Path: gmdzi!unido!mcsun!uunet!samsung!xylogics!transfer!lectroid!lennox
From: len...@lectroid.sw.stratus.com (Craig Scott Lennox)
Newsgroups: comp.os.minix
Subject: another naive suggestion
Message-ID: <1581@lectroid.sw.stratus.com>
Date: 20 Jun 90 18:41:44 GMT
Sender: use...@lectroid.sw.stratus.com
Lines: 20
Posted: Wed Jun 20 19:41:44 1990
Well my last suggestion (about overcoming the 64K I&D barrier through
changes in cc) sure got me a deluge of mail. The upshot of it is, that
this is a serious problem that shouldn't be just kludged around at the
application level, and if that's what I want to do, "go back to playing with
MS-DOS!" :-) A native 32-bit '386 MINIX kernel really is the way
to go here.
I know that several of you have running '386 kernels, but of course the
P-H copyright forbids you from posting or mailing them to others. However,
I'm wondering if it would be legally kosher to puchase PC-MINIX and mail
the distribution diskettes to one of you with a working '386-protected
mode MINIX, in exchange for which you would send me your kernel and source.
Thus P-H would not be stiffed a license fee.
--
| flame me at: len...@shire.hw.stratus.com, Craig Lennox, Stratus Computer |
|"Oh boy, virtual memory! Now I'm gonna make myself a REALLY BIG ram disk!" |
| Disclaimer: My opinions are covered by section 2b of the Gnu Public |
| License and thus do not belong to Stratus Computer. |
Path: gmdzi!unido!mcsun!sunic!uupsi!rpi!dali.cs.montana.edu!samsung!know!
cs.utexas.edu!usc!snorkelwacker!mit-eddie!uw-beaver!ssc-vax!bcsaic!paula
From: pa...@bcsaic.UUCP (Paul Allen)
Newsgroups: comp.os.minix
Subject: Re: another naive suggestion
Message-ID: <26888@bcsaic.UUCP>
Date: 22 Jun 90 01:08:43 GMT
References: <1581@lectroid.sw.stratus.com>
Organization: Boeing Computer Services AI Center, Seattle
Lines: 27
Posted: Fri Jun 22 02:08:43 1990
In article <1...@lectroid.sw.stratus.com> len...@lectroid.sw.stratus.com
(Craig Scott Lennox) writes:
>
>[...] A native 32-bit '386 MINIX kernel really is the way
>to go here.
>
>I know that several of you have running '386 kernels, but of course the
>P-H copyright forbids you from posting or mailing them to others. However,
>I'm wondering if it would be legally kosher to puchase PC-MINIX and mail
>the distribution diskettes to one of you with a working '386-protected
>mode MINIX, in exchange for which you would send me your kernel and source.
>Thus P-H would not be stiffed a license fee.
Take a look at section 9.10 in the monthly Minix Information Sheet, and
then ftp to hobbes.cs.umd.edu. I just checked, and the latest (6/20/90)
version of Bruce's 32-bit kit is now there. Note that this is not for
the faint of heart. I'm not done building it yet, so I think I'll reserve
further comment until I've got the whole thing nailed down.
I believe James da Silva is responsible for the little Minix archive
on hobbes. Thanks, James!
Paul Allen
--
------------------------------------------------------------------------
Paul L. Allen | pal...@atc.boeing.com
Boeing Advanced Technology Center | ...!uw-beaver!bcsaic!pallen
Path: gmdzi!unido!mcsun!uunet!nems!mimsy!jds
From: j...@mimsy.umd.edu (James da Silva)
Newsgroups: comp.os.minix
Subject: Minix-386 availability (was Re: another naive suggestion)
Message-ID: <25088@mimsy.umd.edu>
Date: 24 Jun 90 15:09:54 GMT
References: <1581@lectroid.sw.stratus.com> <26888@bcsaic.UUCP>
Reply-To: j...@cs.umd.edu (James da Silva)
Organization: University of Maryland, Department of Computer Science
Lines: 34
Posted: Sun Jun 24 16:09:54 1990
In article <26...@bcsaic.UUCP> pa...@bcsaic.UUCP (Paul Allen) writes:
>Take a look at section 9.10 in the monthly Minix Information Sheet, and
>then ftp to hobbes.cs.umd.edu. I just checked, and the latest (6/20/90)
>version of Bruce's 32-bit kit is now there. Note that this is not for
>the faint of heart. I'm not done building it yet, so I think I'll reserve
>further comment until I've got the whole thing nailed down.
>
>I believe James da Silva is responsible for the little Minix archive
>on hobbes. Thanks, James!
You are welcome, Paul.
Actually, the little hobbes archive is now pretty much redundant. Bruce
Evans is putting his latest stuff on adelaide in Australia, and Glen Overby
keeps a current copy of the adelaide stuff on plains.nodak.edu (in the
pub/Minix/oz directory).
So the discriminating FTP user can find the Minix-386 patches and a raft
of other Minix-related files on plains.nodak.edu. For the Net-less I still
maintain a Minix BBS archive @ 1-301-277-9408. The latest 32 bit stuff is
available there as well.
By the way, this latest version is well polished; armed with cppmake and
shoelace I was able to bring up a simple Minix-386 setup in a couple hours,
and it worked the first time! A hearty thanks (once again) to Bruce Evans
for an awesome contribution. And thanks to Earl Chew for cppmake and
especially shoelace. And to Glen Overby and Andrew Cagney for maintaining
great minix archives on opposite sides of the world. A round of drinks for
everyone, on me! :-)
Jaime
...........................................................................
: domain: j...@cs.umd.edu James da Silva
: path: uunet!mimsy!jds Systems Design & Analysis Group
Path: gmdzi!unido!mcsun!uunet!lll-winken!uwm.edu!wuarchive!udel!mmdf
From: rdc30...@nmrdc1.nmrdc.nnmc.navy.mil, , (LCDR Michael E. Dobson)
Newsgroups: comp.os.minix
Subject: Re: 32-bit Minix (was another naive suggestion)
Message-ID: <22861@nigel.udel.EDU>
Date: 26 Jun 90 12:16:58 GMT
Sender: m...@udel.EDU
Lines: 23
Posted: Tue Jun 26 13:16:58 1990
pal...@atc.boeing.com (Paul Allen) writes:
> Take a look at section 9.10 in the monthly Minix Information Sheet, and
> then ftp to hobbes.cs.umd.edu. I just checked, and the latest (6/20/90)
> version of Bruce's 32-bit kit is now there. Note that this is not for
> the faint of heart. I'm not done building it yet, so I think I'll reserve
> further comment until I've got the whole thing nailed down.
>
It is actually rather easier than the 1.3->1.5.10 upgrade. I pulled the
stuff from ftp.adelaide.oz.au about a week and one-half ago and built a
32-bit kernel in one evening. The shell scripts that Bruce included really
help to keep things straight. The hardest part is getting proper makefiles
and keeping the 16-bit binaries separate from the 32-bit. Separate 10-meg
partitions for each would be a big plus. The only thing lacking right now
is a 32-bit cpp. If we had that, I could throw the 16-bit binaries completely
off my 386 system.
--
Mike Dobson
Sys Admin for nmrdc1.nmrdc.nnmc.navy.mil (AT&T 3B2/600G running Sys V R 3.2)
Internet: rdc30...@nmrdc1.nmrdc.nnmc.navy.mil or
UUCP: ..uunet!nmrdc1.nmrdc.nnmc.navy.mil!rdc30med or
BITNET: dob...@usuhsb.BITNET or
MCI Mail: 377-2719 or 0003772...@NRI.mcimail.com
Path: gmdzi!unido!mcsun!uunet!munnari.oz.au!metro!natmlab.dap.csiro.au!
ditsyda!evans
From: ev...@ditsyda.oz (Bruce.Evans)
Newsgroups: comp.os.minix
Subject: Patch 1 for 386 kernel
Message-ID: <2674@ditsyda.oz>
Date: 2 Jul 90 13:06:03 GMT
References: <3246@leah.Albany.Edu>
Organization: CSIRO DIT Sydney, Australia
Lines: 198
Posted: Mon Jul 2 14:06:03 1990
Although my 386 Minix kernel has not been posted yet, this bug fix seems
important enough to post. Pages were being remapped from low memory to
extended memory beginning at 2M even when 2M was not the end of extended
memory. This causes the corresponding number of pages at the top of the
final memory memory map to be bad. With more than 1M of extended memory,
the top pages are rarely allocated so the bug would not bite often. The
bug only affects the most recent (June 1990) version of the kernel.
There are bugs in test8 that seem to be more dangerous with this kernel.
Test8 calls mknod with the wrong number of arguments. This creates
directories and files with "random" sizes, including possibly negative
sizes. Big sizes cause directory seaches to take "forever", and negative
sizes may cause panics. Test8 should not be run until these calls and
mknod are fixed.
Patch
--
Bruce Evans ev...@ditsyda.syd.dit.csiro.au
Path: gmdzi!unido!mcsun!uunet!bcstec!voodoo!ssc-vax!bcsaic!paula
From: pa...@bcsaic.UUCP (Paul Allen)
Newsgroups: comp.os.minix
Subject: Progress report
Message-ID: <29883@bcsaic.UUCP>
Date: 11 Aug 90 02:08:36 GMT
Organization: Boeing Computer Services ATC, Seattle
Lines: 48
Posted: Sat Aug 11 03:08:36 1990
I've been able to spend a little time lately on Minix, and here's
a brief summary of where I'm at.
I've had a 32-bit 1.5.10 kernel running on my 386 for a couple weeks
now. I think it would perform better if I could figure out how to
get more than 30 buffers. If I build a kernel with more than 30
buffers, it gets a GP fault as it starts loading the ramdisk. Has
anybody using Bruce's 386 kernel stuff found the key to this little
puzzle? Other than that little snag, Bruce's new package slid right
in. Thanks Bruce!
I installed Gordon Irlam's virtual consoles patch a couple weeks
ago and love it. I've got things set up with two virtual screens
and log in as myself on one and root on the other. It's real handy.
Thanks Gordon!
I installed Ken Hendrickson's 43-line EGA patches recently. It took
a bit of care, 'cause some of Ken's changes collided with Gordon's,
especially in console.c. It turns out that Ken's code puts my
clone VGA in 50-line mode. I spent the night before last fixing
things so everybody knows the screen is 50 lines long, and it's
working well. Ken's patches also set the interrupt character to
^C and the kill character to ^U, just like my Sun at work. The only
problem I've seen so far is the fact that mined has the size of
the screen hard-coded into it. But who needs mined, when I've got
vi? Thanks Ken!
Last night I installed elvis. I started with the sources that were
posted early this year by Steve Kirkendall and applied Steve's patches
of Jan 23 and Peter Valkenburg's patches of Feb 12. I compiled it
with Bruce's compiler, and had one little problem. The vi.h header
file declares redrawpre and redrawpost 'extern', but vars.c includes
vi.h and then declares actual storage for redrawpre and redrawpost.
The linker complained about unresolved externals. I #defined a
symbol in vars.c and used it to conditionally compile the extern
declarations of those variables in vi.h. It was a quick hack, so
I don't think I'll post diffs. I guess this wasn't a problem for
any compilers other than bcc? It's great to be back down to just
one editor for all the Unix machines I use! Thanks Steve!
And thank you, Andy, for Minix! This is fun!
Paul Allen
--
------------------------------------------------------------------------
Paul L. Allen | pal...@atc.boeing.com
Boeing Advanced Technology Center | ...!uw-beaver!bcsaic!pallen
Path: gmdzi!unido!mcsun!uunet!decwrl!wuarchive!cs.utexas.edu!rutgers!
neptune.rutgers.edu!frogpond.rutgers.edu!rbthomas
From: rbtho...@frogpond.rutgers.edu (Rick Thomas)
Newsgroups: comp.os.minix
Subject: Re: Bruce's 32 bit kernel
Message-ID: <Aug.21.17.05.50.1990.4426@frogpond.rutgers.edu>
Date: 21 Aug 90 21:05:52 GMT
References: <27801@nigel.ee.udel.edu>
Organization: Rutgers Engineering Supercomputer Lab
Lines: 26
Posted: Tue Aug 21 22:05:52 1990
Cc: rbtho...@jove.rutgers.edu
What would be the reaction to a "386-PC-clone" Minix distribution,
distinct from the 8086/80286 version, but available from P-H just like
them and the Atari and Mac (and so on) versions?
The big problem would seem to be getting a compiler that generates
32-bit 386 code and can be freely distributed by P-H. I understand
that GCC cannot be so distributed, because of the requirement of the
copy-left that source be made available. (Making source available on
demand would be more hassle and expense than P-H would be willing to
undertake.) I also understand Bruce's reluctance to have his compiler
distributed to those who do not understand the experimental nature of
the 'product', who would try to use it for 'production' and complain
when it didn't work. Perhaps there is an ACK 386 compiler that could
be distributed.
The problem of keeping the source code in sync would seem to be not
significantly worse for n versions (PC/Atari/Mac/Amiga/etc) than for
n+1 versions (all of the above plus 386), so I don't think there should
be any objection from that score.
How about it? Andy, is such a thing possible? Everybody else, would
you buy it if it were available?
Enjoy!
Rick
Path: gmdzi!unido!mcsun!hp4nl!star.cs.vu.nl!ast
From: a...@cs.vu.nl (Andy Tanenbaum)
Newsgroups: comp.os.minix
Subject: Re: Bruce's 32 bit kernel
Message-ID: <7374@star.cs.vu.nl>
Date: 23 Aug 90 18:31:31 GMT
References: <27801@nigel.ee.udel.edu>
<Aug.21.17.05.50.1990.4426@frogpond.rutgers.edu>
Sender: n...@cs.vu.nl
Organization: Fac. Wiskunde & Informatica, VU, Amsterdam
Lines: 13
Posted: Thu Aug 23 19:31:31 1990
In article <Aug.21.17.05.50.1990.4...@frogpond.rutgers.edu>
rbtho...@frogpond.rutgers.edu (Rick Thomas) writes:
>What would be the reaction to a "386-PC-clone" Minix distribution,
Maybe some day, but we are not there yet. If and when everything works,
and has been distributed on the net in the usual way, and we are convinced
it is a solid, stable system, I might think about integrating into the
main system. I don't think I want the headache of maintaining any more
distinct systems than I have to. Actually, from what I understand, the
386 shouldn't require a large number of changes to FS or MM.
This would argue more for integrating the 386 stuff into the main distribution
than making a separate one.
Andy Tanenbaum (a...@cs.vu.nl)
Path: gmdzi!unido!mcsun!sunic!uupsi!rpi!zaphod.mps.ohio-state.edu!
uakari.primate.wisc.edu!uflorida!mailer.cc.fsu.edu!sun13!
sun8.scri.fsu.edu!nall
From: n...@sun8.scri.fsu.edu (John Nall)
Newsgroups: comp.os.minix
Subject: Upgrade to Minix-386
Message-ID: <1091@sun13.scri.fsu.edu>
Date: 11 Oct 90 16:54:21 GMT
Sender: n...@sun13.scri.fsu.edu
Reply-To: n...@sun8.scri.fsu.edu (John Nall)
Organization: SCRI, Florida State University
Lines: 60
Posted: Thu Oct 11 17:54:21 1990
The time has finally come (the walrus said...) to proceed
with the upgrade to Minix-386. I've started, and have run
into a minor problem -- perhaps someone would be so kind as
to send me a brief note as to the problem.
I have done the following:
(1) Pulled everything from the "oz" subdirectory
of plains.nodak.edu (yes, I know that I will
not need it all for Minix-386, but most of it
I wanted in any case).
(2) Did the necessary unpacking etc.
(3) Made a directory structure that follows the
outline of the "runtime.crc" file which is
in the mx386_1.1 subdirectory.
(4) Begun moving files to flesh out the directory
structure, and checking crc's to be sure they
match.
This is where the problem is. Bruce seems to have two sets
of include nad library files. For example:
/usr/include/dist/lib.h
and
/usr/include/lib.h
The /usr/include/dist/lib.h appears to be the distributed
version (presumably that is what the dist stands for), and the
crc checks ok.
However, I cannot find in any of the oz files that I got another
copy of lib.h. There are quite a few that are in this same
category. They are:
(under /usr/include/)
fcntl.h, lib.h, limits.h, math.h, stdlib.h, and unistd.h
(under /usr/src/lib/ansi/)
assert.c, ctime.c, difftime.c, strtol.c, strtoul.c
(under /usr/src/lib/other/)
lock.c, nlist.c
(under /usr/src/lib/posix/)
execlp.c, getcwd.c
Also, so long as I am asking, I am going on the presumption that
I should use bcc32 from the start (in order to build the 386 kernel),
and that there is no need to use bcc16 at all. Is that correct? Or
is there some sort of bootstrapping necessary which utilizes bcc16?
Thanks for any info. Please e-mail me directly. I'm going to try and
keep a good set of notes as I go, and will try and make up a tutorial
after I finish for net distribution (following Glen Overby's example).
--
John W. Nall | Supercomputation Computations Research Institute
n...@sun8.scri.fsu.edu | Florida State University, Tallahassee, FL 32306
"Real programmers can write assembly code in any language." - Larry Wall
Path: gmdzi!unido!mcsun!sunic!uupsi!rpi!zaphod.mps.ohio-state.edu!
samsung!emory!gatech!prism!sun13!sun8.scri.fsu.edu!nall
From: n...@sun8.scri.fsu.edu (John Nall)
Newsgroups: comp.os.minix
Subject: Re: Upgrade to Minix-386
Message-ID: <1093@sun13.scri.fsu.edu>
Date: 11 Oct 90 20:27:29 GMT
References: <1091@sun13.scri.fsu.edu>
Sender: n...@sun13.scri.fsu.edu
Organization: SCRI, Florida State University
Lines: 41
Posted: Thu Oct 11 21:27:29 1990
In article <1...@sun13.scri.fsu.edu> n...@sun8.scri.fsu.edu I wrote:
>The time has finally come (the walrus said...) to proceed
>with the upgrade to Minix-386. I've started, and have run
>
>This is where the problem is. Bruce seems to have two sets
>of include and library files. For example:
>
> /usr/include/dist/lib.h
> and
> /usr/include/lib.h
>
Stupid, stupid, stupid! I just didn't look hard enough.
They are there (in the bcclib stuff). :-<
However, although I have all the library files cleared up,
and crc's check on almost everything, I STILL have some bad crc's
on some files in /usr/src/tools (as indicated in Bruce's list). To
wit:
/usr/src/tools bootblok.s, build.c and menu1.s
Are there some .cdif files for these?
Also, the crc on /usr/src/mm/makefile.cpp is wrong. Bruce shows
63284 2009 /usr/src/mm/makefile.cpp
and I show
07128 1942 /usr/src/mm/makefile.cpp
Since this file came directly from the archives, I'm not sure why
there is a problem. Anyone know?
Thanks for any any and all suggestions.
--
John W. Nall | Supercomputation Computations Research Institute
n...@sun8.scri.fsu.edu | Florida State University, Tallahassee, FL 32306
"Real programmers can write assembly code in any language." - Larry Wall
Path: gmdzi!unido!mcsun!uunet!lll-winken!sun-barr!cs.utexas.edu!swrinde!
zaphod.mps.ohio-state.edu!wuarchive!emory!gatech!prism!sun13!
sun8.scri.fsu.edu
From: n...@sun8.scri.fsu.edu (John Nall)
Newsgroups: comp.os.minix
Subject: Volunteer to test Minix-386 tutorial
Message-ID: <1319@sun13.scri.fsu.edu>
Date: 2 Nov 90 14:47:24 GMT
Sender: n...@sun13.scri.fsu.edu
Organization: SCRI, Florida State University
Lines: 27
Posted: Fri Nov 2 15:47:24 1990
I have successfully completed installing the Bruce Evans modifications
for 32-bit Minix (Minix-386) to standard Minix 1.5.10, thanks to the
help of quite a few people on the net. While the various problems which
were encountered are still fresh in mind, I've written a tutorial which
may keep the wheel from constantly being reinvented. (It is amazing how
many details which are puzzling are perfectly obviously once you know how
to do it). Needless to say, this tutorial is a composite of information
received from the people who have helped me, with some additions which are
added for the sake of clarity.
Before releasing this tutorial for general consumption, I need someone
to test it out for me. I've gone through and checked it, but that does
not mean it would necessarily be clear to someone else who is hitting
it cold.
The necessary qualifications are (a) a 386, (b) some time to work on this,
which I would estimate at about four or five hours total, (c) a willingness
to make comments on the tutorial, and (d) an already installed Minix 1.5.10.
What you get out of it is (a) a chance to do a good deed :-), (b) an extra
amount of hand-holding as necessary, (c) a Minix-386 system.
Please e-mail direct to me. Address is n...@sun8.scri.fsu.edu.
--
John W. Nall | Supercomputation Computations Research Institute
n...@sun8.scri.fsu.edu | Florida State University, Tallahassee, FL 32306
"Real programmers can write assembly code in any language." - Larry Wall
Path: gmdzi!unido!mcsun!uunet!zaphod.mps.ohio-state.edu!ncar!gatech!prism!
mailer.cc.fsu.edu!sun13!sun8.scri.fsu.edu
From: n...@sun8.scri.fsu.edu (John Nall)
Newsgroups: comp.os.minix
Subject: Minix-386 Tutorial
Message-ID: <1362@sun13.scri.fsu.edu>
Date: 7 Nov 90 20:51:19 GMT
Sender: n...@sun13.scri.fsu.edu
Organization: SCRI, Florida State University
Lines: 688
Posted: Wed Nov 7 21:51:19 1990
As promised, here is the tutorial for moving from standard
Minix 1.5.10 to Minix-386. Many thanks to Bruce Evans, who
took time out from his schedule to read the draft and point
out some (obvious after the fact!) mistakes.
Hopes this help move more people over to Minix-386. It is
a very exciting system to use, and has the potential for
being a very powerful tool. Since I'm a teacher, I see the
benefits from the education side, but think that there are
also a lot of pluses for power users.
The following is a shar file.
Tutorial
--
John W. Nall | Supercomputation Computations Research Institute
n...@sun8.scri.fsu.edu | Florida State University, Tallahassee, FL 32306
"Real programmers can write assembly code in any language." - Larry Wall