Subject: Linux-Development Digest #414
From: Digestifier <Linux-Development-Request@senator-bedfellow.MIT.EDU>
To: Linux-Development@senator-bedfellow.MIT.EDU
Reply-To: Linux-Development@senator-bedfellow.MIT.EDU
Date:     Thu, 27 Jan 94 23:13:05 EST

Linux-Development Digest #414, Volume #1         Thu, 27 Jan 94 23:13:05 EST

Contents:
  Re: Winders NT on MC68K? (David C. Niemi)
  Re: Mkimage (David C. Niemi)
  Re: Winders NT on MC68K? (was Re: Which is better? 680x0 or 80x86?) (James A. Crotinger)
  Video7 with XFree86 ? (Alen Tihi)
  Re: Upper Memory Blocks ?? (David C. Niemi)
  CSLIP Server running! PROXY ARP instead of static routes ????? (David-Michael Lincke)
  Re: CD-rom's for Linux (Lee Heins)
  Re: Winders NT on MC68K?  (was Re: Which is better? 680x0 or 80x86?) (Jerry Shekhel)
  Re: Winders NT on MC68K?  (was Re: Which is better? 680x0 or 80x86?) (Jerry Shekhel)
  Re: Winders NT on MC68K? (Miguel de Icaza)
  Re: Winders NT on MC68K?  (was Re: Which is better? 680x0 or 80x86?) (Miguel de Icaza)

----------------------------------------------------------------------------

From: niemidc@YP.lab (David C. Niemi)
Subject: Re: Winders NT on MC68K?
Date: 27 Jan 1994 17:07:10 GMT
Reply-To: niemidc@oasis.gtegsc.com

In article 14q@sol.ctr.columbia.edu, jerry@msi.com (Jerry Shekhel) writes:
>David C. Niemi (niemidc@YP.lab) wrote:
>: >Windows NT (which is multi-tasking, though not that Unix-like) is
>: >designed to be portable, unlike earlier 80x86 only MS systems. A 680x0
>: >port is possible (or so someone at Microsoft talking about the
>: >technical feasibility of an Amiga port) (though not necessarily
>: >likely).
>:
>: While it is true that Windows NoT is not very UN*X-like, it is not
>: as portable as most flavors of UN*X, or even as somewhat hardware-
>: specific flavors of UN*X like Linux.
>:
>
>Wrong.  NT runs on three different architectures right now.  Linux doesn't.
>In fact, no single version of Unix does.  Let's not speculate on Linux's
>portability (relative to NT's) until Linux is actually ported to as many
>architectures as NT supports.

Look, Jerry, I agree this is all speculation.  However, simply because NT
runs on 3 architectures does not indicate that is as "portable" as any
given flavor of UN*X, several of which have been ported to 3 or more
processor families.  BSD, SVR3, and SVR4 have each been ported to over a
DOZEN different processors, of both and mixed endiannesses.  Try that with
NT!



[...]
>: NT's downfall is that it is hard-coded to little-endian processors.
>: Endianness (aka byte sex or the "NUXI" issue) is a problem that UN*X
>: solved 15 years ago, when IBM mainframes where big endian and PDPs were
>: little-endian, who knows what Cybers and Honeywells were...
>:
>: Microsoft indicated no willingness to fix their code to run on
>: big-endian processors, so a 68k port is RIGHT OUT.  All the talk
>: about running on RISC chips is on either bi-endian RISC, or on
>: special little-endian versions of the mainstream big-endian RISC
>: chips.

>Your comments are misleading.  True, MS indicated no willingness to port
>NT to big-endian CPUs.  However, that doesn't mean that NT is as you say
>"hard-coded to little-endian processors".  There is no evidence either way
>whatsoever.  MS may simply be avoiding binary file incompatibilities and
>byte swapping issues.

No, my comments are NOT misleading.  MS is not just avoiding some theoretical
binary file incompatibilities (even if they were, there are much better ways
to handle binary data issues); Microsoft's code is simply not designed to work
on big-endian processors, and it would take significant effort to fix it, per
our friendly Microsoft rep.  If you have info to the contrary, please cite
your source rather than just speculating as to Microsoft's motives.


>As far as a 68K port is concerned, what would be the point?  A 68K version
>of NT would run on what, exactly?  Old discontinued Sun workstations?  The
>small percentage of Amigas and Ataris which are equipped with MMUs and at
>least 16MB RAM?  You can't be serious.

I agree the 68K port is no longer of critical importance.  That has nothing
to do with my point, which is that NT is not designed to be as portable as
it could have been.  It is a far more serious issue for SPARCs, and may
affect PowerPC and others as well.  Obviously UN*X has been ported to a lot
of processor architectures that are no longer major players, but that doesn't
mean it's not portable!
---
David C. Niemi  David.Niemi@oasis.gtegsc.com
======================================================
Now I must sit here and ponder the yonder
Herbivores ate well 'cause their food didn't never run



------------------------------

From: niemidc@YP.lab (David C. Niemi)
Subject: Re: Mkimage
Date: 27 Jan 1994 17:13:10 GMT
Reply-To: niemidc@oasis.gtegsc.com

In article js0@vishnu.jussieu.fr, mangon@gla.ecoledoc.ibp.fr (Olivier MANGON) writes:
>Hi every body,
>where i can find the file mkimage for GCC for recompile libc.4.5.8

As Simon Cooper was kind enough to tell me, "mkimage" is part of the shared
library "tools" package under /pub/linux/GCC/src on tsx-11 (it is probably
on other archive sites too; they key is that it is in "tools").
---
David C. Niemi  David.Niemi@oasis.gtegsc.com
======================================================
Now I must sit here and ponder the yonder
Herbivores ate well 'cause their food didn't never run



------------------------------

From: jac@moonshine.llnl.gov (James A. Crotinger)
Subject: Re: Winders NT on MC68K? (was Re: Which is better? 680x0 or 80x86?)
Date: 27 Jan 94 02:22:54 GMT

jerry@msi.com (Jerry Shekhel) writes:
> : Eh?  What about SunOS 4.x.x ?  I figure that it must have run on x86,
> : m68k and sparc, no?  What about SVR4?

> Yep, you're right, I forgot about that.  But is SunOS 4.x.x still a living
> OS?  I thought it had been superceded by Solaris.
> --

  I believe that 4.0 was the last version of the x86 SunOS. Also, I
seem to recollect that the x86 version of SunOS was never really the
same OS as the 68K/sparc. There were many differences, and the two did
not share the same source (though they probably had a common parent).

  What about BSD? What architectectures does it run on? Or SVR5? If
you count Solaris 2 as SVR5, then SVR5 is available at least on MC 68K
and 88K, and on SPARC and x86 (via Solaris).

  BTW, in spite of Suns wishes, there are still MANY sparc platforms
still running SunOS 4.1.x. Previous to 2.3, Solaris was still too
buggy for us to even consider (plus many vendors were slow to port
their software). Now that their new compilers will only be available
on Solaris 2, we're finally considering upgrading.

  Jim


--
=================================================/\==========================
James A. Crotinger     Lawrence Livermore N'Lab // \ The above views are mine
jac@moonshine.llnl.gov P.O. Box 808;  L-630 \\ //---\  and are not neces-
(510) 422-0259         Livermore CA  94550   \\/Amiga\  sarily those of LLNL.

------------------------------

From: alen@myhost.subdomain.domain (Alen Tihi)
Subject: Video7 with XFree86 ?
Date: 27 Jan 1994 16:52:05 GMT

I have a Video7 ( it's pretty old !!!) and want to run Xwindows on my Linux-box.
I couldn't find this card in the support list, so I don't know if this means that
I have to calculate the clock frequencies by myself or  that I have to give up ?

Could anybody help me ?

Thanks

alen (alen@lucky.rz.uni-konstanz.de)






------------------------------

From: niemidc@YP.lab (David C. Niemi)
Subject: Re: Upper Memory Blocks ??
Date: 27 Jan 1994 19:37:00 GMT
Reply-To: niemidc@oasis.lab

Due to quite a few posts on this topic lately, it is clear that there are
quite a few different motherboard behaviors around with respect to upper
memory (640KB through 1MB).

For those who don't know, Linux simply marks the entire upper memory area
as "reserved" and does not try to use any of it as real memory.  This is
safe and reasonable behavior, as there are considerable problems with
actually using memory in that region even when it exists.

It appears there are at least 4 major types of motherboards:

1)      A very few let you tack your upper memory physically onto the
        end of memory.  I.e. the spare 384KB would go immediately after
        your 4MB, 8MB, 16MB, or whatever.

2)      A few (including mine) let you use read/write shadowing of any or
        all of the 384KB via BIOS parameters.  In other words, the memory
        is always there but is usually disabled so that you can instead
        use the ROMs or adapter RAM at those locations.

3)      Like 2) except that there is no easy way to enable read/write memory
        in the upper memory area.

4)      Some motherboards (a minority) seem to have no way of accessing the
        memory physically present in the upper memory area (mostly <= '286es).

I made a kernel patch for machines with 2), and was getting 150 KB of that
memory available, but gave up on it when I found most people couldn't use it
because they have a type 3) motherboard.

Under DOS, there are at least 2 approaches taken by memory managers:

a)      They use the virtual addressing of the 386/486 to make extended (>1MB)
        memory APPEAR to DOS that it is in the gaps of the upper memory area.
        This is quite useless for Linux, as it can use extended memory just
        as well as upper memory already.

b)      They know intimately about the chip set and enable read/write
        access to portions of memory in the upper memory area.  This
        seems to work for both type 2) and 3) motherboards, but it requires
        detailed knowledge of each chip set and sounds rather difficult.

If anyone knows how to do approach b) above under Linux, hey, let me know.
Some people would really enjoy being able to get another 256KB, or even 64KB
without paying for anything extra (though generally I think it's not worth the
trouble).
---
David C. Niemi  David.Niemi@oasis.gtegsc.com
======================================================
Now I must sit here and ponder the yonder
Herbivores ate well 'cause their food didn't never run



------------------------------

From: dlincke@sgcl1.unisg.ch (David-Michael Lincke)
Crossposted-To: comp.os.linux.admin
Subject: CSLIP Server running! PROXY ARP instead of static routes ?????
Date: 27 Jan 94 20:48:51 MET


I have my CSLIP Server running now, the solution isn't very comfortable 
though.

Since the cslip-client is, practically seen, on its own (point-to-point)
network/subnet routing has to be set up on all hosts and the default gateway 
router of the local net (in order to get out on the internet).

After a
         route add client-ipaddr gw cslip-server-ipaddr metric 1

on other machines on the network I had full connectivity to those.

I'd be really glad though to hear that there are alternatives to setting up 
this routing on all the machines on the local net, for I'll have a really 
hard time trying to talk our network administrator into setting up that route 
on the default gateway router, the nameserver machine, the VAXcluster etc....

I've heard that with pl14u PROXY-ARP was incorporated into the kernel.
Now couldn't the linux-CSLIP-server arp for itself as well as for the 
clsip-client? Then all those painful routes dexcribed above wouldn't be 
necessary!

Has anyone out there expoerience with this? 
How do I have to configure procy-arp?

Any help and hints  are highly appreciated.

Thanks a lot,
David
--
David-Michael Lincke
dlincke@sgcl1.unisg.ch, dave@IRC

------------------------------

From: leeh@i-link.com (Lee Heins)
Subject: Re: CD-rom's for Linux
Date: 26 Jan 1994 20:58:02 -0600

 
>I'm thinking of buying a CD-rom for linux.
> 
>I know some installation packages are on CD-rom, but I can't find any
>information.  Can somebody tell me what's available?
> 
>Thanks,
>Byron
> 
>P.S.  I've heard varying opinions on which CD-rom is the best.  I really
>      can't efford fancy stuff, so I was looking into a cheap'o Mitsumi.
> 
>      Should I looking into any other cheap 'standard'?  or what?
> 
>      Basically all I'm looking for is something that will read CD-Roms
>      and will be reasonably supported.  Speed is NOT anything to worry
>      about.  $$$$ is.    (I'm a student)  :)
>--
>PGP 2.3 key available (in plan file) at:        Support public code:
>btf57346@uxa.cso.uiuc.edu &                     Use GNU software
>btf57346@sumter.cso.uiuc.edu                    and others.

Byron,

   I don't know if they still have 'em for what I paid for it...  but I
got an Apple CD150 (Sony 8002 mechanism)...  It is an external SCSI CD
ROM drive.  It seems to work fine with Linux (I have a Yggdrasil LGX CD and
an InfoMagic Linux CD).  I got mine from MacWarehouse for $150 a couple of
weeks ago.  It is only single speed (no big yank for me, as I won't be using
it for heavy duty use and it still beats the pants off getting stuff off
floppies (if nothing else there aren't a zillion disk swaps to install
Linux!) or waiting for a slow FTP and then downloading V.32), and only
single session photo-CD compatible (again, I could care less, as I have no
interest in photo-CD).  I couldn't find a proprietary interface cheapo CD-ROM
drive for less than what I got the CD-150 for...  And I looked around at most
of the mail order suppliers...  cheapest I could find the Mitsumi for was
about $20 more than the CD150, and it would have used up another slot in my
machine, and cost more to make external, as my machine's drive bays are
limited and are already full (1 Teac 3.5-5.25 combo space saver drive,
1 Seagate 320M SCSI drive, and 1 Colorado Jumbo 250 tape fill the three
1/2 ht 5.25" bays)...
   Of course the CD-150 is a lot less attractive if you don't already have
a SCSI interface on your Linux machine (since I run only Linux and only
SCSI hard drives, I already had SCSI) or if you have your heart set on an
interal drive (although you could pull the mechanism from the case, install
it internal and sell the nice SCSI case for $25-$50 which seems to be the
going rate for that type of case, dropping your cost to $100 to $125).
   I also use the CD-150 on my Macintosh, so it does double duty for me (yes,
some of us own more than one computer!  :-) ).

                                        Lee Heins
                                        leeh@i-link.com


------------------------------

From: jerry@msi.com (Jerry Shekhel)
Subject: Re: Winders NT on MC68K?  (was Re: Which is better? 680x0 or 80x86?)
Date: 27 Jan 1994 17:21:36 GMT

David Barr (barr@pop.psu.edu) wrote:
: >
: >Wrong.  NT runs on three different architectures right now.  Linux doesn't.
: >In fact, no single version of Unix does.
:
: Oh come now this is absurd.  Do you realize how many architectures
: BSD 4.3 has been ported to?  I've lost count.
:

OK.  Let's get something straight.  I thought we were talking about OS's
that were completely source-code compatible across several architectures.

Sure, BSD 4.3 has been ported to everything from here to the moon, but the
translation process has left us not with one portable OS, but with a
hundred different BSD-like operating systems.

: >
: >Let's not speculate on Linux's
: >portability (relative to NT's) until Linux is actually ported to as many
: >architectures as NT supports.
:
: Let's not speculate on NT's portability until it is actually ported to
: as many architectures as UNIX supports.  Fair?
:

Exactly which UNIX are you talking about?

: >
: >Your comments are misleading.  True, MS indicated no willingness to port
: >NT to big-endian CPUs.  However, that doesn't mean that NT is as you say
: >"hard-coded to little-endian processors".  There is no evidence either way
: >whatsoever.
:
: Bull.  Why have companies introduced bisexual chips rather than get
: NT fixed?
:

What does that prove?  For the first time hardware and OS companies are
working together to eliminate one nasty source of incompatibility.  This
is a bad thing?

: >
: >MS may simply be avoiding binary file incompatibilities and
: >byte swapping issues.
:
: Bzzt.  UNIX dealt with endian incompatibilities for 20 years.  Why is
: NT crippled?
:

You're still confused.  UNIX is not a portable operating system.  To
convince yourself, just take a look at the "#ifdef" jungle that comprises
a large part of any GNU program.

: >
: >As far as a 68K port is concerned, what would be the point?  A 68K version
: >of NT would run on what, exactly? Old discontinued Sun workstations?  The
: >small percentage of Amigas and Ataris which are equipped with MMUs and at
: >least 16MB RAM?  You can't be serious.
:
: Can you say M-a-c-i-n-t-o-s-h?  Hello!
:

The Mac is no different.  Many models don't have an MMU.

: --Dave
--
+-------------------+----------------------------+---------------------------+
| JERRY J. SHEKHEL  | Molecular Simulations Inc. | Cowboy Junkies, Phish,    |
| Drummers do it... |     Burlington, MA USA     | Tribe, Guns N' Roses,     |
|    ... In rhythm! |        jerry@msi.com       | TAMA, Zildjian, Linux...  |
+-------------------+----------------------------+---------------------------+

------------------------------

From: jerry@msi.com (Jerry Shekhel)
Subject: Re: Winders NT on MC68K?  (was Re: Which is better? 680x0 or 80x86?)
Date: 27 Jan 1994 17:24:31 GMT

Alan Cox (iiitac@swan.pyr) wrote:
: >
: >Your comments are misleading.  True, MS indicated no willingness to port
: >NT to big-endian CPUs.  However, that doesn't mean that NT is as you say
: >"hard-coded to little-endian processors".  There is no evidence either way
: >whatsoever.  MS may simply be avoiding binary file incompatibilities and
: >byte swapping issues.
:
: Snigger.. Unix fixed this 15 years ago too. I would think its a bit more
: serious than a few byte swapping issues somehow. 
:

Stop it!  Unix didn't fix squat.  Try writing some integers out to a binary
file on one machine and then reading them back in on another machine of
different endian-ness.

: Alan
--
+-------------------+----------------------------+---------------------------+
| JERRY J. SHEKHEL  | Molecular Simulations Inc. | Cowboy Junkies, Phish,    |
| Drummers do it... |     Burlington, MA USA     | Tribe, Guns N' Roses,     |
|    ... In rhythm! |        jerry@msi.com       | TAMA, Zildjian, Linux...  |
+-------------------+----------------------------+---------------------------+

------------------------------

From: miguel@roxanne.nuclecu.unam.mx (Miguel de Icaza)
Subject: Re: Winders NT on MC68K?
Date: 27 Jan 1994 22:33:29 GMT


> : Look, Jerry, I agree this is all speculation.  However, simply because NT
> : runs on 3 architectures does not indicate that is as "portable" as any
> : given flavor of UN*X,
> :
> 
> It is *more* portable than the lot of operating systems we call "UNIX",
> since it's designed to be 100% source compatible across all supported
> architectures.  The portability problem is the thing that NT solves for
> developers.

What about Berkeley Unix? It's also source compatible across all
supported architectures. 

If you do system programming, maybe you need to use architecture
dependant code, and you'll have the same problem with NT.

> : several of which have been ported to 3 or more
> : processor families.  BSD, SVR3, and SVR4 have each been ported to over a
> : DOZEN different processors, of both and mixed endiannesses.  Try that with
> : NT!
> :
> 
> All these UNIX systems you're talking about are not 100% source code
> compatible.  You're confusing "being portable" with "having been ported".

Berkeley Unix is compatible with Berkeley Unix, SV3 with SV3, and so
on. 

They are just different flavors, just like Windows 3.0, Windows 3.1,
Windows NT and Win32, different APIs. Or try compiling a Win32 program
with Windows 3.1...

> Sure, there are versions of UNIX available for lots of different platforms,
> but writing a portable UNIX application is a nightmare!  NT promises to
> solve that problem by being 100% portable across all architectures, and if
> that means that one of the endian-nesses has to die, well, IMHO let it die.

They said the same when they delivered OS/2 1.5...

Miguel.

--

Miguel.


------------------------------

From: miguel@roxanne.nuclecu.unam.mx (Miguel de Icaza)
Subject: Re: Winders NT on MC68K?  (was Re: Which is better? 680x0 or 80x86?)
Date: 27 Jan 1994 22:36:13 GMT



> You're still confused.  UNIX is not a portable operating system.  To
> convince yourself, just take a look at the "#ifdef" jungle that comprises
> a large part of any GNU program.

It seems that you have not read any new GNU source code, now they use
autoconf. You just need to unpack, type configure, then make install
and you're done on almost all machines I have.


Miguel.

--

Miguel.


------------------------------


** FOR YOUR REFERENCE **

The service address, to which questions about the list itself and requests
to be added to or deleted from it should be directed, is:

    Internet: Linux-Development-Request@NEWS-DIGESTS.MIT.EDU

You can send mail to the entire list (and comp.os.linux.development) via:

    Internet: Linux-Development@NEWS-DIGESTS.MIT.EDU

Linux may be obtained via one of these FTP sites:
    nic.funet.fi				pub/OS/Linux
    tsx-11.mit.edu				pub/linux
    sunsite.unc.edu				pub/Linux

End of Linux-Development Digest
******************************
