![[MINIX LOGO]](/m4-4.png)
From: Jac0b <jac0b@thisaintreal.com>
Newsgroups: comp.os.minix
Subject: Update Procedure for Minix-vmd 1.7.0
Date: Mon, 21 Oct 2002 16:08:53 -0400
Organization: Posted via Supernews, http://www.supernews.com
Message-ID: <ur8nnimn7jcs72@corp.supernews.com>
User-Agent: KNode/0.7.1
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7Bit
X-Complaints-To: abuse@supernews.com
Lines: 126
Hello. I recently undertook the task of updating my Minix-vmd 1.7.0
installation to the current source. I would like to thank Philip Homburg
and Kees Bot for their support and answering my many many questions. I call
this an update procedure because I have not read any mention of an actual
upgrade of Minix-vmd 1.7.0. This said, here are the steps I took to
'update' my installation of Minix-vmd 1.7.0
Assumptions:
1. Minix-vmd 1.7.0 is installed.
2. You are able to compile programs and the kernel under Minix-vmd 1.7.0
To test, try:
# cd /usr/src
# make world <-- check for any errors
# build <-- check for any errors
If you don't see any errors from the above commands, all is good.
Update Procedure
1. With a web browser, go to
http://www.minix-vmd.org/cgi-bin/raw/current/vmd/
2. Download the following files
a. include.tgz
b. man.tgz
c. src.tgz
d. usretc.tgz
e. usrlib.tgz
** This next step (3) is for those who do not have network support on their
Minix-vmd. If you have network support, skip to step 4
*** For those without network support in Minix-vmd, steps 3 and 4 can be
alternated to save on the number of disks required. For example, say we
need to transfer src.tgz. Instead of using up 8 disks or so, we can simply
just do fdvol 1440 A: src.tgz (or dd if=src.tgz bs=1440k count=1
of=/dev/fd0 skip=0). When the first disk is done, do step 4 on that disk.
While that disk is decompressing, insert the second disk to get the next
portion of the file and so on. This will be more clear after you read steps
3 and 4.
3. Now we need to get those files over to Minix-vmd. It differs slightly
depending on what OS you are using.
a. Windows
fdvol 1440 A: include.tgz
fdvol 1440 A: man.tgz
fdvol 1440 A: src.tgz
etc
b. Linux
dd if=include.tgz bs=1440k count=1 of=/dev/fd0 skip=0
dd if=man.tgz bs=1440k count=1 of=/dev/fd0 skip=0
dd if=src.tgz bs=1440k count=1 of=/dev/fd0 skip=0
etc.
** Note, you will need to increment _skip_ until a partial disk is
written.
For example,
dd if=src.tgz bs=1440k count=1 of=/dev/fd0 skip=1
dd if=src.tgz bs=1440k count=1 of=/dev/fd0 skip=2
dd if=src.tgz bs=1440k count=1 of=/dev/fd0 skip=3
etc
I noticed that this happened when I saw the output:
0+1 records in
0+1 records out
But, I'm not really sure. However, you'll know when to stop when the
'vol' command, in step 4, stops asking for disks.
4. Next we use the Minix-vmd vol command to get the file onto the Minix-vmd
installation.
** Note: Existing files will be over-written. This should be alright since
we are performing an update.
a. For include.tgz
i. # cd /usr
ii. # vol 1440 /dev/fd0 | gzcat | tar xvfp -
b. For man.tgz
i. # cd /usr
ii. # vol 1440 /dev/fd0 | gzcat | tar xvfp -
c. For src.tgz
i. # cd /usr
ii. # vol 1440 /dev/fd0 | gzcat | tar xvfp -
d. For usretc.tgz
i. # cd /usr
ii. # vol 1440 /dev/fd0 | gzcat | tar xvfp -
iii. # rm -rf /usr/etc
iv. # mv /usr/usretc /usr/etc
e. For usrlib.tgz
i. # cd /usr
ii. # vol 1440 /dev/fd0 | gzcat | tar xvfp -
iii. # rm -rf /usr/lib
iv. # mv /usr/usretc /usr/lib
5. Now we need to update /etc and /var
** Note that any customizations to /etc/rc will be lost
a. # cd
b. # checkhier | sh
6. Minix-vmd uses the same device naming convention as Minix 2.0.3. For
example:
hd1a ==> c0d0p0s0
hd1b ==> c0d0p0s1
hd1c ==> c0d0p0s2
Therefore, we need to make the proper devices
a. # cd /dev
b. # MAKEDEV std
7. Now we compile the libraries, commands and kernel
a. # cd /usr/src
b. # make world
c. # build
8. We need to update the bootloader as well.
a. # cd /
b. # mv boot boot.old
c. # cp /usr/mdec/boot .
d. # installboot -d /dev/c0d0p0s0 /usr/mdec/bootblock boot
9. Now reboot.
a. # shutdown -r now
As long as I haven't forgotten anything, you should now be running Minix-vmd
2.0r0. Be sure to keep checking this thread for replies/comments. There may
be errors in this document and will be identified along with the necessary
corrections.
Enjoy,
Jacob
From: Martijn van Buul <pienjo@c64.org>
Newsgroups: comp.os.minix
Subject: Re: Update Procedure for Minix-vmd 1.7.0
Date: 28 Oct 2002 13:59:37 GMT
Organization: Eindhoven University of Technology
Lines: 84
Message-ID: <slrnarqgm8.kkq.pienjo@mud.stack.nl>
References: <ur8nnimn7jcs72@corp.supernews.com>
Reply-To: pino+comp_os_minix@dohd.org (Martijn van Buul)
NNTP-Posting-Host: mud.stack.nl (131.155.141.70)
X-Trace: fu-berlin.de 1035813577 2242716 131.155.141.70 (16 [32225])
X-Favourite-Drink: Alcoholic, yet fluid
User-Agent: slrn/0.9.7.4 (FreeBSD)
It occurred to me that Jac0b wrote in comp.os.minix:
Nice work. A few comments.
> 2. You are able to compile programs and the kernel under Minix-vmd 1.7.0
> To test, try:
> # cd /usr/src
> # make world <-- check for any errors
> # build <-- check for any errors
> If you don't see any errors from the above commands, all is good.
Bit of an overkill, right? Especially on slower machines, this takes quite
a while, and it's really not required to rebuild the entire system to check
whether the installed system is halfway sane, especially since you're going
to overwrite major parts of it (/usr/include, /usr/lib...) in the next
part.
> ** Note, you will need to increment _skip_ until a partial
> disk is written. For example,
> dd if=src.tgz bs=1440k count=1 of=/dev/fd0 skip=1
> dd if=src.tgz bs=1440k count=1 of=/dev/fd0 skip=2
> dd if=src.tgz bs=1440k count=1 of=/dev/fd0 skip=3
> etc
> I noticed that this happened when I saw the output:
> 0+1 records in
> 0+1 records out
> But, I'm not really sure. However, you'll know when to
> stop when the 'vol' command, in step 4, stops asking for
> disks.
Check. The manpage on dd is rather clear about that. That said, there might
be other reasons dd bails out early, most notably a bad disk. You can estimate
the number of required disks, ofcourse, by divide the filesize by 1.4MB
> 5. Now we need to update /etc and /var
No, you should do that *after* the compilation of commands, since you
end up using the 1.7.0 checkhier now! Furthermore, I may be a bit paranoid,
but I *always* check the output of 'script-generating scripts' like
checkhier for possibly disastrous effects.
> 6. Minix-vmd uses the same device naming convention as Minix 2.0.3. For
> example:
> hd1a ==> c0d0p0s0
> hd1b ==> c0d0p0s1
> hd1c ==> c0d0p0s2
> Therefore, we need to make the proper devices
> a. # cd /dev
> b. # MAKEDEV std
Again, you should do that *AFTER* the commands has been built, as you're now
using 1.7.0's Makedev, which will not generate any new device nodes!
> 7. Now we compile the libraries, commands and kernel
> a. # cd /usr/src
> b. # make world
> c. # build
That didn't work right away. Make failed while trying to rebuild make the
first time I tried. A manual installation worked. Gzip doesn't install;
don't know why, as a manual 'make install' works like a charm, while the
'make install' done by the master makefile bombs out. I had to patch the
Makefile in /usr/src/gnu/cmd. :x
I later found out that the various makefiles (especially build!) expect
/usr/sbin to be in the search path. 1.7.0 didn't even have a /usr/sbin,
IIRC, so you need to manually add it to /root/.ashrc (and $user/.ashrc if
you're su-ing to root). There's a PATH line there. If you don't, you'll
end up with things like "build: not found" or "installboot: not found".
Between the 'make world' and the 'build', I strongly suggest doing a
'hash -r', to ensure that the shell is using the correct binaries.
After the build, it might be a good time to do the checkhier and MAKEDEV
stuff. Note that if you updated /etc stuff yourself, you cannot expect
checkhier to work - check things manually! /etc/rc is a good example,
for instance. In my case, it didn't get copied; my own /etc/rc was newer..
--
Martijn van Buul - Pino@dohd.org - http://www.stack.nl/~martijnb/
Geek code: G-- - Visit OuterSpace: mud.stack.nl 3333
Kees J. Bot: The sum of CPU power and user brain power is a constant.
From: ctantignone@hotmail.com (Claudio Tantignone) Newsgroups: comp.os.minix Subject: Re: Update Procedure for Minix-vmd 1.7.0 Date: 7 Nov 2002 03:45:57 -0800 Organization: http://groups.google.com/ Lines: 31 Message-ID: <c5f4fa83.0211070345.1c989191@posting.google.com> References: <ur8nnimn7jcs72@corp.supernews.com> <slrnarqgm8.kkq.pienjo@mud.stack.nl> NNTP-Posting-Host: 200.10.105.42 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit X-Trace: posting.google.com 1036669557 28381 127.0.0.1 (7 Nov 2002 11:45:57 GMT) X-Complaints-To: groups-abuse@google.com NNTP-Posting-Date: 7 Nov 2002 11:45:57 GMT Martijn van Buul <pienjo@c64.org> wrote in message news:<slrnarqgm8.kkq.pienjo@mud.stack.nl>... > It occurred to me that Jac0b wrote in comp.os.minix: > > Nice work. A few comments. > I tried the update procedure. I have VMD working in a dosfile. I update everything as descripted in your postings. The first problem with make world was that LLgen is missing trying to compile LLgen. I do not know how compile it if LLgen is missing. The readme indicates using make firstinstall but this key does not exists. So I comment it into Makefile. After that, every works, I compile and install all commands and kernel. But it does not boot. Booting with the old DOSBOOT.COM, the kernel indicates a problem with some structure in memory. As I remember a similar problem with standard minix 2.0.3, I copy the BOOT.COM generated in this new compilation (I booted and old copy of VMD and copy it with mtools). The new BOOT.COM also does not work, it loads the kernel but reboots inmediattely so I can not get any error message. It reboot while loading some tak like fs/mm or inet, I can not see exactly the error because it cleans the screen and restarts the PC (to BIOS tests). Did someone test the 2.0r0 release using a dosfile? Thanks in advance. Claudio.
From: "Jacob" <jacob@thisaintreal.com> Newsgroups: comp.os.minix Subject: Re: Update Procedure for Minix-vmd 1.7.0 Date: Thu, 7 Nov 2002 10:05:11 -0500 Organization: Posted via Supernews, http://www.supernews.com Message-ID: <usl09kbpp1dl4e@corp.supernews.com> References: <ur8nnimn7jcs72@corp.supernews.com> <slrnarqgm8.kkq.pienjo@mud.stack.nl> <c5f4fa83.0211070345.1c989191@posting.google.com> X-Priority: 3 X-MSMail-Priority: Normal X-Newsreader: Microsoft Outlook Express 6.00.2800.1106 X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2800.1106 X-Complaints-To: abuse@supernews.com Lines: 40 "Claudio Tantignone" <ctantignone@hotmail.com> wrote in message > The first problem with make world was that LLgen is missing trying to > compile LLgen. I do not know how compile it if LLgen is missing. The > readme indicates using make firstinstall but this key does not exists. > > So I comment it into Makefile. After that, every works, I compile and > install all commands and kernel. That's what I had to do too. I believe that LLgen will be removed. I'm not sure though. I would be neat to play with cause I'm writing a Recursive-Descent parser for a compiler course. It would be neat to compare what it produces with what I produce. :) > But it does not boot. Booting with the old DOSBOOT.COM, ... I copy the BOOT.COM > generated in this new compilation (I booted and old copy of VMD and > copy it with mtools). I hate to suggest this, but have you tried the update procedure from scratch again? But, this time copying the newly compiled BOOT.COM? I haven't used Minix via a dosfile so my suggestions may be problematic. > The new BOOT.COM also does not work, it loads the kernel but reboots > inmediattely so I can not get any error message. It reboot while > loading some tak like fs/mm or inet, I can not see exactly the error > because it cleans the screen and restarts the PC (to BIOS tests). Did you run the checkhier command to make sure all the /dev/ and /etc/rc is updated? the 2.0r0 uses a different scheme for the hard drivers. Hope some of this is usefull, Jacob
From: ctantignone@hotmail.com (Claudio Tantignone) Newsgroups: comp.os.minix Subject: Re: Update Procedure for Minix-vmd 1.7.0 Date: 8 Nov 2002 04:07:14 -0800 Organization: http://groups.google.com/ Lines: 46 Message-ID: <c5f4fa83.0211080407.2dbba178@posting.google.com> References: <ur8nnimn7jcs72@corp.supernews.com> <slrnarqgm8.kkq.pienjo@mud.stack.nl> <c5f4fa83.0211070345.1c989191@posting.google.com> <usl09kbpp1dl4e@corp.supernews.com> NNTP-Posting-Host: 200.10.105.42 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit X-Trace: posting.google.com 1036757234 2532 127.0.0.1 (8 Nov 2002 12:07:14 GMT) X-Complaints-To: groups-abuse@google.com NNTP-Posting-Date: 8 Nov 2002 12:07:14 GMT "Jacob" <jacob@thisaintreal.com> wrote in message news:<usl09kbpp1dl4e@corp.supernews.com>... > > That's what I had to do too. I believe that LLgen will be removed. I'm not > sure though. ok. > > I hate to suggest this, but have you tried the update procedure from scratch > again? But, this time > copying the newly compiled BOOT.COM? I haven't used Minix via a dosfile so > my suggestions may > be problematic. Yes, i tried various times, first I had no space in the partition, and need to make a bigger dosfile. Also I was checking for DOSFILE support and everything seems to be enabled. > > Did you run the checkhier command to make sure all the /dev/ and /etc/rc is > updated? > the 2.0r0 uses a different scheme for the hard drivers. Yes. I update all. > > Hope some of this is usefull, > Any idea is usefull. Yesterday I tried with a new kernel, compiled step by step. I disable inet server also, but it does not work. So I copied the boot partition using dd to a floppy and make it booteable with installboot, and it boots ok, but I have not very clear how I need to configure the monitor variables to point to my dosfile in FAT. I believe that can be some misconfiguration because the same kernel can boot from floppy. Or some missing support that I need to enable to use dosfile. If I point the image variable to 1.7.0 kernel it works fine. The dosfile is in FAT 16, 2nd IDE hard disk, primary partition. Any ideas is welcome. Regards, Claudio.
From: kjb=733043@cs.vu.nl (Kees J Bot) Newsgroups: comp.os.minix Subject: Re: Update Procedure for Minix-vmd 1.7.0 Date: Fri, 8 Nov 2002 11:20:37 +0100 Organization: One Minix-vmd system, it works! Lines: 28 Message-ID: <l53gqa.5fj.ln@jetsam.cs.vu.nl> References: <ur8nnimn7jcs72@corp.supernews.com> <slrnarqgm8.kkq.pienjo@mud.stack.nl> <c5f4fa83.0211070345.1c989191@posting.google.com> NNTP-Posting-Host: jetsam.cs.vu.nl In article <c5f4fa83.0211070345.1c989191@posting.google.com>, Claudio Tantignone <ctantignone@hotmail.com> wrote: > >The new BOOT.COM also does not work, It does indeed not work. We now have the strange situation that BOOT.COM is built under Minix-vmd, but can only be used with Minix. The old BOOT.COM did some pretty horrible things to allow Minix to run. It moved things around in memory, so that Minix could sit in its normal place. If you made a BIOS call then Minix and DOS stuff would be exchanged, and unexchanged to make the call. To get rid of this idiocy I changed Minix so that the kernel can run anywhere in low memory and like it. Alas Minix-vmd isn't so flexible. The virtual memory code is written around the assumption that kernel data addresses are equal to absolute memory addresses. The way I imagined that this could be solved is to link the kernel at a high address (1 gig for instance), to let the bootstrap load the kernel in extended memory, set up a minimum page table to allow the kernel to run and run it. This is quite a bit of work, a lot of fun work even, but I never got to it because I think that DOSMinix has become an obsolete relic for only being able to run under DOS, Windows 95, and Windows 98. -- Kees J. Bot, Systems Programmer, Sciences dept., Vrije Universiteit Amsterdam
From: ctantignone@hotmail.com (Claudio Tantignone) Newsgroups: comp.os.minix Subject: Re: Update Procedure for Minix-vmd 1.7.0 Date: 8 Nov 2002 10:46:10 -0800 Organization: http://groups.google.com/ Lines: 25 Message-ID: <c5f4fa83.0211081046.3fb00dad@posting.google.com> References: <ur8nnimn7jcs72@corp.supernews.com> <slrnarqgm8.kkq.pienjo@mud.stack.nl> <c5f4fa83.0211070345.1c989191@posting.google.com> <l53gqa.5fj.ln@jetsam.cs.vu.nl> NNTP-Posting-Host: 200.10.105.42 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit X-Trace: posting.google.com 1036781170 31865 127.0.0.1 (8 Nov 2002 18:46:10 GMT) X-Complaints-To: groups-abuse@google.com NNTP-Posting-Date: 8 Nov 2002 18:46:10 GMT kjb=733043@cs.vu.nl (Kees J Bot) wrote in message news:<l53gqa.5fj.ln@jetsam.cs.vu.nl>... > In article <c5f4fa83.0211070345.1c989191@posting.google.com>, > Claudio Tantignone <ctantignone@hotmail.com> wrote: > > > >The new BOOT.COM also does not work, > > It does indeed not work. We now have the strange situation that > BOOT.COM is built under Minix-vmd, but can only be used with Minix. > ... > > This is quite a bit of work, a lot of fun work even, but I never got to > it because I think that DOSMinix has become an obsolete relic for only > being able to run under DOS, Windows 95, and Windows 98. ok, dosminix allows me to avoid repartition hd. My final idea is to use part of the pci code to access net card in an virtual environment like vmware. Well, I will continue booting the floppy, and I hope to discover how to point /usr to dosfile if possible. Regards, Claudio.
From: kjb=733043@cs.vu.nl (Kees J Bot) Newsgroups: comp.os.minix Subject: Re: Update Procedure for Minix-vmd 1.7.0 Date: Fri, 8 Nov 2002 22:15:17 +0100 Organization: One Minix-vmd system, it works! Lines: 31 Message-ID: <5h9hqa.ppl.ln@jetsam.cs.vu.nl> References: <ur8nnimn7jcs72@corp.supernews.com> <c5f4fa83.0211070345.1c989191@posting.google.com> <l53gqa.5fj.ln@jetsam.cs.vu.nl> <c5f4fa83.0211081046.3fb00dad@posting.google.com> NNTP-Posting-Host: jetsam.cs.vu.nl In article <c5f4fa83.0211081046.3fb00dad@posting.google.com>, Claudio Tantignone <ctantignone@hotmail.com> wrote: >kjb=733043@cs.vu.nl (Kees J Bot) wrote in message news:<l53gqa.5fj.ln@jetsam.cs.vu.nl>... >> >> This is quite a bit of work, a lot of fun work even, but I never got to >> it because I think that DOSMinix has become an obsolete relic for only >> being able to run under DOS, Windows 95, and Windows 98. > >Well, I will continue booting the floppy, and I hope to discover how >to point /usr to dosfile if possible. With the help of the fatfile driver you should be able to access the disk file. An outtake from the boot(8) manual page: c0 = fatfile c1 = at fatfile-d0 = at:1:/minix/minix.mnx The 'c1' setting isn't necessary under Minix-vmd (it is under Minix), but you may want to access the real disk as well using mtools or so. The big drawback compared to the BOOT.COM way is that the boot parameters and kernel image come from the floppy instead of the root device. Other than that the system functions normally, with / and /usr on the disk file. The dosfile driver is dead weight under Minix-vmd. It compiles, it probably works, but you can't use it. If I were making a distribution now I'd probably kick it out. -- Kees J. Bot, Systems Programmer, Sciences dept., Vrije Universiteit Amsterdam
From: ctantignone@hotmail.com (Claudio Tantignone) Newsgroups: comp.os.minix Subject: Re: Update Procedure for Minix-vmd 1.7.0 Date: 10 Nov 2002 06:05:21 -0800 Organization: http://groups.google.com/ Lines: 16 Message-ID: <c5f4fa83.0211100605.25a499cf@posting.google.com> References: <ur8nnimn7jcs72@corp.supernews.com> <c5f4fa83.0211070345.1c989191@posting.google.com> <l53gqa.5fj.ln@jetsam.cs.vu.nl> <c5f4fa83.0211081046.3fb00dad@posting.google.com> <5h9hqa.ppl.ln@jetsam.cs.vu.nl> NNTP-Posting-Host: 200.16.146.165 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit X-Trace: posting.google.com 1036937121 19530 127.0.0.1 (10 Nov 2002 14:05:21 GMT) X-Complaints-To: groups-abuse@google.com NNTP-Posting-Date: 10 Nov 2002 14:05:21 GMT kjb=733043@cs.vu.nl (Kees J Bot) wrote in message news:<5h9hqa.ppl.ln@jetsam.cs.vu.nl>... > > With the help of the fatfile driver you should be able to access the > disk file. An outtake from the boot(8) manual page: > > c0 = fatfile > c1 = at > fatfile-d0 = at:1:/minix/minix.mnx > ok, this works fine. It is working. Also under VMware. Now I need to read the man pages about the configurations parameters. Regards, Claudio.
From: Martijn van Buul <pienjo@c64.org>
Newsgroups: comp.os.minix
Subject: Re: Update Procedure for Minix-vmd 1.7.0
Date: 8 Nov 2002 13:53:50 GMT
Organization: Eindhoven University of Technology
Lines: 13
Message-ID: <slrnasngfe.1du4.pienjo@mud.stack.nl>
References: <ur8nnimn7jcs72@corp.supernews.com> <slrnarqgm8.kkq.pienjo@mud.stack.nl> <c5f4fa83.0211070345.1c989191@posting.google.com>
Reply-To: pino+comp_os_minix@dohd.org (Martijn van Buul)
NNTP-Posting-Host: mud.stack.nl (131.155.141.70)
X-Trace: fu-berlin.de 1036763630 10216603 131.155.141.70 (16 [32225])
X-Favourite-Drink: Alcoholic, yet fluid
User-Agent: slrn/0.9.7.4 (FreeBSD)
It occurred to me that Claudio Tantignone wrote in comp.os.minix:
> The first problem with make world was that LLgen is missing trying to
> compile LLgen. I do not know how compile it if LLgen is missing. The
> readme indicates using make firstinstall but this key does not exists.
You need to fetch the original tarball - the version shipped with Minix-VMD
is incomplete in this respect. See the pointer in that README.
--
Martijn van Buul - Pino@dohd.org - http://www.stack.nl/~martijnb/
Geek code: G-- - Visit OuterSpace: mud.stack.nl 3333
Kees J. Bot: The sum of CPU power and user brain power is a constant.
| [HOME] | [HINTS/FAQ] | [MINIX DOWNLOADS] | [CONTRIB SOFTWARE] |
| [NET SOFTWARE] | [MINIX-VMD] | [TEXTBOOK] | [LINKS] |
![[Valid HTML 4.0!]](/validh_40.png){kind=small}