From:     Digestifier <Linux-Activists-Request@news-digests.mit.edu>
To:       Linux-Activists@news-digests.mit.edu
Reply-To: Linux-Activists@news-digests.mit.edu
Date:     Fri, 3 Apr 92 12:00:21 EST
Subject:  Linux-Activists Digest #13

Linux-Activists Digest #13, Volume #2             Fri, 3 Apr 92 12:00:21 EST

Contents:
  Re: What is linux? (Lance Pickup)
  Re: Help, can't compile 0.95a! (Suzuki Shinji)
  hello? (Lawrence C. Foard)
  Help Mr. Fortran (James L Nance)
  Re: HD timeout Errors (with .95a) (Linus Benedict Torvalds)
  Re: Linux and SX, maybe FAQ? (Dragan Cvetkovic)
  Re: What is linux? (William Michael Lye)
  newgcc (DAVE EINSTEIN)
  Selection compile oddities (I Reid)
  OAK SVGA recognition (80x86 assembler question) (I Reid)
  Re: Help Mr. Fortran (Paul Mather)
  Re: Problem with gcc2 (Joseph Knapka)

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

From: lpickup@xanadu.btv.ibm.com (Lance Pickup)
Subject: Re: What is linux?
Date: Fri, 03 Apr 92 13:20:49 GMT
Reply-To: lpickup@vnet.ibm.com

In article <IMHW400.3@INDYVAX.IUPUI.EDU>, IMHW400@INDYVAX.IUPUI.EDU (Mark H. Wood) writes:
|> Is this just because nobody's done it, or LINUX' structure would make it too
|> difficult, or is it a matter of principle :-) ?  Seriously, if anybody is
|> working on MCA mods, I'd like to know.  If not, I may take a stab at it
|> myself, so my poor underemployed PS/2 can have a real operating system, in
|> addition to MeSs-DOS.

I'm not aware of any development on an MCA version (although I would
to hear the contrary myself).  If I can find any free time, I'll be
taking a stab at it, but I won't guarantee my success!  I've gotten
most of the required tools together (notably missing as86 which is
kind of a bummer!) that work on DOS and have at least recompiled the
code successfully.  Once I find a way to get the bootsect and setup
stuff available without as86, maybe I'll be able to actually try
building my own boot disk!

I am optimistic that an MCA version can be done without too much
hassle (i.e. I will be able to port it).  I'll keep people posted
if I make any progress, but keep in mind that I have very little
free time, so progress will be slow.
--
    ...Lance

=======================================================================
Lance Pickup                VNET/IBM Internet: lpickup@btv
IBM/Vendor Systems                   Internet: lpickup@vnet.ibm.com
Technology Products
Burlington, VT                          Phone: (802) 769-7104 (tie 446)
=======================================================================

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

From: suzuki@tkl.iis.u-tokyo.ac.jp (Suzuki Shinji)
Subject: Re: Help, can't compile 0.95a!
Date: 3 Apr 92 10:41:00 GMT

Hi. First of all, my thanks to all of you driving the evolution of Linux.

Seems some have seen trouble recompiling 0.95a kernel.
This is my positive acknowledgement on the matter.

In article <magister.702263905@garfield.catt.ncsu.edu>
magister@garfield.catt.ncsu.edu (Mike Dodds) writes:
 >nicolai@lysator.liu.se (Nicolai Wadstr|m) writes:
 >
 >>I'm unable to compile Linux 0.95a, I have installed the new sources
 >>and bootup with the 0.95a binary kernel image, using the old gcc (1.40?),
 >>...<the rest omitted>
 >
 >Well, 
 >        I can only offer my solution, not any reason why.  It seems
 >that my 4Meg of RAM is insufficient for this mem hog.  So, what
 >I did was to go down the tree stucture manually, finding all the
 >sub-Makefiles, and "make"-ing them.  ... <the rest omitted>

When I tried to rebuild the kernel using gcc1.4, I had to compile
two files w/o -O flag and -Iinclude had to be added to linux/Makefile.
Worse, booting with new kernel got stack-limit exception, halting the
machine. Then I found a comment in linux/kernel/Makefile, which says
#gcc2 doen't have these:
#GCC_OPT = -fcombine-regs
 After switching to gcc2, rebuilding the kernel was a snap.
If I remember, adding '-Iinclude' was everything. I have 16M, though.

 By the way, I have a quick (also dirty) hack on keybord.S to
exchange CTRL and CAPS-LOCK, and to have ALT work as meta key.
If you're interested, just drop me a note. However, I understand
that 'kernel debugger package' has an easily configurable driver.
Using it might be a better idea.

cheers,
-shinji suzuki
I have a .Signature but it's in kanji.

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

Crossposted-To: alt.test
From: entropy@wintermute.WPI.EDU (Lawrence C. Foard)
Subject: hello?
Date: Fri, 3 Apr 1992 07:32:41 GMT

Is comp.os.linux dead everywhere?
-- 
Disclaimer: Opinions are based on logic rather than biblical "fact".   ------
This is a mutated signature virus, if you don't put it in your .sig    \    /
file you may lose your job, your dog may be run over, and you may die.  \  /
If you repent and add the .sig you may win the lottery and get laid.     \/ 

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

From: jlnance@eos.ncsu.edu (James L Nance)
Subject: Help Mr. Fortran
Reply-To: jlnance@eos.ncsu.edu (James L Nance)
Date: Fri, 3 Apr 1992 14:20:14 GMT

I remember reading a post here not too long ago from the person who was writing
gnu FORTRAN.  I was wondering if this person, whose name I can not recall, has
considered porting FORTRAN to Linux.  I would like to see if I can get Spice2g6
ported to Linux, but parts of it (lots of it) are written in FORTRAN and
I think
it would be a Major job to translate them into C.




                                  Thanks,
                                  Jim Nance 

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

From: torvalds@klaava.Helsinki.FI (Linus Benedict Torvalds)
Subject: Re: HD timeout Errors (with .95a)
Date: 3 Apr 92 14:41:29 GMT

In article <1992Apr3.030517.29450@afterlife.ncsc.mil> sdbaker@afterlife.ncsc.mil (Stewart Baker) writes:
>Is there anyone else having problems with their disk in .95a?
> [ description deleted ]

I'm afraid there are people still having problems with the 0.95a
harddisk drivers: the problems show up in "unexpected HD interrupt" and
"HD timeout" messages.  These messages sometimes result in read-errors,
it seems: general protection errors and sometimes even bad filesystems. 

I'll make a new alpha-patch available tomorrow, which has some
corrections to the harddisk driver: they aren't guaranteed to help you,
but I hope the few persons experiencing these errors will try them out
and report to me what happened.

The upcoming alpha-patches will also contain other corrections: the
387-emulation had bugs (corrections were already sent to the gcc-2
beta-testers) and there has been some further work done on the VFS-code
(thanks to entropy and card) as well as some other corrections (hedrick,
bruce evans etc pointed out bugs..)

I'm afraid the patches will be against a clean 0.95a once more: people
with other patches might have problems. The ps-patches are incorporated,
btw.

                Linus

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

From: dragan@coli.uni-sb.de (Dragan Cvetkovic)
Subject: Re: Linux and SX, maybe FAQ?
Date: 3 Apr 92 14:00:36 GMT

dragan@coli.uni-sb.de (Dragan Cvetkovic) writes:

>       Hi! Maybe it is a FAQ, but I am new to this group. So, my question is:
>I am going to get one 386SX/25MHz PC compatible with 2 Mb RAM, and 80 Mb
>hard disk, so I ask: is it possible to run Linux in this enviroment? I just
>get it over FTP, but if I can not use it, I will better not try to.
>               Thanks. D. Cvetkovic
>                       dragan@coli.uni-sb.de
        I got a few answers on my question. Thanks.
                                        Dragan Cvetkovic 

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

From: lye@fraser.sfu.ca (William Michael Lye)
Subject: Re: What is linux?
Date: Fri, 3 Apr 1992 15:18:11 GMT

johnsonm@stolaf.edu (Michael K. Johnson) writes:

>Could someone with these instructions please mail them to me?  I don't
>want to use it right now, but I should have some appendix in the
>manuals that explains how to do this with appropriate warnings.

>(I know how to change the PT, but how to modify the FAT was
>non-obvious when I looked at it yesterday)

>Thanks,

>michaelkjohnson
>johnsonm@stolaf.edu

It's actually quite simple.  When you modified the partition table, you
reduced the number of sectors in the partition in addition to adjusting
the starting & ending cylinder/head/track numbers.  You actually don't have
to modify the FAT.  You just go into the DOS Boot Record, in DOS Logical
Sector 0.  Look at the sector with a hex editor.  The two bytes at offsets
13h and 14h give the total number of sectors (if there are fewer than 64K
sectors).  If you have more than 64K sectors, the 4 bytes at 20h through 23h
give the total number of sectors.  You simply change the appropriate value
to give the same number of sectors as you had in the Partition Table.  Re-
boot, and it should work.  This means that there is some space in the FAT
that is not used (the portion that was for the tail end of the partition)
but if you were to reduce the size of the FAT, you'd have to mess with the
file link-ups in the root directory, and start moving files around, major
FAT surgery.

If you want to be sure, you can use something like NU to write garbage to
the very last logical DOS sector, and then find it again on the disk in 
the absolute cylinder/head/sector to make sure that it is indeed at the end
of the partition.  Just be aware that DOS can leave a small number of un-
used physical sectors at the end of the partition because DOS always works
in terms of an integral number of clusters.




-- 
Bill Lye, lye@sfu.ca
SFU claims these views....NOT!

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

From: EINSTEIN@plh.af.mil (DAVE EINSTEIN)
Subject: newgcc
Reply-To: EINSTEIN@plh.af.mil (DAVE EINSTEIN)
Date: Fri, 3 Apr 1992 16:08:25 GMT


To those installing gcc-1.40 this should be of help. 


        These are instructions for installing gcc-1.40 on Linux-0.95a.
Lunux-0.95a contains everything you need to install and run gcc 1.40
however an editor would certainly be helpful for testing.

You will need to get include-0.12.tar and newgcc.tar to your linux system.
It is probably easiest to uncompess them  and Rawrite them to floopies under
DOS. You should also read newgcc.FAQ.

Now cd to / and untar the include files (include-0.12.tar)
        For example if you have rawritten your uncompressed include-0.12
files to a 5 1/4 inch disk put your floppy in drive a: and type

tar -tvf /dev/at0 | more
   (This shows you what files are in the tar, as well as detecting some errors. 
     If Murphy's laws do not apply to you, you can skip it.)

tar -xvf /dev/at0
   (This should write a bunch of *.h files into /usr/include . If you are 
      paranoid go to /usr/include and look at some of them.)

Now cd to /usr and untar newgcc.tar.
        If you are working with rawritten floppies put in the floppy with
                newgcc.tar and

tar -tvf /dev/at0 | more (for people with healthy paranoia)
tar -xvf /dev/at0

        This will overwrite some of the include files from include-0.12,
don't panic, thats why you untarred the include files first.)

Now gcc looks for ccp and cc1 in /usr/local/lib while they are actually in
/usr/lib so you will need to create symbolic links with

ln -s /usr/lib/cc1 /usr/local/lib
ln -s /usr/lib/cpp /usr/local/lib


Now to test it create the file hello.c

#include <stdio.h>

main()
{
  printf("Hello Linux\n")
}


and compile it by

gcc -o hello hello.c

        If you get a gp-fault here then the problem is probably 'as'
the as that comes with newgcc seems to have problems on some machines,
(it works fine on mine). Try getting the 'as' from gccbin.tar.Z, that
has helped some people with this problem.

        There are still some problems with gcc-1.40. It will occasionally
die during compiles with optimization turned on, this is usually curable
by turning optimization off entirely ( no -O flag), sometimes by adding
-fcombine-regs or some other -f switch.

gcc 2.x:
        There are various versions of gcc 2.x around. The gcc 2.x releases
come with much more extensive include files, libraries etc. However, please
realize that gcc-2.x and the the gnu libc and the new binutils are considered
to be PRE-BETA at the FSF end, and so wierd problems will occurr. In general
I have found gcc-1.40 to be quite sufficient for most uses, but if you really
need 2.x contact hlu@eecs.wsu.edu to get on his list of testers. ( a copy
of gcc-2.x that you happen across on the net is almost gauranteed to be out
of date.)
        Enjoy :-)
                Dave Einstein 

        
 

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

From: eonu24@castle.ed.ac.uk (I Reid)
Subject: Selection compile oddities
Date: 3 Apr 92 12:58:33 GMT

I decided to try the selection patches last night so, having patched
the kernel (0.95a + ps + HD timeout hack) I rebuilt and went off for a
mug of tea. Came back to find a general protection error but a
seemingly ok kernel build (i.e. make completed everything). Logged out
(well, tried to but the machine hung, waited 30 seconds to give it a
chance to sync... it didn't, then rebooted with the NEW kernel, compiled
up selection and all looked good (selection is brilliant.... if you
don't have it you should get it :-)).

I decided to rebuild the kernel to see if I got another general protection
error. Got two this time + a message about page tables (sorry forget the exact
wording). Tried a rebuild with the OLD kernel but the new kernel source- more
protection errors + a message about possibly corrupt memory.

I set up a fresh copy of 0.95a source, compiled it up under the old kernel
without problems. Rebooted with that and all was well. Patched in the ps
stuff, rebuilt and all went well. Rebooted and all was well. Saved all
the working kernel stuff and patched in selection. Compiled up and all
went well, rebuilt and all was well, rebooted and all was well (and is
still running as I write this). The only difference I can see between
what I did originally and what I did this time is that I did a kernel
clean (make clean) before patching the first time but not before the
second. It's all a bit odd anyway 'cos as far as I know my kernel source tree
was exactly the same on both occasions. Any ideas anyone?

Iain

p.s. Selection wanted TIOCLINUX which my /usr/include/termios.h didn't seem to
have so I copied the appropriate line from /usr/src/linux/include/termios.h and
everything seems to be ok. I thought all my include files were up to date but
maybe not.

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

From: eonu24@castle.ed.ac.uk (I Reid)
Subject: OAK SVGA recognition (80x86 assembler question)
Date: 3 Apr 92 12:59:58 GMT

I am trying to get Linux (0.95a) to recognise my SVGA board (Oak 0-67 Video
Bios with 512K RAM) and, after much searching, have finally managed to find
some details on how to recognise it. Unfortunately, my abysmally limited
knowledge of 80x86 assembly language is causing me problems and I need some
help. I would be eternally grateful if someone could give me a hand with this.

Thanks,
      Iain

============================================================
Info: (from vgadoc.zip)

here's a bit of hacked up pascal that correctly recognises my board
p.s I'm not a pascal programmer so I'm working on guesswork :-(

=====
procedure _oak;
begin
  if testreg2($3de,$d,$ff) then
  begin
    name:='OAK 037C';
    if testreg2($3de,$11,$ff) then name:='OAK-067';
  end;
end;

function tstrg(pt,msk:word):boolean;       {Returns true if the bits in MSK
                                            of register PT are read/writable}
var old,nw1,nw2:word;
begin
  old:=port[pt];
  port[pt]:=old and not msk;
  nw1:=port[pt] and msk;
  port[pt]:=old or msk;
  nw2:=port[pt] and msk;
  port[pt]:=old;
  tstrg:=(nw1=0) and (nw2>0);
end;

function testreg2(pt,rg,msk:word):boolean;
var old,nw1,nw2:word;
begin
  port[pt]:=rg;
  testreg2:=tstrg(pt+1,msk);
end;
----

I reckon I'm ok with

mov ax,#0xd                     <--- testreg2
mov dx,0x3de
out dx, ax

inc dx                          <--- testrg
in ax,dx

but I'm not sure how to handle the and not msk bit (and the couple of lines
after it). Which registers can I use? Can I push ax and cx out of the way, use
them as temp space and then pop them back?

The other info I've got is

----
   OAK Technologies OTI-067 and 037C.

   OTI037C apparently has max 256Kb and no banking registers.
   OTI-067 has 256K or 512K and banking registers.

  3DEh index  Dh (R/W):
    bit 7  512K if set, 256 else.

  3DEh index 11h (R/W):
  bit 0-3  Bank no    (Read/Write or two separate banks ??)
      4-7  Bank no
  Note: Some sources gives this register as 3DFh with no index.


  Bank Switching:      (OAK OTI-067 Only)

  Two banks exist in reg 3DEh index 11h.


  ID Oak VGA:

  if exist($3DE,$d) then
    if exist($3DE,$11) then OAK_037
                       else OAK_067 

  Modes:

  (OTI 067)

  4Fh  T   132   60  16 (8x8)
  50h  T   132   25  16 (8x14)
  51h  T   132   43  16 (8x8)   
  52h  G   800  600  16 planar
  53h  G   640  480 256 packed
  54h  G   800  600 256 packed
  55h  G  1024  768   4 ??
  56h  G  1024  768  16 planar
  57h  G   768 1024   4 ??

  (OTI037C "UNIQUE" VGA)

  50h  T   132   25  16 (8x14)
  51h  T   132   43  16 (8x8)
  52h  G   800  600  16 planar


  (OTI037C with NEL Electronics BIOS)

  50h  G   640  480  16 planar
  51h  T    80   30     (16x9)
  52h  G  1024  480  16 planar
  53h  T    80   25     (16x8)
  54h  T   132   43     (8x8)
  55h  T   132   25     (14x8)
  56h  T   132   43     (8x9)
  57h  T   132   25     (14x9)
  58h  T    80   43     (8x8)
  59h  T    80   43     (8x9)
  5Ah  T    80   60     (8x8)
  5Bh  G   800  600  16
  5Ch  T   100   37     (16x8)
  5Dh  T   100   75     (8x8)
  5Eh  G   800  600  16
  6Ah  G   800  600  16
  6Bh  T   100   37     (16x8)

============================================================

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

From: mather@sees.bangor.ac.uk (Paul Mather)
Subject: Re: Help Mr. Fortran
Date: Fri, 3 Apr 1992 16:09:50 GMT

In article <1992Apr3.142014.28253@ncsu.edu>, jlnance@eos.ncsu.edu (James L Nance) writes:
> I remember reading a post here not too long ago from the person who was writing
> gnu FORTRAN.  I was wondering if this person, whose name I can not recall, has
> considered porting FORTRAN to Linux.  I would like to see if I can get Spice2g6
> ported to Linux, but parts of it (lots of it) are written in FORTRAN and
> I think
> it would be a Major job to translate them into C.

Have you tried running the FORTRAN stuff through f2c, the FORTRAN to C
translator?  You never know, it might work.

F2c is (I'm fairly sure) available from the netlib library (on
research.att.com[?]) and on other good archive sites, as they say...

Cheers,

Paul.

-- 
e-mail: p.mather@sees.bangor.ac.uk

If your mailer can't reach me, I'm obviously not worth talking to.

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

From: knapka@athena.cs.uga.edu (Joseph Knapka)
Subject: Re: Problem with gcc2
Date: 3 Apr 92 16:03:01 GMT

In article <1992Apr3.061149.12337@mnemosyne.cs.du.edu> zmbenhal@isis.cs.du.edu (Zeyd M. Ben-Halim) writes:
>schulte@thp.uni-koeln.de (Martin Schulte) writes:
>
>Could somebody post an answer to this question? It has come up at least half
>a dozen time already. (Un)fortunately I never got these error when I use gcc2.
>
>Zeyd
>
>
>>I took gcc2misc.tar.Z and gcc2lib.tar.Z from tsx-11. I untared them
>>in /usr.  /usr/gcc2/bin/gcc of a script which only contains the
>>line "#include <stdio.h>" leads to the error-message:
>>In file incluced from gcc2test.c:1:
>>/usr/gcc2/include/stdio.h:99: parse error before '('
>
>>Any help appreciated, Martin
>>--

I had this problem early on with gcc-2.0. Someone posted about the tar
on the 0.95a distribution disk corrupting files. I switched to another
tar (I think from fileutils.tar.Z on tsx-11, but I'm not sure) and
untarred the gcc-2.x stuff with that, and everything worked after
that. Other people seem to be having problems like this that aren't
fixed by switching to another version of tar, though.

Joseph

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


** 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-Activists-Request@NEWS-DIGESTS.MIT.EDU

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

    Internet: Linux-Activists@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
    tupac-amaru.informatik.rwth-aachen.de	pub/msdos/replace

The current version of Linux is 0.95a released on March 17, 1992

End of Linux-Activists Digest
******************************