which will have to be put into the \u0027/bin\u0027 directory on the\nspecified root-device (specified in config.h). Bash must be found under\nthe name \u0027/bin/sh\u0027, as that\u0027s what the kernel currently executes. Happy\nhacking.\n\n\t\tLinus Torvalds\t\t\"torvalds@kruuna.helsinki.fi\"\n\t\tPetersgatan 2 A 2\n\t\t00140 Helsingfors 14\n\t\tFINLAND\n\n-----\n\nFrom: torvalds@klaava.Helsinki.FI (Linus Benedict Torvalds)\nNewsgroups: comp.os.minix\nSubject: What would you like to see most in minix?\nSummary: small poll for my new operating system\nKeywords: 386, preferences\nMessage-ID: \u003c1991Aug25.205708.9541@klaava.Helsinki.FI\u003e\nDate: 25 Aug 91 20:57:08 GMT\nOrganization: University of Helsinki\n\nHello everybody out there using minix -\n\nI\u0027m doing a (free) operating system (just a hobby, won\u0027t be big and\nprofessional like gnu) for 386(486) AT clones. This has been brewing\nsince april, and is starting to get ready. I\u0027d like any feedback on\nthings people like/dislike in minix, as my OS resembles it somewhat\n(same physical layout of the file-system (due to practical reasons)\namong other things).\n\nI\u0027ve currently ported bash(1.08) and gcc(1.40), and things seem to work.\nThis implies that I\u0027ll get something practical within a few months, and\nI\u0027d like to know what features most people would want. Any suggestions\nare welcome, but I won\u0027t promise I\u0027ll implement them :-)\n\n\t\tLinus (torvalds@kruuna.helsinki.fi)\n\nPS. Yes - it\u0027s free of any minix code, and it has a multi-threaded fs.\nIt is NOT protable (uses 386 task switching etc), and it probably never\nwill support anything other than AT-harddisks, as that\u0027s all I have :-(.\n" } ] }