From: [email protected] (Barr3y Jaspan)

Subject: Questions about answers in Linux FAQ w.r.t. OS/2
Message-ID: <[email protected]>
Date: 4 Apr 92 22:15:34 GMT
Sender: [email protected] (News system)
Organization: Massachusetts Institute of Technology
Lines: 28
Nntp-Posting-Host: hodge.mit.edu

I have two comments/questions about the Linux FAQ. I have not (before now)
read this newsgroup.

(1) The question "How does Linux operate in an OS/2 environment?" is answered
with "Fine." What exactly does this mean? The most obvious answer is that
OS/2 and Linux can coexist on a disk, and that you can boot either one of them
(probably using OS/2's boot manager).

(2) The answer to "How big is the 'complete' Linux package?" leaves out a lot
of information. The totals given do not appear to contain, say, all the
programs from /bin, /usr/bin, /etc, and so forth that Unix needs to operate.
Is a binary distribution of all this stuff available? If so, how large is it?

Comparing the Linux boot and root disks, kernel source, and libc source to the
entire OS/2 LA distribution just isn't fair. OS/2 contains binary
distributions for all OS/2 stuff, plus a complete DOS and Windows 3.0
distribution, plus a large number of fonts for the Adobe Type Manager,
migration tools for Windows programs, and so forth. The user can choose how
much of this to install, and it usually isn't necessary to use all 20 disks.

Note that I am not claiming that OS/2 is not much larger than Linux. However,
the FAQ appears to be misleading, and that is not a FAQ's job. I'd like to
know the real answer.

Thanks!

--
Barr3y Jaspan, [email protected]


[next article]
From: [email protected] (Karl R. Buck)

Subject: Re: uncompress / OOM errors
Message-ID:
Date: 4 Apr 92 23:29:43 GMT
References: <[email protected]> <[email protected]
du>
Organization: Kansas State University
Lines: 54
NNTP-Posting-Host: matt.ksu.ksu.edu

[email protected] (Joel M. Hoffman) writes:

>More often than not, I'm unable to uncompress .Z files from the net.
>(Yes, I donwload them in binary mode every time.) I suspect this is a
>case of the already-discussed problem with kermit corrupting files.

>At any rate, the real problem is that when I uncompress a bad .Z file,
>uncompress takes all available memory, and then some, and so kills the
>system. Other processes get killed, e.g. The whole idea is that one
>process getting screwed up isn't supposed to interfere with another.

>Will this be fixed soon?

Hello Joel. I have been experiencing the same behavior with kermit.
After much heartache and many corrupt files I tried simply changing
how I was using kermit.

I don't know if this will fix your problems, but this has fixed
*downloads* for me.

Normally when I wanted to download a file or a batch of files I'd run
kermit on the other end and put it in server mode, escape back to the
local Linux system, and do a "get filename". This almost always (after
installing version 0.95) resulted in a corrupt file; that is the sizes
would not even match. Instead, now I just invoke kermit at the remote
system (make sure I set file type binary if needed), type "send
filename", escape back to the linux system, and do a "receive".

Since using this method I've had no trouble at all. Just in case, here
is my .kermrc:

set line /dev/modem
set speed 2400
set modem hayes
set file type binary
set send packet-length 900
set receive packet-length 900
set parity

BTW, those tired of your remote system moving the bottom line to the
middle of the screen in 50 line mode, take a look at where "tset"
(most likely in your .login for Sun and probably for most BSD systems)
and add a "-I" option (tset -I). This will stop it from initializing
the screen. Works for me. 🙂








--
Karl Buck, [email protected] 🙂


[next article]
From: [email protected] (Hongjiu Lu -- Graduate Student)

Subject: Re: gcc2.1, no as (gas)?
Message-ID: <[email protected]>
Date: 4 Apr 92 22:32:22 GMT
References: <[email protected]
.Edu>
Sender: [email protected] (USENET News System)
Organization: Washington State University
Lines: 16

In article <[email protected]> [email protected] (Gr
eg Lee) writes:
>In article [email protected] (Jiansheng Zhao
) writes:
>}I installed gcc2.1 according to FAQ in 2misc.tar. However, I can't compile
>}0.95a kernel with it (no problem for gcc1.4). It asked for as, cpp. ...
>
>I've kept cc, cc1, cpp, as, ld in /bin from gcc 1.40 through 2.1 and
>never had any problems with (g)cc finding what it needs.
>

as and ld are ok. But you should dump cc, cc1 and cpp. BTW, as and ld
are not the part of gcc.

>--
>Greg Lee

H.J.


[next article]
From: [email protected] (Joseph Knapka)

Subject: Gnu emacs
Message-ID: <[email protected]>
Date: 4 Apr 92 23:23:28 GMT
Organization: University of Georgia, Athens
Lines: 13

I notice that ESC-X shell doesn't work under GNU emacs. Does anyone
know why? Are there plans to correct it? I'd also like to be able to
get at SWI-prolog using ESC-X run-prolog. Anyone know how to arrange
such a thing, or even if it's possible? I imagine Prolog would have to
be hacked at some...

Thanks for any info,

Joseph

BTW: Re: my previous compaint about compress: thanks all who answered.
Summary: compress leaves a file alone if it would get larger as a
result of compression.


[next article]
From: [email protected] (Joel M. Hoffman)

Subject: Re: uncompress / OOM errors
Message-ID: <[email protected]>
Date: 5 Apr 92 01:33:26 GMT
References: <[email protected]> <[email protected]
du>
Sender: [email protected] (USENET News system)
Organization: University of Maryland, College Park
Lines: 22
Nntp-Posting-Host: next.wam.umd.edu

In article [email protected] (Karl R. Buck) wr
ites:
>[email protected] (Joel M. Hoffman) writes:
>
>>More often than not, I'm unable to uncompress .Z files from the net.
>>(Yes, I donwload them in binary mode every time.) I suspect this is a
>>case of the already-discussed problem with kermit corrupting files.
>
>>At any rate, the real problem is that when I uncompress a bad .Z file,
>>uncompress takes all available memory, and then some, and so kills the
>>system. Other processes get killed, e.g. The whole idea is that one
>>process getting screwed up isn't supposed to interfere with another.
>
>>Will this be fixed soon?
>
>Hello Joel. I have been experiencing the same behavior with kermit.
>After much heartache and many corrupt files I tried simply changing
>how I was using kermit.

Many people have written with suggestion about how to improve kermit's
behavior. I appreciate the feedback. But I'm more concerned about
the behavior of Linux itself, which starts killing innocent processes
after a broken one uses up all available memory.


[next article]
From: [email protected] (Drew Eckhardt)

Subject: Re: priority levels in Linux?
Message-ID: <[email protected]>
Date: 5 Apr 92 02:05:43 GMT
References: <[email protected]>
Sender: [email protected] (The Daily Planet)
Organization: University of Colorado at Boulder
Lines: 11
Nntp-Posting-Host: romeo.cs.colorado.edu

In article <[email protected]> [email protected] (Joel M. Hoffman
) writes:
>Does Linux support different priority level for different processes
>(as used with nice, renice, etc.)?
>
>-Joel
>([email protected])
>
>

Yes, but as of this time, to my knowledge, no
one has written a working renice utility.


[next article]
From: [email protected] (Drew Eckhardt)

Subject: Re: Questions about answers in Linux FAQ w.r.t. OS/2
Message-ID: <[email protected]>
Date: 5 Apr 92 02:20:25 GMT
References: <[email protected]>
Sender: [email protected] (The Daily Planet)
Organization: University of Colorado at Boulder
Lines: 58
Nntp-Posting-Host: romeo.cs.colorado.edu

In article <[email protected]> [email protected] (Barr3y
Jaspan) writes:
>I have two comments/questions about the Linux FAQ. I have not (before now)
>read this newsgroup.
>
>(1) The question "How does Linux operate in an OS/2 environment?" is answered
>with "Fine." What exactly does this mean? The most obvious answer is that
>OS/2 and Linux can coexist on a disk, and that you can boot either one of them
>(probably using OS/2's boot manager).

Linux will coexist with any other operating system(s) which respects
the "standard" PC partitioning scheme - this includes DOS, OS/2,
Minix, etc.

>(2) The answer to "How big is the 'complete' Linux package?" leaves out a lot
>of information. The totals given do not appear to contain, say, all the
>programs from /bin, /usr/bin, /etc, and so forth that Unix needs to operate.
>Is a binary distribution of all this stuff available? If so, how large is it?
>

Yes.

>Comparing the Linux boot and root disks, kernel source, and libc source to the
>entire OS/2 LA distribution just isn't fair. OS/2 contains binary
>distributions for all OS/2 stuff, plus a complete DOS and Windows 3.0
>distribution, plus a large number of fonts for the Adobe Type Manager,
>migration tools for Windows programs, and so forth. The user can choose how
>much of this to install, and it usually isn't necessary to use all 20 disks.
>

That depends on your definition of complete. Obviously, the OS/2
distribution includes neither a 'C' compiler nor debugger, compiler
construction tools like Bison and Flex, etc. However, a compiler,
assembler, linker, etc and Make are required in a Unix environment
where most software is available in a source distribution only.

For me, a comfortable system is a 32M root partition (with a floating
3-10M free), plus another swap partition. This includes
everything except the full-blown Emacs. That means EVERYTHING,
all of the Gnu file, shell, text, and bin utilities, bison, flex,
bash, shadow password login and utilities, awk, sed, perl, gcc 2.1 / gas
and full libraries / include files, elvis, micro emacs, vile,
ed, full kernel / library source, /vmunix and /vmunix.bak, cron, RCS,
compress, tar, kermit.

Full library and kernel source is there, plus the object code from
the last kernel build.

You don't have the same wealth of utilities under OS/2.


I should also mention that before I patched the kernel so it wouldn't
get stuck on hard errors in the hd driver, so I could make a
file system on the non-error free partition that is now my root,
I was running it on
my 10M swap partition. In this configuration, I still had enough
space for all of the "normal" utilities (no choice of three
editors, perl, online library source, or games) and to do kernel
builds with GCC1.39.


[next article]
From: [email protected] (Joel M. Hoffman)

Subject: gcc can't find ___stderr. Where is it??
Message-ID: <[email protected]>
Date: 5 Apr 92 03:04:05 GMT
Sender: [email protected] (USENET News system)
Organization: University of Maryland, College Park
Lines: 14
Nntp-Posting-Host: next.wam.umd.edu

This is probably some inane omission on my part, but while trying
to compile linux/tools/build.c, I get an error message about an
undefined symbol ___stderr referenced from the text segment.

What am I doing wrong?

On a more general note:
I seem to recall this having been discussed once before in
alt.os.linux, but of course the article has expired, and I have no way
to find the information from the archives. It would be nice if there
were some way to grep the archives while using FTP. Is this possible?

-Joel
([email protected])


[next article]
From: [email protected] (cm445a17)

Subject: MGR --> Anyone???
Message-ID: <[email protected]>
Date: 5 Apr 92 06:57:33 GMT
Organization: University of Maryland University College
Lines: 3

Is anyone working on porting MGR for Linux? Thanks,

Jaime


[next article]
From: [email protected] (Zeyd M. Ben-Halim)

Subject: adduser (again!)
Message-ID: <[email protected]>
Date: 5 Apr 92 06:34:48 GMT
Sender: [email protected] (netnews admin account)
Organization: University of Denver, Dept. of Math & Comp. Sci.
Lines: 10


Sorry to be such a nag, but has anybody managed to use adduser correctly.
From looking at the sources and those for login.c (poeigl .12) they do not
agree on the salt to use in encryption and as a result never generate thae
same result for a given password. From login.c salt is the first 2 char. from
the users name. adduser uses a random salt each time it's envoked.
I changed adduser to use uname[0] and uname[1] but no joy.
So what gives?

Zeyd


[next article]
From: [email protected] (Tor Lillqvist)

Subject: Re: Since I haven't seen an FAQ...
Message-ID:
Date: 5 Apr 92 14:45:41 GMT
References: <[email protected]>

Sender: [email protected]
Organization: Technical Research Centre of Finland, Laboratory for Information
Processing (VTT/TIK)
Lines: 23
In-reply-to: [email protected]'s message of 4 Apr 92 06:42:36 GMT

In article [email protected]
utgers.edu (Charles Hedrick) writes:
The only thing I can think of that would be different about the Gnu
OS when it's released is that it will have a kernel written by Gnu.

Aren't you forgetting that Linux (at least according to the info I
have read here, I haven't looked at the sources yet) is written for
the 386 architecture only, while Mach is presumably more
architecture-independent. On the other hand, I don't know if the
architecture-specific bits of Mach are less or more, line-count-wise,
than those parts of Linux that are 386-specific. The 386 can't show
through on every level of Linux, can it?

For the "hacker" wanting a cheap but powerful system with free UNIX
complete with sources, nowadays the question is probably moot. There
doesn't seem to be much question that a generic 386 or 486 box gives
the best MIPS/$ ratio. I certainly don't have any warm and fuzzy
feeling about the 386 or the AT architecture per se, but am still
beginning to accept that maybe I haven't got much choice when I
finally some day buy a home system... (But certainly, I won't run DOS
on it.) But who knows what happens in five years?
--
Tor Lillqvist,
working, but not speaking, for the Technical Research Centre of Finland


[next article]
From: [email protected] (A. V. Le Blanc)

Subject: Re: Compress badness
Message-ID: <[email protected]>
Date: 4 Apr 92 18:31:29 GMT
References: <[email protected]>
Followup-To: [email protected]
Organization: Computing Centre, University of Manchester
Lines: 17


In article <[email protected]> [email protected] (J
oseph Knapka) writes:
>
>... And then one day, compress went flaky. Works for everything except
>one file, /home/tar/misc.tar, which has within it shoelace.tar, some
>patches for shoelace, and some other odds and ends. I say, "compress
>/home/tar/misc.tar", and compress pounds on the disk a litte while,
>and then back comes the root: prompt --- and misc.tar is totally
>unchanged --- there's not even a .Z on the end of its name.

This is standard behaviour when the 'compressed' file is larger than
the 'uncompressed' file. You can force compression by
compress misc.tar.Z
and then see how big the two files are.

-- Owen
[email protected]


[next article]
From: [email protected] (Andy I. Liaw)

Subject: partitioning problem
Message-ID: <[email protected]>
Date: 5 Apr 92 07:28:00 GMT
Organization: Texas A&M University, College Station
Lines: 12

Hi! everyone,

I finally installed linux on my IDE drive. The strange thing is that
I created both dos and linux partitions with pfdisk, and activate the
dos partition with pfdisk also (if i activate it with dos's fdisk,
linux won't boot up and i have to re-partition and re-install). When
I boot up to dos, I can go to the D: prompt! And it shows nothing in
there. How do I restrict access to the linux partition from DOS? I
am really not crazy about going through the installation for the 20th
time, so I don't want to mess up the linux partition any more.

Andy


[next article]
From: [email protected] (Andy I. Liaw)

Subject: serial port in linux
Message-ID: <[email protected]>
Date: 5 Apr 92 07:30:37 GMT
Organization: Texas A&M University, College Station
Lines: 5

Can someone please tell me how to set line in kermit? My
hayes compatible modem is set to COM1. What is the device
name for it?

andy


[next article]
From: [email protected] (BURNS)

Subject: Re: Compress badness
Message-ID: <[email protected]>
Date: 5 Apr 92 07:44:25 GMT
References: <[email protected]>
Organization: Georgia Institute of Technology
Lines: 15

in article <[email protected]>, [email protected] (
Joseph Knapka) says:

> /home/tar/misc.tar", and compress pounds on the disk a litte while,
> and then back comes the root: prompt --- and misc.tar is totally
> unchanged --- there's not even a .Z on the end of its name. I do a

Compress does that if the compressed file turns out to not be significantly
smaller than the original - perhaps your tar file has a lot of binaries in
it? To test this out, use the -f (force) switch in compress, and see if the
.Z shows up.
--
BURNS,JIM (returned student)
Georgia Institute of Technology, 30178 Georgia Tech Station,
Atlanta Georgia, 30332 | Internet: [email protected]
uucp: ...!{decvax,hplabs,ncar,purdue,rutgers}!gatech!prism!gt0178a


[next article]
From: [email protected] (Heikki Karhunen)

Subject: Re: make and terminal problems
Message-ID:
Date: 5 Apr 92 10:41:59 GMT
References: <[email protected]>
Sender: [email protected] (Uutis Ankka)
Organization: Department of Theoretical Physics, University of Helsinki,
Finland
Lines: 29
In-Reply-To: [email protected]'s message of 3 Apr 92 21: 23:18 GMT

In article <[email protected]> [email protected] (Alex R.N. W
etmore) wrote:
> What should the permissions bits be set to for make? I used to get an error:
> make: setuid: not owner (or something like that), now I get make:
> setgid: whateverr. I have tried setting make g+s but with no luck, any
> ideas.
>
Odd... My make's permissions are as follows:

-rwxr-xr-x 2 root root 11620 Nov 20 05:30 /usr/bin/make*

I have had no difficulties whatsoever with it. I've used it both as root and
as a normal user.

> Also, is it just me or is the terminal emulation having some trouble
> with cursor movement. Also, sometimes I get it in a mode were the
> terminal output all goes to one line, with no way of fixing it that I
> can think of (except rebooting). I have played around with setterm with
> no luck.

I've had my share of these, too. Especially after tinkering with curses. My
solution is to type: stty sane. It helps, provided you have a stty 🙂

>
> alex

/-----------------------------+-----------------------------------------\
| [email protected] | There is always a job for a theoretical |
| [email protected] | physicist -- at least in theory. |
\-----------------------------+-----------------------------------------/


[next article]
From: [email protected] (Joel M. Hoffman)

Subject: mismatched processes. Kernel bug?
Message-ID: <[email protected]>
Date: 5 Apr 92 15:29:52 GMT
Sender: [email protected] (USENET News system)
Organization: University of Maryland, College Park
Lines: 20
Nntp-Posting-Host: next.wam.umd.edu

I was compiling tools/build.c with ``gcc build.c'' and I got something
like the following:

Warning, pid 143 older sibling has mismatched ys
Warning, pid 143 younger sibling has mismatched os
Warning, pid 143 older sibling has mismatched ys
Warning, pid 143 younger sibling has mismatched os

I forget the exact wording. I sync'd, rebooted, and tried it again,
but couldn't repeat the error.

I had a similar problem yesterday, but I ignored it. Several hours
later, the process table seems to have been completely correupted.
All I had was Scroll-Lock information to go by, but there were lots of
processes whose parent process no longer existed. I also tried
killing the processes with ``kill -9'' (logged in as root), but to no
avail. What does one do with illigitamte mis-linked immortal
processes?

-Joel


[next article]
From: [email protected] (Greg Lee)

Subject: 95c write problem
Message-ID: <[email protected]>
Date: 5 Apr 92 17:43:23 GMT
Sender: [email protected] (News Service)
Organization: University of Hawaii
Lines: 17
Nntp-Posting-Host: uhunix.uhcc.hawaii.edu

With the 95c patches and compiled with gcc 2.1, linux
does not display login prompts on the tty2, tty3, tty4,
though my inittab puts getty's on them. I think the
problem is with the write statments in agetty.c. The
following test program is extracted from agetty.c -- it
should display " login: " on the screen, but it doesn't
display anything.
-----------test.c-------------
#define LOGIN " login: "
main()
{
(void) write(1, LOGIN, sizeof(LOGIN) - 1);
}
------------------------------

--
Greg Lee


[next article]
From: [email protected] (Bowen (Laughing Cheetah))

Subject: Re: partitioning problem
Summary: Hiding linux partition from MSDOS
Keywords: MINIX LINUX PARTITIONS
Message-ID: <[email protected]>
Date: 5 Apr 92 18:38:27 GMT
References: <[email protected]>
Sender: [email protected]
Organization: Purdue University Computing Center
Lines: 21

In article <[email protected]> [email protected] (Andy I. Liaw) writes
:
>Hi! everyone,
>
>I finally installed linux on my IDE drive. The strange thing is that
>I created both dos and linux partitions with pfdisk, and activate the
>dos partition with pfdisk also (if i activate it with dos's fdisk,
>linux won't boot up and i have to re-partition and re-install). When
>I boot up to dos, I can go to the D: prompt! And it shows nothing in

You can restrict the 'D:' partition quite easily. When using pfdisk to
partition the drive, set the Linux partition's type ID to Minix. DOS
will ignore this partition, and if you look at it using FDISK, it will
report 'non-DOS' partition. Works rather well.

-bg

--
________________________________________________________________________________
Bowen Goletz
Purdue University Midi 'n Music
[email protected] [email protected]


[next article]
From: [email protected] (Douglas E. Quale)

Subject: Re: priority levels in Linux?
Message-ID: <[email protected]>
Date: 5 Apr 92 18:45:57 GMT
References: <[email protected]> <1992Apr5.020543.15790@colorado.
edu>
Sender: [email protected] (The News)
Organization: Undergraduate Projects Lab, UW Madison
Lines: 13

In article <[email protected]> [email protected] (Drew
Eckhardt) writes:
>
>Yes, but as of this time, to my knowledge, no
>one has written a working renice utility.
>

Renice(1) requires an addition to the kernel. Linux provides nice(2), but
what we really want is BSD's getpriority(2)/setpriority(2) and then renice(1)
is easy and nice(2) can be retired.

--
Doug Quale
[email protected]


[next article]
From: [email protected] (Greg Lee)

Subject: Re: 95c write problem
Message-ID: <[email protected]>
Date: 5 Apr 92 19:19:59 GMT
References: <[email protected]>
Sender: [email protected] (News Service)
Organization: University of Hawaii
Lines: 25
Nntp-Posting-Host: uhunix.uhcc.hawaii.edu

}following test program is extracted from agetty.c -- it
}should display " login: " on the screen, but it doesn't
}display anything.

The test program I gave does display correctly. My mistake.
However, it might still be that the problem is with the write()
call.

Here is another difficulty with linux 95c: the TIOCGWINSZ
ioctl no longer tells the screen size. The following program
returns three 0's:
------------------------
#include
#include
main()
{
struct winsize ws;

printf("ioctl result = %d\n", ioctl(1, TIOCGWINSZ, &ws));
printf("ws.ws_col = %d\n", ws.ws_col);
printf("ws.ws_row = %d\n", ws.ws_row);
}
------------------------
--
Greg Lee


[next article]
From: [email protected] (Steve Sykes)

Subject: Re: priority levels in Linux?
Message-ID: <[email protected]>
Date: 5 Apr 92 21:04:13 GMT
References: <[email protected]> <1992Apr5.020543.15790@colorado.
edu> <[email protected]>
Sender: [email protected] (Network News)
Organization: Department of Computer Science, Warwick University, England
Lines: 15
Nntp-Posting-Host: stone

I've noticed some comment about renice. I have added the getpriority
and setpriority calls to the kernel and they seem to work reasonably
well. I was reluctant to post them because I did notice some odd
behaviour occaisionally....although nothing in the last 2 months.

The only annoying thing about adding set/getpriority was that linux
regards a lower 'NICE' value as lower priority rather than the
other way round.

Still...it worked for me (I think :-)) If you're interested send me
e-mail and I'll consider trying to knock up a patch for it and send it
to Linus for inclusion in the next release.

Laurence Culhane
[email protected]


[next article]
From: [email protected] (Bill Bogstad)

Subject: Where is the source for init/getty/login?
Message-ID: <[email protected]>
Date: 5 Apr 92 20:59:11 GMT
Organization: Department of Cognitive Science, JHU
Lines: 10


I'm trying to do some work on system startup and shutdown scripts
and it would really help to know which source files generated the
init/getty/login files on the 0.95a rootimage. I THINK it might be the
"poeigl 1.2" release, but I'm not sure. Extra karma points go to the person
who knows which compiler/library was used to generate them.

Thanks,
Bill Bogstad
[email protected]



[next article]
From: [email protected] (Joseph Laval Tremblay)

Subject: Re: serial port in linux
Message-ID: <[email protected]>
Date: 5 Apr 92 21:33:23 GMT
References: <[email protected]>
Sender: [email protected]
Organization: Computer Science Department, Texas A&M University
Lines: 16

[email protected] (Andy I. Liaw) writes:
>Can someone please tell me how to set line in kermit? My
>hayes compatible modem is set to COM1. What is the device
>name for it?

I am also having problems with this. By the way default
values are set in kermit and pcomm I am thinking that com1
and com2 are tty64 and tty65 -- but when I try to dial the
machine just sits there and counts down until timeout. I'm
sure many of you have these working -- any guesses what
the problem could be?

Joe

--
Joe Tremblay, [email protected]


[next article]
From: [email protected] (The WOZ)

Subject: documentation for shoelace
Message-ID: <[email protected]>
Date: 5 Apr 92 20:55:24 GMT
Organization: California Polytechnic State University, San Luis Obispo
Lines: 40

I want to boot linux off my hard drive. I understand I need
shoelace to do this (someone correct me if I'm wrong).

So I went off to tsx-11 and grabbed it, uncompressed and untared it,
and attempted to read the documentation. It appears as if the docs
haven't been updated since the port from minix, and it is mostly a
'by route' type of document (i.e. it says type this, do that, etc...
without really telling WHAT you're doing or why it's necessary). It
also mentions files which simply do not exist in the 0.95a release
(maybe I need to ftp more stuff?).

Maybe I'm just stupid, but I like to know what the final product is
supposed to look like before I start putting it together. Can
someone post or email me the details of how shoelace works?

Some more specific questions:
1) does shoelace _need_ to be in the root directory? Can I remove
it later?
2) The readme and config files refer to files in /etc/system (mm,
kernal, fs, and init) The 0.95a dist does not have a /etc/system Do I
need to get these files from somewhere? (I'm guessing the answer
is no, and this is a minixism, or alternately that the files
_are_ already there somewhere and I just missed them...)
3) What is the file bootlace used for? It's not mentioned in the
readme file.
4) What is the file /usr/src/linux/linux/Image ? (Please tell me I'm
not going to have to recompile the kernal or something...)
5) When is the file disktab used?
6) How exactly does winiboot work? What does it do?

Sorry if I seem a little clueless about all this (gotta learn
somehow I guess 🙂

Thanks in advance.
--Adam
--
You don't understand, Doran. | The Lord Knight
Bugs do not happen because you do something. | Errant Grammer
Bugs just are. | Checker of the BEM
-Explorer_Bob to me, [email protected] \ and proud of it!


[next article]
From: [email protected] (Andy I. Liaw)

Subject: Re: serial port in linux
Message-ID: <[email protected]>
Date: 5 Apr 92 21:52:43 GMT
References: <[email protected]> <[email protected]>
Organization: Texas A&M University, College Station
Lines: 15

In article <[email protected]> [email protected] (Joseph Laval Tremblay)
writes:
>>Can someone please tell me how to set line in kermit? My
>>hayes compatible modem is set to COM1. What is the device
>>name for it?
>
>I am also having problems with this. By the way default
>values are set in kermit and pcomm I am thinking that com1
>and com2 are tty64 and tty65 -- but when I try to dial the
>machine just sits there and counts down until timeout

I can set line /dev/ttys1, which is the device name for com1,
thanks to those reply. But when I dial there's no reaction
from the modem. What's wrong?

Andy


[next article]
From: [email protected] (Andy I. Liaw)

Subject: index or descriptions for files?
Message-ID: <[email protected]>
Date: 5 Apr 92 21:56:49 GMT
Organization: Texas A&M University, College Station
Lines: 7

I went on to tsx-11.mit.edu and got VERY confused by the files.
Is there an index or maybe a file-description somewhere? If
not, can we have one?

Thanx!

Andy


[next article]
From: [email protected] (david nugent)

Subject: config-95a
Message-ID:
Date: 5 Apr 92 11:07:10 GMT
Reply-To: [email protected]
Organization: Unique Computing Pty Ltd, Melbourne, Australia
Lines: 15

Did anyone have problems with this? After applying it and successfully
building the kernel, it apparently broke the serial ports. I get
a general protection interrupt when running kermit, for example.
Everything else seems to work ok though.

I reversed out the diffs for now, but I support the idea of having the
standard device driver interface that the package offers.

david

..............................................................................
david nugent Public Access Usenet "Only Nixon can go to China"
[email protected] +61-3-792-3507 - ancient Vulcan proverb
3:632/400@fidonet, 58:4100/1@intlnet, 199:4242/5@rainbownet, 33:300/6@admin
PO Box 260, Endeavour Hills, Victoria, Australia, 3802.


[next article]
From: [email protected] (Alex R.N. Wetmore)

Subject: bash question
Message-ID: <[email protected]>
Date: 5 Apr 92 22:45:06 GMT
Organization: Freshman, Electrical and Computer Engineering, Carnegie Mellon, Pi
ttsburgh, PA
Lines: 11

I got bash.Z from funet in the /bin area, and noticed that it was really
buggy. It seemed to crash with about every 4th command I typed (usaully
resulting in a general protection fault). Is there another one that has
had better luck being compiled? Also, has anyone considered porting csh
to linux?

Also, is there a site that carries the 0.95a kernel with the ps
extensions? I just don't have the time to go through kernel rebuilds
right now.

alex


[next article]
From: [email protected] (The News Guru)

Subject: Re: serial port in linux
Message-ID:
Date: 5 Apr 92 23:58:31 GMT
References: <[email protected]> <[email protected]> <[email protected]
du>
Organization: Kansas State University
Lines: 13
NNTP-Posting-Host: matt.ksu.ksu.edu

[email protected] (Andy I. Liaw) writes:

>In article <[email protected]> [email protected] (Joseph Laval Tremblay)
writes:
[...]

>I can set line /dev/ttys1, which is the device name for com1,
>thanks to those reply. But when I dial there's no reaction
>from the modem. What's wrong?

I had a similar problem. I booted dos, (yech) ran telix (which worked
fine) and then booted linux and kermit again and it worked. I don'nt
know what the real cause of the problem was but you might give this a
try.


[next article]
From: [email protected] (Gary Bjerke)

Subject: ?? need advice on linux/486 compatibility
Message-ID: <[email protected]>
Date: 5 Apr 92 23:25:21 GMT
Sender: [email protected]
Reply-To: [email protected] (Gary Bjerke)
Distribution: usa
Organization: Tandem Computers; Austin TX
Lines: 23
Originator: garyb@anasazi

I'm about to buy, or perhaps put together, a 486DX system. I'd like to get
an EISA motherboard and SCSI drives. I'd like to run both DOS and Unix, but
I don't want my hardware choices to be limited by what linux and bsd386
currently support. I'd appreciate any advice from gurus out there concerning
the following questions:

1. Is the EISA problem merely one of driver support for 32-bit buswidth, or
are there other problems? Is EISA support expected in the near future
(say, next 6-9 months)?
2. If the problem is just in drivers, can I work around it by tempoarily using
a 16-bit ESDI or IDE controller in the EISA bus?
3. Someone mentioned timing problems with support for some 486 motherboards.
Can anyone elaborate?
4. I plan to buy a 14" or 15" 1024x768 monitor. Will there be problems sup-
porting such a monitor? I am considering the Orchid Fahrenheit 1280
graphics card.
5. Someone posted a SCSI driver for bsd386. I presume it was for an ISA bus.
At the risk of repeating question 1 - is EISA/SCSI support expected soon,
or are there major problems?

All advice and suggestions will be greatly appreciated.

-- Gary Bjerke


[next article]
From: [email protected]
Newsgroups: comp.os.linux,alt.os.linux

Subject: compress/tar problems on SCSI installation
Message-ID: <[email protected]>
Date: 16 Apr 92 18:16:15 GMT
Organization: Penn State University
Lines: 18

I seem to have a nasty little problem on my hands here. I'm trying to
install Linux on the second partition of my SCSI drive ( ST01 controller ).
So I FTP'ed the patched 0.95c+ boot image. This part works fine, it reports
finding the SCSI drive at ID=0 and that it fails read capacity and
everything else. My problem is that I need the SCSI fdisk to do the
installation and it is a tar file. I know to rawrite it onto floppy and how
set up the proper dev. I have two root images, 0.12 and 0.95a, and on both
of them tar is compressed and I keep getting a "ENOSPC" error when I try to
uncompress tar.Z .

What I really need to know is - what does ENOSPC mean and how do I get
around it?

--
Steve Bairstow

[email protected]
or [email protected] (NeXT mail ok)


[next article]
From: chojnows@osiris (Brian Chojnowski)

Subject: Re: ESDI drives
Message-ID: <[email protected]>
Date: 16 Apr 92 18:46:32 GMT
References: <[email protected].
uiuc.edu> <[email protected]>
Sender: [email protected] (News)
Organization: University of Illinois at Urbana
Lines: 5

Duhh, I just assumed that since the doc didnt specify to mkdir /user, that
mount would. That solved the problem. Thank You.

Brian Chojnowski
[email protected]


[next article]
From: [email protected]
Newsgroups: comp.os.linux,alt.os.linux

Subject: Re: compress/tar problems on SCSI installation
Message-ID: <[email protected]>
Date: 16 Apr 92 19:21:40 GMT
References: <[email protected]>
Organization: Penn State University
Lines: 2

I just had a flash of insight here. "ENOSPC" means error no space left on devic
e, in other words the disk is full. But I still need to get around this.


[next article]
From: [email protected] (Sergio Slobodrian)
Newsgroups: comp.os.linux,alt.os.linux

Subject: Re: Keyboard dies when I hit Caps-Lock (+SCSI)
Message-ID: <[email protected]>
Date: 16 Apr 92 20:13:55 GMT
References: <[email protected]> <[email protected]
ku.dk> <[email protected]>
Sender: [email protected] (usenet)
Organization: Bell-Northern Research, Ottawa, Canada
Lines: 22
Nntp-Posting-Host: bcars282

In article <[email protected]> [email protected] (Niels Skov Ols
en) writes:
>I wrote:
>
>>Following keys gives TOTAL lock up but with Linux stil running:
>> Pause
>> Scroll Lock (gives a task-info)
>> Caps Lock
>> Num Lock
>
>Clarification: Everything keep running but it seems as if the
>wiring of the keyboard interrupt to the service routine is
>being broken after one of these keys are hit.
>
I have the same problem with the keyboard, but I can clarify it further.
Once the keyboard locks up on my machine, even a hardware reset leaves
it locked (hitting the reset button, not turning on and off the PC)!! Only
when the bios gets around to initializing the keyboard does it start
to behave once again. The problem with the Caps Lock key is that on my
keyboard it is exactly where the key is on my SPARC at work
and that makes for very frusterating vi sessions.

Does any one have any ideas?


[next article]
From: [email protected] (Hongjiu Lu -- Graduate Student)
Newsgroups: alt.os.linux,comp.os.linux

Subject: gcc 2.1 beta is available
Message-ID: <[email protected]>
Date: 6 Apr 92 05:17:24 GMT
Sender: [email protected] (USENET News System)
Organization: Washington State University
Lines: 214
Originator: [email protected]

Hi,

I am putting gcc 2.1 beta on headrest.woz.colorado.edu under
/pub/linux/gcc and banjo.concert.net under /pub/Linux/GCC. They should
be there in a few hours. Drew, could you please move 2.1lib.tar.Z,
2.1misc.tar.Z, libc-2.1.tar.Z and 2.1shared-A.tar.Z to /pub/linux/gcc?

Although gcc 2.1 is beta, it is quite stable. During the alpha testing
stage, we fixed lots of bugs. It should run well for most of you.
But I cannot guarantee it has no bugs. If you find any bug, please
send it to me at "[email protected]" as well as the channel "GCC" on the
Linux mailing list.

THANKS TO ALL THE ALPHA TESTERS. You did a great job and help me a lot,
Without you guys, I cannot image how I could find all those bugs.

I am enclosing the FAQ for gcc 2.1. PLEASE DO READ IT. It may have the
answers for your questions. You must follow the instructions in the FAQ.

I also put a patch, 2.1kernel.tar.Z, for kernel to compile with
gcc 2.1. You have to apply Linus' second patch first. And another
package, 2.1ps.tar.Z, for ps is modified by me to gcc 2.1.

BTW, make sure tar and compress you have are working right. Some
problems for the previous gcc 2.1 seemed to be caused by bad
tar/compress.

H.J.
--
School of EECS Internet: [email protected]
Washington State University BITNET: [email protected]
Pullman, WA 99164 Phone: (509) 335-6470 (O)
USA (509) 334-6315 (H)
---------FAQ-------
QUESTION: Where is the gcc 2.1 beta?

ANSWER: It's on banjo.concert.net under /pub/Linux/GCC and
headrest.woz.colorado.edu under /pub/linux/?. You may find it on the
other sites. But I am not sure if they are up to date. There are two
files, 2.1lib.tar.Z and 2.1misc.tar.Z. Another file, binutils.tar.Z,
is not the part of gcc 2.1, but used by gcc 2.1. It contains all the
binary utilities you need to run gcc 2.1.

There may be also some files called 2.1shared-x.tar.Z. They have some
programs compiled in .a format. Please take a look at

/usr/shared/obj/README*

after unpack 2.1shared-A.tar.Z under /usr.

QUESTION: What are the contents of them?

ANSWER: 2.1lib.tar.Z contains cc1, cc1plus and cpp. 2.1misc.tar.Z
contains gcc 2.1 drivers, header files, libraries, manual pages and
installation instructions.

QUESTION: How do I install it?

ANSWER: First of all, backup the old compiler if you have one and
remove it from your hard drive. YOU MUST BE ABSOLUTELY SURE THERE IS NO
ANY OTHER C COMPILER ON YOUR SYSTEM INSTALLED. I heard too much stories
about it. By compiler, I mean the executables, libraries (crt0.o) and
header files. This gcc 2.1 will only work with the kernel above 0.95.
You may have to have 2 MB ram to run it and have some swap space
(> 4MB) since cc1 and cc1plus are about 1 MB. First do

cd /usr
tar xvpfvz xxxx/2.1misc.tar.Z

Then read FAQ in /usr/install/gcc2. After

cd /usr
tar xvpfvz xxxx/2.1lib.tar.Z

run the shell script "inst2.x". If you need binutils.tar.Z, do a

cd /usr
tar xvpfvz xxxx/binutils.tar.Z

Please check the file permission for all files you have unpacked. Maybe
I should write a shell script to do that.

QUESTION: How do I install a new release of gcc 2.1?

ANSWER: Depends on the nature of the new release, if the gcc itself is
changed, you have to install everything in 2.1lib.tar.Z and
2.1misc.tar.Z come with the new release. If just the libraries are
changed, you can just pick 2.1update.tar.Z and unpack it in /usr. Then
do

cd /usr
tar xvpfvz xxxx/2.1update.tar.Z
cd /usr/install/gcc2
update2.x

In both cases, it is recommended to relink all the shared executables.
If the gcc itself is changed, a new set of .a's may be released too.

QUESTION: What are the main differences with the old release?

ANSWER: Here are the features

1. gcc and g++ are in one.

2. ANSI and BSD 4.4 compatible stdio.

3. iostream for g++. It has

* Full implementation of ANSI C stdio.
* Full implementation of the streambuf layer of AT&T's
iostream library for C++.
* Full compatibility between stdio and streambufs.

and libg++.a 2.0.

4. Libc.a is almost new. Most of glibc.a 1.02 (nonnetwork stuffs) are
here.

a. ctype
b. grp
c. locale
d. malloc (libmcheck.a)
e. posix
f. pwd
g. stdlib
h. string

Some other functions are added.

b. cvt
b. drand48
c. getlogin (poeigl-1.2)
d. getpass (poeigl-1.2)
e. mkfifo (is that supported in kernel?)
f. regex (GNU)
g. curses (BSD)
h. swab
i. time (BSD, untested)
j. ufc
k. utmp stuffs (poeigl-1.2)

These old functions are not changed very much.

a. soft math (fmod is added)
b. math (some are new)
c. termcap
d. unistd

5. The shared libraries are here. To link with the static libraries,
add the -static flag to gcc.

QUESTION: Is stdio ANSI compatible?

ANSWER: Yes, please test it.

QUESTION: Is g++ in 2.1?

ANSWER: Yes. Use g++ to compile the C++ code. Libg++.a 2.0 is here.
One thing missing is curses. I will take a look if I have time. And
fix.o failed to compile.

QUESTION: What options can I use for gcc?

ANSWER: Read manual page, gcc.ps or gcc.man. -O3, -O4, -O5 or even
higher should work with 2.1. -O3 is said to be the best. Don't ask me
why. Try it yourself. Also -static tells gcc to use the static
libraries. The default is the shared libraries.

QUESTION: Where is the source code of the new libc.a?

ANSWER: The same place you find this file. It is called libc-2.1.tar.Z.

QUESTION: How can I use gcc 2.1 to compile the kernel?

ANSWER: The kernel has to be compiled with the header files come with
it. I will ask Linus to make it easy for you. Please keep in mind,
kernel uses the different header files. Do not include any header files
from gcc 2.1. It can be done by using

cc -nostdinc -I....

QUESTION: Why does gcc say "foo.h not found" when compiling the
kernel and I am following the instruction above?

ANSWER: Not every source code under ./linux is kernel code. By kernel,
I mean those linked together and written to boot image. If there is a
file which is compiled to run under Linux, that is an application and
you should follow the instruction below.

QUESTION: How can I use gcc 2.1 to compile applications?

ANSWER: Any application has to be compiled with the header files come
with gcc 2.1. If you need some header files from the kernel sources,
do a

#include "/foo/src/linux/include/linux/bar.h"

It is strongly recommended to prototype all the library functions by
including the appropriate head files. At least, it will save you
trouble of malloc (0).

QUESTION: Why does a program compiled ok by gcc 2.1 and the new binary
utilities get an error from kernel saying something about binary file
format?

ANSWER: Linux has changed a.out.h. The new a.out.h is compatible with
the applications compiled with the old a.out.h. But if an application is
compiled with the new a.out.h, you have to use the new a.out.h. If you
get this error message, you should get the new a.out.h and old ld from
where you get this and use them to recompile the kernel.

QUESTION: Why does g++ complain?

ANSWER: You need "expr". It is in GNU shell utilities 1.6.


[next article]
From: [email protected] (Hongjiu Lu -- Graduate Student)
Newsgroups: alt.os.linux,comp.os.linux

Subject: diff for gcc 2.1 beta
Message-ID: <[email protected]>
Date: 6 Apr 92 08:06:43 GMT
Sender: [email protected] (USENET News System)
Organization: Washington State University
Lines: 29
Originator: [email protected]

Hi,

I was just informed I should make some change to the libraries. Here
is a diff for libc.a 2.1. If you can apply this diff, please do get a
new 2.1misc.tar.Z.

BTW, please wait for a new hours. I am uploading them now.

H.J.
------gcc 2.1 diff-----
*** compiler/lib/time/Makefile.old Mon Apr 6 02:03:35 1992
--- compiler/lib/time/Makefile Mon Apr 6 01:19:33 1992
***************
*** 8,14 ****
AS =/usr2/linux/cross/bin/as
LD =/usr2/linux/cross/bin/ld
LDFLAGS =-s -x
! TIME_FLAGS=-DTZ='"/usr/lib/zoneinfo"'
CC =/usr2/linux/cross/bin/gcc -B/usr2/linux/cross/lib/gcc-
CC =/local/bin/gcc-i386-linux -DPRE_GCC_2
INC =-nostdinc -I. -I/usr2/linux/src/libc/include \
--- 8,14 ----
AS =/usr2/linux/cross/bin/as
LD =/usr2/linux/cross/bin/ld
LDFLAGS =-s -x
! TIME_FLAGS=-DTZDIR='"/usr/lib/zoneinfo"'
CC =/usr2/linux/cross/bin/gcc -B/usr2/linux/cross/lib/gcc-
CC =/local/bin/gcc-i386-linux -DPRE_GCC_2
INC =-nostdinc -I. -I/usr2/linux/src/libc/include \


[next article]
From: [email protected] (David Engel)
Newsgroups: alt.os.linux,comp.os.linux

Subject: Re: gcc 2.1 beta is available
Message-ID: <[email protected]>
Date: 6 Apr 92 18:19:58 GMT
References: <[email protected]>
Organization: Optical Data Systems, Inc.
Lines: 12

[email protected] (Hongjiu Lu -- Graduate Student) writes:
: THANKS TO ALL THE ALPHA TESTERS. You did a great job and help me a lot,
: Without you guys, I cannot image how I could find all those bugs.

It took longer that I think any of us expected, but we now have one heck
of a compiler and library. Thanks H.J.

-David
--
David Engel Optical Data Systems, Inc.
[email protected] 1101 E. Arapaho Road
(214) 234-6400 Richardson, TX 75081


[next article]
From: [email protected] (Alan B Clegg)
Newsgroups: alt.os.linux,comp.os.linux

Subject: ABC Release Soon! (Was: Re: gcc 2.1 beta is available)
Message-ID: <[email protected]>
Date: 6 Apr 92 19:21:56 GMT
References: <[email protected]> <1992Apr6.181958.22339@ods
.com>
Sender: [email protected]
Organization: Concert Network -- Internet Operations Group
Lines: 23

In article <[email protected]> [email protected] (David Engel) writes:
>[email protected] (Hongjiu Lu -- Graduate Student) writes:
>: THANKS TO ALL THE ALPHA TESTERS. You did a great job and help me a lot,
>: Without you guys, I cannot image how I could find all those bugs.
>
>It took longer that I think any of us expected, but we now have one heck
>of a compiler and library. Thanks H.J.

I second this emotion (8-) and would like to ask the following question:

HAS ANYONE FOUND ANY PROBLEMS WITH THE CURRENT VERSION OF 2.1(beta)?

If not, I am going to use it as the base for the ABC Release of Linux .95a
which I hope to have available by this Friday.

Thanks to those of you that submitted code to me for the release!

-abc
BTW: Anyone at the University of Hawaii around reading this? If so, send me
some mail ASAP.
--
[email protected] Alan Clegg - Network Programmer
KD4JML (just my luck!) MCNC -- Center for Communications


[next article]
From: [email protected] (Hongjiu Lu -- Graduate Student)
Newsgroups: alt.os.linux,comp.os.linux

Subject: diff of FAQ for gcc 2.1 (beta)
Message-ID: <[email protected]>
Date: 6 Apr 92 20:16:31 GMT
Sender: [email protected] (USENET News System)
Organization: Washington State University
Lines: 34
Originator: [email protected]

Hi,

Here is the diff of FAQ for gcc 2.1 (beta). Please use this one.

H.J.
----------------
*** FAQ.old Sun Apr 5 21:48:02 1992
--- FAQ Mon Apr 6 11:59:59 1992
***************
*** 169,179 ****
utilities get an error from kernel saying something about binary file
format?

! ANSWER: Linux has changed a.out.h. The new a.out.h is compatible with
! the applications compiled with the old a.out.h. But if an application is
! compiled with the new a.out.h, you have to use the new a.out.h. If you
! get this error message, you should get the new a.out.h and old ld from
! where you get this and use them to recompile the kernel.

QUESTION: Why does g++ complain?

--- 169,180 ----
utilities get an error from kernel saying something about binary file
format?

! ANSWER: Linux has changed a.out.h. The new a.out.h in Linus' second
! patch is compatible with the applications compiled with the old
! a.out.h. But if an application is compiled with the new ld which uses
! the new a.out.h, you have to use the kernel compiled with the new
! a.out.h. You have to recompile the kernel with Linus' second patch
! applied and old ld. And then use new ld to compile the programs.

QUESTION: Why does g++ complain?



[next article]
From: [email protected] (Hongjiu Lu -- Graduate Student)
Newsgroups: alt.os.linux,comp.os.linux

Subject: IMPORTANT: gcc 2.1, libc.a 2.1a and Linux
Message-ID: <[email protected]>
Date: 6 Apr 92 21:15:00 GMT
Sender: [email protected] (USENET News System)
Organization: Washington State University
Lines: 21
Originator: [email protected]

Hi,

I heard there were some problems with binaries linked with my new ld.
That is caused by the new a.out.h introduced with Linus's second patch.
The new a.out.h is compatible with the old one, But not vice versa.

My new ld (binutils.tar.Z) uses the new a.out.h. The binaries
linked by it can only run under the kernel with Linus's second patch,
which has to be built by the old ld.

Any programs using the old a.out.h, e.g., gdb, will not recognize the
binaries linked by the new ld. You can either use the old ld or use
the new version compiled with the new a.out.h.

From Linus's second patch, we are moving to VFS. As a result, there
will be some changes in libc.a. That means next release of gcc 2.x and
libs will not run under the kernel below Linus's second patch to 0.95a.
I strongly recommend you move to 0.95a with Linus's second patch if you
haven't done so.

H.J.


[next article]
From: [email protected] (Ed Carp)
Newsgroups: alt.os.linux,comp.os.linux

Subject: Taylor-uucp 1.03 beta - porting status
Message-ID: <[email protected]>
Date: 7 Apr 92 01:59:19 GMT
Organization: Bleeding Edge Technologies - 'Sliding Down The Razor Blade Of Life
...'
Lines: 32

This is a followup to a previously posted message.

I've fixed the problems with uuxqt and uux "memory fault".
Now all that remains is to fix the debug problem with
uucico. And test the whole thing, of course! 🙂

Note: I've configured the package for a BNU-style (HDB)
uucp, since that's what seems to be prevalent. Others
will have to reconfigure by hand.

I'll be uploading the (hacked) sources and binaries
to tsx-11 either this evening or tomorrow. The binaries
(with H. J.'s gcc 2.1 shared library stuff) are about
200K in a tar file - probably a bit less compressed.
The sources are about 600K compressed - this includes
all the docs, as well.

I've also included a sample Systems and Devices file,
for those who just can't wait (like me!) 🙂

When Ian makes the diffs to the real 1.03 release available,
I'll be releasing the whole thing over again.

A big THANK YOU goes to H. J. for his work in getting
gcc 2.1 to run! Thanks!!

Stay tuned...
--
Ed Carp N7EKG/6 [email protected] [email protected]
Cupertino, CA 408/252-5947

-- Absolutely unabashed Gates McFadden groupie! --


[next article]
From: [email protected] (Will Clark)
Newsgroups: alt.os.linux,comp.os.linux

Subject: Re: IMPORTANT: gcc 2.1, libc.a 2.1a and Linux
Message-ID: <[email protected]>
Date: 7 Apr 92 17:56:40 GMT
References: <[email protected]>
Sender: [email protected] (News Service)
Organization: Stanford University
Lines: 37

In article <[email protected]> [email protected] (Hong
jiu Lu -- Graduate Student) writes:
>Hi,
>
>I heard there were some problems with binaries linked with my new ld.
>That is caused by the new a.out.h introduced with Linus's second patch.
>The new a.out.h is compatible with the old one, But not vice versa.
>
>My new ld (binutils.tar.Z) uses the new a.out.h. The binaries
>linked by it can only run under the kernel with Linus's second patch,
>which has to be built by the old ld.
>
>Any programs using the old a.out.h, e.g., gdb, will not recognize the
>binaries linked by the new ld. You can either use the old ld or use
>the new version compiled with the new a.out.h.
>
>From Linus's second patch, we are moving to VFS. As a result, there
>will be some changes in libc.a. That means next release of gcc 2.x and
>libs will not run under the kernel below Linus's second patch to 0.95a.
>I strongly recommend you move to 0.95a with Linus's second patch if you
>haven't done so.
>
>H.J.


A plea from the C-impaired: (Am I the only one using Linux?)

Someone please put a 0.95c boot image somewhere for FTP.

I am trying to learn enough about C to feel at least a little
comfortable with kernel patching. I'd like to use gcc 2.1, but to do
that I have to patch the kernel ...

Will Clark (Baseball season is here!)
[email protected]
------------------------
SAIC does not share my opinions.
harlick@lindy is a friend's account, but you can send me email here, too.


[next article]
From: [email protected] (Haohui Wang)
Newsgroups: alt.os.linux,comp.os.linux

Subject: Re: problem compiling kernel...
Message-ID: <[email protected]>
Date: 7 Apr 92 18:15:41 GMT
References: <[email protected]>
Sender: [email protected]
Organization: UB
Lines: 15
Nntp-Posting-Host: lictor.acsu.buffalo.edu

In article <[email protected]> [email protected] (gary a moyer) writes:
>Has anyone gotten the following error when compiling
>the kernel:
>(compiling fork.c..)
>caught fatal signal 6 (??!!?)
>
>Has anyone else had this happen on the way to a kernel build ?
>
> Thx. [email protected]

Same here. Can anyone who compiled kernel post a summary or something?

thanks

howie


[next article]
From: [email protected] (Joe Smith)
Newsgroups: alt.os.linux,comp.os.linux

Subject: Re: Problem with tar (?)
Message-ID: <[email protected]>
Date: 7 Apr 92 19:36:29 GMT
References: <[email protected]>
Sender: [email protected]
Followup-To: alt.os.linux
Organization: University of Pennsylvania
Lines: 28
Nntp-Posting-Host: pobox.upenn.edu

In article <[email protected]> [email protected] (Rob
Coleman) writes:
>
>I am apparently one of the small number of people who are having trouble
>with using tar (any version that I can find) to extract certain files
>from their archives uncorrupted.

Well I'm another whose found tar (and/or tar+compress) unreliable.
I'm using the pair from the binutils archive, but I've tried the ones
on the 0.95a root disk as well. I've encountered sporadic failures
both on extracting and creating tar archives. It also refuses to
create multi-volume compressed archives (although it will do
non-compressed ones).

In one case, tar refused (it just stopped, but would respond to ^C) to
create an archive from a directory where the files had bizarre group
ids (some 5-digit number I don't recall) until I changed the group on
all the files to root.

I'm sorry I don't have time to work on this, because a reliable,
standard archiver/backup utility is a very high priority item.


--
Joe Smith
University of Pennsylvania [email protected]
Dept. of Biochemistry and Biophysics (215) 898-8348
Philadelphia, PA 19104-6059


[next article]
From: [email protected] (Hongjiu Lu -- Graduate Student)
Newsgroups: alt.os.linux,comp.os.linux

Subject: FAQ: gcc 2.1 and kernel
Message-ID: <[email protected]>
Date: 7 Apr 92 20:14:20 GMT
Sender: [email protected] (USENET News System)
Organization: Washington State University
Lines: 195
Originator: [email protected]

Hi,

There seems some confusion about how to use gcc 2.1 to compile 0.95a
kernel with Linus's second patch. That may be my fault. Ok. Here is the
new FAQ for gcc 2.1. Please save it. THE ONE ON FTP SITES IS NOT SO
CLEAR AS THIS ONE.


H.J.
----------FAQ FOE GCC 2.1----
QUESTION: Where is the gcc 2.1 beta?

ANSWER: It's on banjo.concert.net under /pub/Linux/GCC and
headrest.woz.colorado.edu under /pub/linux/gcc. You may find it on the
other sites. But I am not sure if they are up to date. There are two
files, 2.1lib.tar.Z and 2.1misc.tar.Z. Another file, binutils.tar.Z,
is not the part of gcc 2.1, but used by gcc 2.1. It contains all the
binary utilities you need to run gcc 2.1.

There may be also some files called 2.1shared-x.tar.Z. They have some
programs compiled in .a format. It has almost everything you want for
Linux. Please take a look at

/usr/shared/obj/README*

after unpack 2.1shared-A.tar.Z under /usr.

QUESTION: What are the contents of them?

ANSWER: 2.1lib.tar.Z contains cc1, cc1plus and cpp. 2.1misc.tar.Z
contains gcc 2.1 drivers, header files, libraries, manual pages and
installation instructions.

QUESTION: How do I install it?

ANSWER: First of all, backup the old compiler if you have one and
remove it from your hard drive. YOU MUST BE ABSOLUTELY SURE THERE IS NO
ANY OTHER C COMPILER ON YOUR SYSTEM INSTALLED. I have heard too many
stories about it. By compiler, I mean the executables, libraries
(crt0.o) and header files. This gcc 2.1 will only work with the kernel
above 0.95. You may have to have 2 MB ram to run it and have some swap
space (> 4MB) since cc1 and cc1plus are about 1 MB. First do

cd /usr
tar xvpfvz xxxx/2.1misc.tar.Z

Then read FAQ in /usr/install/gcc2. After

cd /usr
tar xvpfvz xxxx/2.1lib.tar.Z

run the shell script "inst2.x". If you need binutils.tar.Z, do a

cd /usr
tar xvpfvz xxxx/binutils.tar.Z

Please check the file permission for all files you have unpacked. Maybe
I should write a shell script to do that.

QUESTION: How do I install a new release of gcc 2.1?

ANSWER: Depends on the nature of the new release, if the gcc itself is
changed, you have to install everything in 2.1lib.tar.Z and
2.1misc.tar.Z come with the new release. If just the libraries are
changed, you can just pick 2.1update.tar.Z and unpack it in /usr. Then
do

cd /usr
tar xvpfvz xxxx/2.1update.tar.Z
cd /usr/install/gcc2
update2.x

In both cases, it is recommended to relink all the shared executables.
If the gcc itself is changed, a new set of .a's may be released too.

QUESTION: What are the main differences with the old release?

ANSWER: Here are the features

1. gcc and g++ are in one.

2. ANSI and BSD 4.4 compatible stdio.

3. iostream for g++. It has

* Full implementation of ANSI C stdio.
* Full implementation of the streambuf layer of AT&T's
iostream library for C++.
* Full compatibility between stdio and streambufs.

and libg++.a 2.0.

4. Libc.a is almost new. Most of glibc.a 1.02 (nonnetwork stuffs) are
here.

a. ctype
b. grp
c. locale
d. malloc (libmcheck.a)
e. posix
f. pwd
g. stdlib
h. string

Some other functions are added.

b. cvt
b. drand48
c. getlogin (poeigl-1.2)
d. getpass (poeigl-1.2)
e. mkfifo (is that supported in kernel?)
f. regex (GNU)
g. curses (BSD)
h. swab
i. time (BSD, untested)
j. ufc
k. utmp stuffs (poeigl-1.2)

These old functions are not changed very much.

a. soft math (fmod is added)
b. math (some are new)
c. termcap
d. unistd

5. The shared libraries are here. To link with the static libraries,
add the -static flag to gcc.

QUESTION: Is stdio ANSI compatible?

ANSWER: Yes, please test it.

QUESTION: Is g++ in 2.1?

ANSWER: Yes. Use g++ to compile the C++ code. Libg++.a 2.0 is here.
One thing missing is curses. I will take a look if I have time. And
fix.o failed to compile.

QUESTION: What options can I use for gcc?

ANSWER: Read manual page, gcc.ps or gcc.man. -O3, -O4, -O5 or even
higher should work with 2.1. -O3 is said to be the best. Don't ask me
why. Try it yourself. I was told you could manipulate those optimation
flags to get the best performance. Also -static tells gcc to use the
static libraries. The default is the shared libraries.

QUESTION: Where is the source code of the new libc.a?

ANSWER: The same place you find this file. It is called libc-2.1.tar.Z.

QUESTION: How can I use gcc 2.1 to compile the kernel?

ANSWER: The kernel has to be compiled with the header files come with
it. Please use my 2.1kernel.tar.Z and 2.1ps.tar.Z, which can be found
where you get gcc 2.1. The kernel uses its own header files. Do not
include any header files from gcc 2.1 while compiling the kernel.

QUESTION: Why does gcc say "foo.h not found" when compiling the
kernel and I am following the instructions above?

ANSWER: Not every source code under ./linux is kernel code. By kernel,
I mean those linked together and written to boot image. If there is a
file which is compiled to run under Linux, that is an application and
you should follow the instruction below. ./tools/build is one of them.
It should be compiled with gcc 2.1 header files. Please read my
2.1kernel.tar.Z.

QUESTION: How can I use gcc 2.1 to compile applications?

ANSWER: Any application has to be compiled with the header files come
with gcc 2.1. If you need some header files from the kernel sources,
do a

#include "/foo/src/linux/include/linux/bar.h"

It is strongly recommended to prototype all the library functions by
including the appropriate head files. At least, it will save you
trouble of malloc (0) (read stdlib.h).

QUESTION: Why does a program compiled ok by gcc 2.1 and the new binary
utilities get an error from kernel saying something about binary file
format or some weird things?

ANSWER: Linux has changed a.out.h. The new a.out.h in Linus' second
patch is compatible with the applications compiled with the old
a.out.h. But if an application is compiled with the new ld
(binutils.tar.Z) which uses the new a.out.h, you have to use the
kernel compiled with the new a.out.h. You have to recompile the kernel
with Linus' second patch applied and old ld. And then use new ld to
link the programs.

QUESTION: Why does g++ complain even die?

ANSWER: You need "expr", which is in GNU shell utilities 1.6, echo (?)
and sed.


[next article]
From: [email protected] (tim)
Newsgroups: alt.os.linux,comp.os.linux

Subject: RE:Scott's HD timeout problems
Summary: HD timeoutHD timeoutHD timeout HDtimeout HD timeout
Message-ID: <[email protected]>
Date: 8 Apr 92 03:43:28 GMT
Sender: [email protected] (tim)
Organization: University of Minnesota
Lines: 55

Scott:
:Sorry to be a real pain in the but, but here's the next installment in
:the ongoing saga of 'Scott Tries To Install Linux'.
:Early this morning playing around trying to get things to work. Decided
:to boot up with 0.95 as suggested in I Reid's Beginners Guide. Made an fs,
:rebooted, and it worked, no probs. Tried 0.95a and :
:8 virtual consoles
:4 pty's
:HD timeoutHD timeout .......
:dev 0300, block 0
:Unable to read partition table of device 0300
:HD timeoutHD timeout .......
:dev 0340, block 0
:Unable to read partition table of device 0340
:Partition tables ok. /* Wonderful !!!! */
:harddisk I/O error
:dev 0341, block 1
:Kernel panic : Unable to mount root

:Before switching off, decided to reboot with 0.95a, just in case.
:Amazing, it worked. Closed eyes, took a deep breath, tried again, it
:worked...

i got the same errors the first time i tried 0.95[a]. have you talked
to Linus? he helped me alot with those errors. i still get some error
messages but now i can run it without worrying about it crashing often. the 95c
patchces might help some i think.

:0.95 - works
:0.95a -
:HD timeoutHD timeoutHD timeoutHD timeout Partition tables ok.
:39488/43545 blocks
:14327/14515 inodes
:226 buffers - 241424 bytes buffer space
:Free mem : 1310720 bytes
:All systems functioning
:0.95a seems to work in the above manner 1 in 10 attempts. I am doing
;:nothing strange between reboots.

have you run it more than just past the boot? i got timeouts when
writing or removing a large file[s] or anytime the disk was being
accessed alot.

:Remeber I can get 0.95a to works from hda1. But for the moment at least
:I need to keep DOS as it is installed.

metoo. i have problems with one drive more than the other. it seems to
be a problem reading the drive, so i've been told. well, anyway,
you're not the only one with those problems and there could well be a
few minor changes in the HD driver that will help considerably.

0.12 might work better for you also. have you tried that? i get
unexpected intterupt errors with 0.12 but it doesn't timeout.

tim


[next article]
From: [email protected] (Hongjiu Lu -- Graduate Student)
Newsgroups: alt.os.linux,comp.os.linux

Subject: fixes for header files for gcc 2.1 and kernel
Message-ID: <[email protected]>
Date: 8 Apr 92 03:15:48 GMT
Sender: [email protected] (USENET News System)
Organization: Washington State University
Lines: 74
Originator: [email protected]

Hi,

I was told that the header files of gcc 2.1 and kernel are quite
confusing. Here is my solution. I am enclosing a program, conv. That
program will ensure only one set of header files is needed for
compiling both applications and kernel.

Before you run it, you have to have gcc 2.1 beta, mv, mvdir, rm and
tar installed on your system.

Before you do anything, please backup gcc 2.1 and kernel sources. Here
goes.

Step 1: Apply Linus's second patch and then my 2.1kernel.tar.Z.

Step 2: Read conv and understand what it is doing. Then go to the
toplevel of kernel source tree and do

conv

Step 3: Edit the toplevel Makefile and add -DPRE_GCC_2 to CFLAGS.
You may also want to get rid of "-nostdinc -Ixxxx" stuff since
./include has been moved to /usr/include.

Step 4: Type

make


H.J.
--------conv----------
#!/bin/sh
# you have to run this script under ./linux
INCS="errno.h fcntl.h limits.h signal.h stdarg.h stddef.h \
termios.h time.h unistd.h utime.h"

# move ctype.h and string.h in kernel source, which are different from
# gcc 2.1 to ./linux.
mv ./include/ctype.h ./include/string.h ./include/linux

# make a new directory
mkdir ./include.old

# move duplicated header files to ./include.old
for h in $INCS
do
mv ./include/$h ./include.old
done

# move ./include/sys to ./include.old
mvdir ./include/sys ./include.old

# move those header files which don't exist in gcc 2.1 to /usr/include
tar cf - ./include | (cd /usr; tar xpf - )

# remove ./include
/bin/rm -rf ./include

# change
#include
#include
# to
#include
#include
#
DIRS="boot fs init kernel lib mm"
for d in $DIRS
do
for f in $d/*.c $d/*.h $d/*/*.c $d/*/*.h
do
sed 's,,,' $f | sed 's,,,'
> 1
mv 1 $f
done
done


[next article]
From: [email protected] (Rob J. Nauta)
Newsgroups: alt.os.linux,comp.os.linux

Subject: alt.os.linux ?
Summary: rmgroup it ?
Keywords: obsolete rmgroup kill-it
Message-ID: <[email protected]>
Date: 6 Apr 92 09:22:39 GMT
Followup-To: alt.os.linux

Organization: Sex, UNIX & Rock 'n Roll
Lines: 12

Hi

I notice that most messages are now posted to both alt.os.linux, and
comp.os.linux. Generally comp groups get better distribution, so maybe
it's time to rmgroup it ? Any comments ?

Rob
--
+-----------------------+
| Rob J. Nauta | "In the end, all time spent is wasted time."
| [email protected] |
| Phone: +31-40-833777 |


[next article]
From: [email protected] (tim)
Newsgroups: alt.os.linux,comp.os.linux

Subject: Re: Job Ctrl && $$->programmers (How....)
Message-ID: <[email protected]>
Date: 8 Apr 92 22:27:00 GMT
Sender: [email protected] (tim)
Organization: University of Minnesota
Lines: 6


you're probably not using bash at all. you're probably using ash from
the 0.95a root disk? you can get bash from the 0.95 or 0.12 root disk
or possibly alone from one of the ftp sites.

tim


[next article]
From: [email protected] (Ed Carp)
Newsgroups: alt.os.linux,comp.os.linux

Subject: uucp porting status report - 04/08/92
Message-ID: <[email protected]>
Date: 8 Apr 92 22:06:07 GMT
Organization: Bleeding Edge Technologies - 'Sliding Down The Razor Blade Of Life
...'
Lines: 16

Well, for the most part, it works! I can't turn off debugging
in uucico (have to use -x4 or above), but it works fine with
smail3. I wrote a simple shell script to deal with local
mail - I'll be including that with both the binaries and the sources.

I just converted to gcc-2.1 from H. J., and now I can't get the
executables to execute! Can someone please email step-by-step
directions on how to do this? Thanks!

If I can get the binaries to run, I'll be sending the whole thing over to
tsx-11 Thursday morning sometime.
--
Ed Carp N7EKG/6 [email protected] [email protected]
Cupertino, CA 408/252-5947

-- Absolutely unabashed Gates McFadden groupie! --


[next article]
From: [email protected] (Ed Carp)
Newsgroups: alt.os.linux,comp.os.linux

Subject: taylor-uucp 1.03 (final) uploaded to tsx-11...
Message-ID: <[email protected]>
Date: 10 Apr 92 00:53:02 GMT
Organization: Bleeding Edge Technologies - 'Sliding Down The Razor Blade Of Life
...'
Lines: 30

Taylor uucp 1.03 (final) has been uploaded to tsx-11. The files
are /pub/incoming/uucp-1.03-src.tar.Z and /pub/incoming/uucp-1.03-src.tar.Z.

To run the binaries, you must have recompiled your kernel with
the new a.out.h file from gcc-2.1.

To install the binaries, cd to / and untar. You might want to
make sure the ownerships and permissions on the binaries are
correct. Here's what they SHOULD be:

-r-sr-xr-x uucp/root 46080 Apr 8 02:30 1992 usr/bin/uucp
-rwxr-xr-x root/root 33792 Apr 8 02:30 1992 usr/bin/uulog
-r-sr-xr-x uucp/root 33792 Apr 8 02:30 1992 usr/bin/uuname
-r-sr-xr-x uucp/root 54272 Apr 8 02:30 1992 usr/bin/uustat
-r-sr-xr-x uucp/root 46080 Apr 8 02:30 1992 usr/bin/uux
-rw-r--r-- root/root 97 Apr 4 13:06 1992 usr/lib/uucp/Devices
-rw-r--r-- root/root 284 Apr 9 09:36 1992 usr/lib/uucp/Systems
-rw-r--r-- root/root 127 Apr 6 09:58 1992 usr/lib/uucp/Systems.distri
-rw-r--r-- root/root 32 Apr 6 18:42 1992 usr/lib/uucp/paths
-rw-r--r-- root/root 154 Apr 6 18:35 1992 usr/lib/uucp/reset.c
-rwxr-xr-x uucp/root 41984 Apr 8 02:29 1992 usr/lib/uucp/uuchk
-r-sr-xr-x uucp/root 95232 Apr 8 02:29 1992 usr/lib/uucp/uucico
-r-sr-xr-x uucp/root 50176 Apr 8 02:29 1992 usr/lib/uucp/uuxqt
-rwxr-xr-x root/root 80 Apr 8 18:50 1992 usr/lib/uucp/zap

--
Ed Carp N7EKG/6 [email protected] [email protected]
Cupertino, CA 408/252-5947

-- Absolutely unabashed Gates McFadden groupie! --


[next article]
From: [email protected] (Ed Carp)
Newsgroups: alt.os.linux,comp.os.linux

Subject: Problems with serial port drivers??
Message-ID: <[email protected]>
Date: 10 Apr 92 08:43:00 GMT
Organization: Bleeding Edge Technologies - 'Sliding Down The Razor Blade Of Life
...'
Lines: 14

Has anyone seen any problems with the serial port drivers under
linux? I'm running 0.95a and am having a hard time making uucp fly
at 9600 baud and above, either straight through or with a modem.
This same machine and comm board runs XENIX HDB UUCP flawlessly,
and there have not been any reported complaints about taylor-uucp
dropping packets, etc.

This might not show up under light loads -- just under a heavy
load, like uucp at 19200.
--
Ed Carp N7EKG/6 [email protected] [email protected]
Cupertino, CA 408/252-5947

-- Absolutely unabashed Gates McFadden groupie! --


[next article]
From: [email protected] (Farhan H Garib)
Newsgroups: alt.os.linux,comp.os.linux

Subject: Re: Problems with serial port drivers??
Message-ID: <[email protected]>
Date: 10 Apr 92 16:29:44 GMT
References: <[email protected]>
Sender: [email protected] (The Network News)
Reply-To: [email protected]
Organization: Columbia University
Lines: 19
Nntp-Posting-Host: cunixb.cc.columbia.edu

In article <[email protected]> [email protected] (Ed Carp) writes:
>Has anyone seen any problems with the serial port drivers under
>linux? I'm running 0.95a and am having a hard time making uucp fly
>at 9600 baud and above, either straight through or with a modem.
>This same machine and comm board runs XENIX HDB UUCP flawlessly,
>and there have not been any reported complaints about taylor-uucp
>dropping packets, etc.
>
>This might not show up under light loads -- just under a heavy
>load, like uucp at 19200.
>--
>Ed Carp N7EKG/6 [email protected] [email protected]
> Cupertino, CA 408/252-5947
>
>-- Absolutely unabashed Gates McFadden groupie! --

Kermit seems to drop packets at 19200 also.

-Greg


[next article]
From: [email protected] (gary a moyer)
Newsgroups: alt.os.linux,comp.os.linux

Subject: got root95c+, now need device driver for lp0...
Keywords: lp, lpd, lpr, lp0
Message-ID: <[email protected]>
Date: 10 Apr 92 22:00:50 GMT
Sender: [email protected]
Lines: 8

Well, I got the .95c+ release of the boot image (kudos to those
responsible). When I boot I get an 'lp0 recognized'. Anyhow,
I still have _no_ device driver for the printer at lp0. Is this
a question of which root image I'm using (am using .95 root image)?
If not, can the driver be created using mknode and if so could someone
please post the necessary info to do this?

thanks. [email protected]


[next article]
From: [email protected] (Charles Hedrick)
Newsgroups: alt.os.linux,comp.os.linux

Subject: Re: Problems with serial port drivers??
Message-ID:
Date: 10 Apr 92 22:54:54 GMT
References: <[email protected]>
Followup-To: alt.os.linux
Organization: Rutgers Univ., New Brunswick, N.J.
Lines: 30

[email protected] (Ed Carp) writes:

>Has anyone seen any problems with the serial port drivers under
>linux? I'm running 0.95a and am having a hard time making uucp fly
>at 9600 baud and above, either straight through or with a modem.

I have had a number of problems with KA9Q. As of 0.95b or c, I
know of three possible problems:

1) There's a tendency for the controller not to bother giving a
transmit done interrupt. Linus added various timeout code that was
supposed to fix this, but I'm not convinced it does. I'm using my own
code. If remembers the time when the last transmit was started. If a
new transmit is done more than 1/2 sec or so after that, it assumes
the done was missed and does the transmit immediately.

2) Someone I know who is an expert in these things claims that
the read routine must loop until there's nothing more to read. He
claims that there is a small buffer even in the dumb controllers,
and if you don't empty the buffer each time, you'll lose.

3) There's a problem in select. If you are in emacs (which
waits for tty input at a select), and other program that uses
select will fail to activate.

I believe Linus is working on problem 3, but has declined my patches
related to the first two. So far I haven't done careful enough tests
to be sure that all three of these problems actually exist. I'm
waiting for him to fix select. Once he does, if problems remain, I'll
try to verify exactly what they are.


[next article]
From: [email protected] (Drew Eckhardt)
Newsgroups: alt.os.linux,comp.os.linux

Subject: Re: got root95c+, now need device driver for lp0...
Keywords: lp, lpd, lpr, lp0
Message-ID: <[email protected]>
Date: 10 Apr 92 23:31:08 GMT
References: <[email protected]>
Sender: [email protected] (The Daily Planet)
Organization: University of Colorado at Boulder
Lines: 19
Nntp-Posting-Host: romeo.cs.colorado.edu

In article <[email protected]> [email protected] (gary a moyer) writes:
>Well, I got the .95c+ release of the boot image (kudos to those
>responsible). When I boot I get an 'lp0 recognized'. Anyhow,
>I still have _no_ device driver for the printer at lp0. Is this

Yes, you have a device driver for line printer devices. You
just don't have a "special file" that lets programs outside of
the kernel see it...

>a question of which root image I'm using (am using .95 root image)?
>If not, can the driver be created using mknode and if so could someone
>please post the necessary info to do this?
>

Line printer devices are character devices, major number 6.

Do a

mknod /dev/lp0 6 0


[next article]
From: [email protected] (Mika Pekka Liljeberg)
Newsgroups: alt.os.linux,comp.os.linux

Subject: Re: Problems with serial port drivers??
Message-ID:
Date: 11 Apr 92 10:40:27 GMT
References: <[email protected]>
Sender: [email protected] (Uutis Ankka)
Organization: Department of Computer Science, University of Helsinki, Finland
Lines: 35
In-Reply-To: [email protected]'s message of 10 Apr 92 08: 43:00 GMT

In article <[email protected]> [email protected] (Ed Carp) wrote:
> Has anyone seen any problems with the serial port drivers under
> linux? I'm running 0.95a and am having a hard time making uucp fly
> at 9600 baud and above, either straight through or with a modem.
> This same machine and comm board runs XENIX HDB UUCP flawlessly,
> and there have not been any reported complaints about taylor-uucp
> dropping packets, etc.
>
> This might not show up under light loads -- just under a heavy
> load, like uucp at 19200.

Yeah, I have experienced problems at high speeds, too. There is packet
loss when transferring files with kermit, and the modem has an irritating
habit of suddenly dropping off-line, if simultaneous heavy disk activity
occurs. The surest way to effect this is to compile something on another
console (thank God I don't need swap in order to use the compiler!).
I even have a 16550AN UART chip, and it doesn't help at all (does anyone
know, if the serial driver supports 16550 fifos?).

My guess is that something in the kernel disables interrupts for too
long, perhaps its the hard disk driver, perhaps the problem occurs when
forking a new process, I don't know. It's annoying all the same.

As an aside, I've been hearing that it isn't possible to use serial ports
using the same IRQ line simultaneously. Isn't it possible to poll all the
UARTS on the IRQ line in the interrupt handler? Ie. you'd need a slightly
smarter interrupt handler. One routine would handle com1 and com3 and
another one would handle com2 and com4. I've never written an actual serial
driver, so I'm not very sure about this, but I think it ought to be possible.

Mika
--
Mika Liljeberg Email: [email protected]
Helsinki University [email protected]
Dept. of Computer Science


[next article]
From: [email protected] (Bowen Goletz)
Newsgroups: alt.os.linux,comp.os.linux

Subject: Re: Problems with serial port drivers??
Message-ID: <[email protected]>
Date: 11 Apr 92 18:56:39 GMT
References: <[email protected]>
Sender: [email protected]
Followup-To: alt.os.linux
Organization: Purdue University Computing Center
Lines: 24

In article [email protected].
FI (Mika Pekka Liljeberg) writes:
>In article <[email protected]> [email protected] (Ed Carp) wrote:
>> Has anyone seen any problems with the serial port drivers under
>> linux? I'm running 0.95a and am having a hard time making uucp fly
>> at 9600 baud and above, either straight through or with a modem.
>
>occurs. The surest way to effect this is to compile something on another
>console (thank God I don't need swap in order to use the compiler!).
>I even have a 16550AN UART chip, and it doesn't help at all (does anyone
>know, if the serial driver supports 16550 fifos?).

I have a 16550 on my second serial port and have major problems with Linux
grabbing incomming data. I have a modem connected to the port, and I
see that characters are sent, but I never see any echo. The modem
lights for RX bink on command completion (eg, OK), but kermit does
not show modem response. Ideas?

-bg

--
________________________________________________________________________________
Bowen Goletz
Purdue University Midi 'n Music
[email protected] [email protected]


[next article]
From: [email protected] (Todd Radel)
Newsgroups: alt.os.linux,comp.os.linux

Subject: Re: got root95c+, now need device driver for lp0...
Keywords: lp, lpd, lpr, lp0
Message-ID: <[email protected]>
Date: 12 Apr 92 02:12:40 GMT
References: <[email protected]> <[email protected]>
Followup-To: alt.os.linux
Organization: University of Delaware
Lines: 28

In article <[email protected]> [email protected] (Dre
w Eckhardt) writes:
>In article <[email protected]> [email protected] (gary a moyer) writes:
>>Well, I got the .95c+ release of the boot image (kudos to those
>>responsible). When I boot I get an 'lp0 recognized'. Anyhow,
>>I still have _no_ device driver for the printer at lp0. Is this
>>a question of which root image I'm using (am using .95 root image)?

>Line printer devices are character devices, major number 6.
>
>mknod /dev/lp0 6 0

Well, I tried that. I typed:

mknod /dev/lp0 c 6 0

and rebooted. The kernel reported something along the lines of "lp1
exists (0)". Then I tried:

cat some_text_file > /dev/lp0

I got an error "Cannot write to device". Any explanations?


--
* Todd Radel Sysop, Dickinson Domain BBS *
| CIS/English FidoNet: 1:150/160; VirtualNET @3023 |
| University of Delaware REGISTERED Alpha Colony VI! |
* "You don't need to be crazy to be a sysop, but it helps!" *


[next article]
From: [email protected] (William A Jones)
Newsgroups: alt.os.linux,comp.os.linux

Subject: Modems and 0.95c+
Message-ID: <[email protected]>
Date: 12 Apr 92 19:15:58 GMT
Sender: [email protected]
Organization: The Ohio State University
Lines: 69
Nntp-Posting-Host: bottom.magnus.acs.ohio-state.edu

I posted a brief question about this yesterday but have had no responses,
so here is a more detailed description of the problems and the solutions
I have attempted so far.

First, when I was running linux 0.12, I was able to connect to my modem
(a Everex Evercom II) and cause it to dial. I never could get the
baud rate set correctly (1200 bps) but I really didn't have a need for
it so I didn't worry about it.

Now that I have 0.95c+ running, kermit and UUCP are available and a port
of Cnews is in the works and I'm working on one for trn, I would like
to run my UUCP node under linux rather than under DOS (Waffle is not
very graceful when it comes to passing mail along, besides the multi-
tasking capcbilities of linux means that I will not be tying my terminal
up for hours on end with my news feed). Now, however, linux does not
seem to recognize that I have a modem at all, much less a COM2 port (mouse
is on COM1).

My inittab file is a barebones file consisting of only the /dev/tty[1-4]
entries which come with a standard distribution. I have the following
line in my /etc/rc file
/bin/stty 1200 < /dev/ttys2 > /dev/ttys2
but this is a late development and everything which I have tried with this
line present has also been tried with the line absent. When I run kermit
(not version 5A as all versions of that are causing segmentation faults),
I type in the following

set line /dev/ttys2
set speed 1200
set mode hayes

and I get no error messages to any of them. But I am unable to communicate
with the modem either through the dial command or through connect. Is this
a lack of knowledge on my part concerning kermit (my only experience is with
MS-kermit) or is there a problem with my modem?

I have also run mknod (command was: mknod /dev/modem c 4 65) and tried
the same thing using /dev/modem instead of /dev/ttys2, tried /dev/ttys1 and
/dev/ttys3 just in case, and in none of the situations do I get any
responses.

One thing that does happen, when I boot linux, it sounds as though my modem
"picks up" on the phone line, yet the number will still ring. When I reboot
into DOS, before I am able to dial out using MS-Kermit, I have to issue a
hangup command. Does this seem strange to anyone else?

If I include the line
ttys2:console:/bin/agetty -m /dev/ttys2 1200
in my inittab file, friends have reported that the modem answers but that
they do not get a login prompt. So, apparently it is possible to get into
my system from the outside (at least partially) but I can't get out :(.

Since so many people have kermit (and apparently UUCP) running properly
on their systems, would someone please give me some pointers on what I
might try. Do I need to get a new modem? I am getting a Boca 14.4 kbps
modem on Wednesday (I hope). Might this solve my problem? I'd hate to
have to wait that long, but I am pulling my hair out over this.

Thanks,

Bill

================================================================
[email protected] ...!osu-mps!cambria!bjones
[email protected]
[email protected]

Three address and counting . . .
=================================================================


[next article]
From: [email protected] (Wayne Whitney)
Newsgroups: alt.os.linux,comp.os.linux

Subject: gcc 2.1 problem: parse error before '('
Keywords: gcc 2.1 problem parse error
Message-ID:
Date: 13 Apr 92 01:27:34 GMT
Lines: 18
Nntp-Posting-Host: husc10.harvard.edu

Hello Linux users,

As this is my first post to the .os.linux group(s), I just
want to thank Linus and everybody else whose been working on this
project for a great job and all the time and effort you've put into
it. Anyway, what brings me to post is a problem I've had trying to
get gcc 2.1 to work. Basically, whenever I try to compile something,
the compiler dies on the first line with a '(', saying: parse error
before '('. I remember there was an earlier post about this problem,
but a solution was never posted. So I was hoping that someone could
tell me how to fix it, as I don't really know where to begin. I'm
sure the archives that I've gotten weren't corrupted, and I've checked
twice to be sure that all of gcc 1.4 was gone; I followed the
instructions in the FAQ on gcc2.1 for installing it.
Thanks.

Wayne Whitney
[email protected]


[next article]
From: [email protected] (Hongjiu Lu -- Graduate Student)
Newsgroups: alt.os.linux,comp.os.linux

Subject: Re: gcc 2.1 problem: parse error before '('
Keywords: gcc 2.1 problem parse error
Message-ID: <[email protected]>
Date: 13 Apr 92 02:28:12 GMT
References:
Sender: [email protected] (USENET News System)
Organization: Washington State University
Lines: 28

In article [email protected] (Wayne Whitne
y) writes:
>Hello Linux users,
>
> As this is my first post to the .os.linux group(s), I just
>want to thank Linus and everybody else whose been working on this
>project for a great job and all the time and effort you've put into
>it. Anyway, what brings me to post is a problem I've had trying to
>get gcc 2.1 to work. Basically, whenever I try to compile something,
>the compiler dies on the first line with a '(', saying: parse error
>before '('. I remember there was an earlier post about this problem,
>but a solution was never posted. So I was hoping that someone could
>tell me how to fix it, as I don't really know where to begin. I'm
>sure the archives that I've gotten weren't corrupted, and I've checked
>twice to be sure that all of gcc 1.4 was gone; I followed the
>instructions in the FAQ on gcc2.1 for installing it.
> Thanks.
>
>Wayne Whitney
>[email protected]

The solution has been posted several times. MAKE SURE YOUR TAR/COMPRESS
WORK CORRECTLY. It is in FAQ of 2.1 too.

NOT ALL TAR/COMPRESS'S ARE WORKING RIGHT. If you are not sure,
uncompress gcc 2.1 on the other machine and move gcc 2.1 binaries to
Linux one at a time.

H.J.


[next article]
From: [email protected] (Brian Olsen)
Newsgroups: alt.os.linux,comp.os.linux

Subject: Just plain questions
Keywords: gcc temp files and where to find TeX & groff guides
Message-ID: <[email protected]>
Date: 13 Apr 92 15:34:06 GMT
Sender: [email protected]
Followup-To: alt.os.linux
Organization: University of Missouri - Rolla
Lines: 13


Hello!
Last night I decided to try some stuff out on Linux, so I started
up three compile jobs at once with gcc2.1. About 15 minutes into the
jobs, I started getting errors regarding the cpp temp files.
When I ran the jobs separately the programs went without minimal error.
Is this a problem with the kernel or with gcc keeping track off temp files?

Another question:
Where is the best site to pick up a TeX and/or groff user's guide?

Brian
[email protected]


[next article]
From: [email protected] (Demian A. Johnston)
Newsgroups: alt.os.linux,comp.os.linux

Subject: Arrow keys and emacs
Message-ID:
Date: 13 Apr 92 19:00:10 GMT
Organization: Sophomore, Physics, Carnegie Mellon, Pittsburgh, PA
Lines: 9

I saw it posted here not too long ago a few lines to add to one's .emacs
file to let you use the arrow keys with emacs ??? Could someone repost
that?


Demian
---> [email protected]
---> [email protected]
---> [email protected]


[next article]
From: [email protected] (Ed Carp)
Newsgroups: alt.os.linux,comp.os.linux

Subject: Re: Problems with serial port drivers??
Message-ID: <[email protected]>
Date: 13 Apr 92 17:31:13 GMT
References: <[email protected]>
Organization: Bleeding Edge Technologies - 'Sliding Down The Razor Blade Of Life
...'
Lines: 14

In article [email protected].
FI (Mika Pekka Liljeberg) writes:

>forking a new process, I don't know. It's annoying all the same.

It's more than annoying -- I'm running a Telebit modem, and I get about
1MB of stuff a day through. I've had to drop down to 2400 baud with
another modem to get it to consistently work. It's only a matter of time
until the sysadmin upstream finds out that I'm talking to him at 2400
baud and blows a gasket 🙂
--
Ed Carp N7EKG/6 [email protected] [email protected]
Cupertino, CA 415/336-0797

-- Absolutely unabashed Gates McFadden groupie! --


[next article]
From: [email protected] (William A Jones)
Newsgroups: alt.os.linux,comp.os.linux

Subject: Where is system-0.95a.tar.Z??
Message-ID: <[email protected]>
Date: 15 Apr 92 01:29:09 GMT
Sender: [email protected]
Organization: The Ohio State University
Lines: 16
Nntp-Posting-Host: top.magnus.acs.ohio-state.edu

Last week a notice of system-0.95a.tar.Z was posted and I retreived the
file. Since then I have had a hard disk crash and am now trying to
rebuild my linux system but system-0.95a.tar.Z seems to have disappeared!
Maybe I'm overlooking something (this archive indexes are all about a week
old). Please repost the location or if it has been removed, tell me where
I can get a copy (I really want the login and passwd routines -- they let
me set up a multiuser system easily).

Thanks,
Bill

======================================================================
[email protected]
[email protected]
[email protected]
======================================================================


[next article]
From: [email protected] (Hongjiu Lu -- Graduate Student)
Newsgroups: alt.os.linux,comp.os.linux

Subject: How to make shared libs and .a
Keywords: shared libs
Message-ID: <[email protected]>
Date: 15 Apr 92 05:19:29 GMT
Sender: [email protected] (USENET News System)
Organization: Washington State University
Lines: 154
Originator: [email protected]

It seems gcc 2.1 is quite stable and we are moving toward to making it
the standard compiler for Linux. Here are the stuffs you need to
build the shared libs. Before you do anything, PLEASE READ README
FIRST. I'd like to see more and more programs be distributed in .a
format.

H.J.
--------CUT HERE---------------------
begin 644 shlib.tar.Z
M'YV0+E[,0<,FC9@7`!(J7,BPH<.'$"-*G$@1`(B+-VK4`&$1!(T8-F)PO$@2
MQDB2*"^:!%$CQ@T9-V;8@`%#QL48,&+,J&&19\6?0(,*'5IG#ITPM:$0Y;][0<4JUJM6K6+-JM1IP8,^64F?.&C9TR+L9L76N5I`T:-$Y^#'FR
M9%V4-FS,L,DR1@T:-VC(K+$RAHR_/$'`8,NXL<.B1Y-R9+H6JE3'F#-KWJQY
M#!JD(%04#7-F+)VD/4`H$!%V;-FS.BY*<6VV#(B!2,N0`1$6:1X08MSE@0:=S0>0."#IHR:9(:#4,GS1@0;=Z0J<-FK`L1J[FX`2_BXNC2.L235[\Z
M09LU9*(W;P.'O8@$8W;3H6\_@9LV(+Q0E!PO?%[email protected]$(8+TU5WG0\@_->@
M=?W]%Z"!OM&->U<)T*#-H(6@M)?95@&*&Y\(9Z.RB0@`(CD%&&&M%U-004023RC`9')CL%&'DS$:%=\;+J#A@YA-/ADE"$T$@04(.-$$@YA&R5''
M&'2`L(>2GH$6W'!CS;$%3ETDB9^)S>4!1QF.%II4&M%HTNJ8`=;Z2Q6QER0"4'"I:&UF<*"NQQ$1QUT#$'"K,F25(9>*1!!PHQI)!D
M'V(F%V@;822'@K+`R7$C"R#$J@)29]BQ11>TVJK`1;'2(6D9B]+4!;5BU&'&
M%C*82VUVVW7'KKL@8'I:'IO2<"YPPA&G**.^@F!$$DP4$9H1(*3F1!5,,!$P
MM&E0JP:U;M3A8J@/*P<"$6\P@6#"B@7\+7)F@(`"MM>%D)H,M%[D[44)P"&'
MLF:@<)YIMLB.G43(*(:!@1`]FO#&I&R=+JVT,YXH@AP@IM#VRK3+3C(((8X3A1H12@2`V
MIB!`V1T()MSP)'9Q0(I07JG&V/\V90A&&TX5UH+UR4-<@P7(XY"&2\\?;(%YE1^J[IFL'"
M'&GH4<8;-0>?`@M&6$V2UB:;<7S-S;>I`7*?W9&M//M#S!8CU+P6ZC1`MH@G\`%#">0SWOY7+,/`[>RHV0K]F70'OD\]RLV&.=^+D,@
M28I3-SXP(E@2O`'/DR=+(/%20$/
M<*"^E,!M9LJIF0CN@!0W).<,L3F4OY`CN.6`Z@UR"%3@),MDYPZ4$J!5^O?"/&'@A-0X00)2XT*.%B&%("`#WPPV0F$@,4>:)&+7@2C&(E0
MQC..*XUAK.(0VA@:-+;0?[A6H+:>B"@E`X!Q;H,%$BG&*?]MC'
M%RDJD"[X8!Y8(,E$CM"1@!0D%T&VQ7$A,25I6,$*/HD2Z#W+C*.ZHPM!H+A=
MB<"'S'+#;U+WE7IE*@UC81LI4Z*YJNWR:E`$`0%1,LS;?>H.:$C#WWPY,C.4
MZ0W&\9KS*G8QI`@N-6D0V47@Q1UR[8]VVB1=&72%@O%E;6O20QZOZ!`?-[@M
MM?7+#&.C33^&)P`4!$9*"='!!G!3T(ASSV#U!)@-]"@QD"VN8[4;&37E],S6U
M&UGN=E?*K87T8V8$`MH&V>BC^?7B>GQ>+IQCJ&(*`*58]$#5X$]416DQ`!!%\X:E+'TU0I0E6J"N2G
M.6L:O>FMLYT?)0D`!7@_4X9`?_RCJJU:ZLTN@%.!IFQ>,)W)!FB609J?1)C"
M&.8PMPKS?E$]X`C]2M@M[$^5MAIJ'PG+@KE.,0W?.TWX*IJ"]^VRLX$\[!2I
M>((WT!&V[^.?"4P@1A?<5CO=!&0+9,`M5:($M_!+S0GF4+]@9K8,P7RAW&98
MPQO&Q@U+!(&1L`/M$0YJM@RTVP,M.L'^SN&"1?37!L?EP4S]QI+/VUH)R7;(%*[0N"11(A/3NTM3
M5O&*6:SC&[\8QS'2L9.3@N,:/XQ&#HMQCAD&<1MTY&74X,H(4Q7%/@8DVK0Y+@R[&$0[%;#DR+R'.$8S-GZT9KM$W*2MMTV>(M<625$@"U*2E#1PVWA\O>QMK[R@W%J-;QPJ+Y(DRRUV,?Y"H-B4)&"S
M6^N:>\,V-A20%@0S9%M>Y2S=&,Z-5:Z*30/Q%JB]W>V[`"9L"'\YQ5YR+LO$
MQ'3?XF9H(U#KHF=HK!C80$'Q3!"F>(!T@>\59E9[>9!&G`."B8EFE/3LN9KF
M-!UD2$,YV-`-.,1;@!/UXCJX83=ANY2!2PU>6][K:5%;+YJ+.;)C)G.9E[Y(
M8EO(6,="UE.JX8RXL=(5@ACD!>]YW+C7K1"29&0C'9F+2#IB%WJC!"8VN,%%
M6G*#G`CF(_K&DTLT8I'%L/O@%8$,:"9CILI$92H(C[C$)[Z5$3`)!$/('AU0
M9Z*D9>@WC'N#XX[B('O*`2FX%%QR0,"$)`CA2U*X^,5!D`0W].E/U5GB''0P
M\XLT07NHJX-8FO,M[*SA<4S2^A,2ERL>VC,/(&I#&UP@'.`M`3$/86^Y$*8@A2'T@#4(4I`]\2X'X\Q!0>`1/.$-3X(8*$#P1QC"$";?@Z_<
M:$B7=WGF#2^"+Z1+F>C]@A/*<(?MFKX.J$_.%YS$!M>?G@VI=P/KC11YEQOA
M"4_H`48Y%#8VA4$!5L*2E@3?`Q+((/0O3X(4>E"@-]2-#1QZP>=;8*`TS``'
M-N!^$_'P`AFXP/(61PKU!R2#`U6,_&.`2G%>((;DO``I"OC/^N<@A_87Y/TO
M$'_0-`?T9W__H0!L0`;[UW_N5P?P)W\$6']N4"!DH`!]D@9PL(#^-WX!"($%
M.($7"`<*<')N\!4:V(`/.(`?^`(D^!4*@"/4UUC7MX+;YWW@)W[OAX`&01Q#
MM(1]1:!`ON!M3Z(7)AAQU!P(D('F%!P*7(U/,00*@@0=F*'@P]P(DD$VJD3NZ
MIP!NF!1P>(8N-X=^*`2C-WEQ*'J:IX<)"`(M0`4@,`-C4%87T0+,,7PDD'Q9
M0@2"5R2Z`SU;8(8_``(JHQ@@T`4*<#IN,#)EX!G,803-TAWZP1P\R")19PM%P:CYAVH&"R!8GE0(B;YH5U9QX1.R#A7Z"`N^(7:582Y008N2`(6HHB[08F6
MN'PN-R+`^`(ELHP*,(:HM7*!2(AK.%,7X8P`$HB`B%HDXB(D9QUK.(Y"PR+#
MEXTP(B,2LHYC$#`D@".*N",,@H\_DA3#)P)U6!X\\@;XV(*K$Y`#&3#"J(AR
M4#(\,@<,*8$LZ(0/J8E)@H=3!2>W,2ECD`9A0&IU8QPOEA3.X5@L9XA#H(=O
M6(C1)P5TB'F:-WDLF11`XI*_]P0N.7I+2)%-Z)`E$X@\28XX"7S66(]8>(^J
MT9!/6)1/D"0D\(Y(%X_5-(\S8H]0J8\Z4B0269,.>8U".9-I6)!0B9!B`)8R
MN7F%ER1,>9$1R98^:9$0:21)@HP8U9/V]Y--V85G*8^ZH8M46X&\G-K("?=07&:X19P(1<@,6\I81+V1A(SX!)QT1MYYBF"0`*)QE+T7!;81D0=YJP&9NQ:7$C0"=`YS>V,8:TM(,H-Q9AIP!8,`1&
MP`1!<`13`#(ML"5'L#!<\@154'@&DYQ=0@1#\`5(4`1!0`1%(`53H`!9,@5/
MP`168#"I(970]01"H`3'F1IB@B4)X(,<*(#S1Y'X%P13\)X"PG\;"(#R&8'V
M%P9S<'E$@)_LAX(=J((4F8`".IS%>9\]T`*"TP)XH`":AY\R&)(TB",M8(/A
M]X-PF)Q0D"5?,'I?\'Q#(`5/,`534*$Q:'T82I$U^'WA1QUT`*%Y,`=VH(A>
MX85)X`1#\)XM@%T#)5&1F`0NH(A)D)\,Z*%!2(1"\@)D8B9.0J$,:IQ`>@4A
M26HM\`0VH(AFT">8<@;.`22Z\2>VT0)ATP;"@J9,5P8M``>IHARL`@)M@`(].@1>=*?!6:530"M#``50\)YWJGE>U`(&M`09D0!VVL0(E`"(EL'8E0`9@<'&*`P)Y\`9"9W*]*3C/(1;44JIU@!V0`0*?
M@70:=7QH.`07T7F)=WQ(D*)>(GT)4WTS:"#;UWTRBH-`:'XQH(PGH@!4@"5'
M4`2_*@7!6J!,6J"/PR'>V03X&9=BDA8,PG,)4*AZ:H8H()R/^J>*>)R2:(81
M:88\\*V/]P;B2JZ'FB-M6)#P^JT)0J]V>J[E6JA^>J^*F*_;10+QJ@!9"@*Q
M<:?HJ9Z,^IWA.9Y>M*-B("9W*K'B600IT+`HH+$4FQ;_>J(INJ+E*J$ZJHA<
MZJ4BMZ:,TSIN"J?*,J?M2@)`8(9`(+(O&$0PH"!ST`:*QK,NT)WC"K",BJX-
M>JB$*03CR9U@X@0]P`4BD'UW*JDNEP)2JX@&4Z<)D`#M&G]TT+,`"B!@V[/=
MN;-ANU3:]7O0.;(!>ZX#JXA+V[13\+11.[4&4K5(('A86QYM**X!D6X=MQL"B0*WF@+EH;=\FRPV=Q1LP`8ZPK-!F[:Z@Q]P4"U"VX8HT*O3
M*7VTXAXY*KB@^ZQ2$*W32BL*D!S34;D>V[AU%[N66[:$TKEW2KN>AR.*=Z>B
MV[K_*@5-<*A!B0*JR[HPEP(M83?U^K'%ZZ75`D2VH:^[B[V%YT6].R3S`0=?P"R+:4\_\`,N4+_W^P-*PHUE
MR"/WF#OF2KS&"P)@0$_&X08P:X9UJ%UC``;]J)%BDCUL4"8!Y`:*]KVG6+0"
M3+[8A09U``=&4G896P3@N;%>Q"-C*R9.`@?B:AR[<0(OP`4C(,,RO#%E,#9.
M(E$I]P)Q@$4RHIB,:1L0LA_P*[_0E0`HP+\KY[\[`,"IB`;,$:0$_,1MN,!A
MY,(@L%PLP`5IL<4,P@(G``;E(0*7\[>%*JB'V@0*G$T:Z44^,,3T$;]`E[OO
M*\?S"\2X^:UL.P1%P'-,,@*U200WC"DY_)%CP7.RFMX,A9X18T(9ET<9F6.44T(`.=O&\N(0,NH1.C&9JB+`.E2M',NR7!4)X`(;AP=TH`!?D%8&QG,]\RTJ``)44'10TGBFDTS`-CBV`MY$.-54]1MYO8XQM'F@0EXQS4`MXQYO8`>D1@+@QP(E$$!XT+6@HT1H,H9%$56K`S7/^\[QK"9?0<]ET"$!#<]F
M"`-XX(AE9=!C,*$)`"TDP-`X``/X',QS<*,;1[8+"]"_'!H@D`!(P"JV04,]
M5'3I$FP;9Q3Q#`,&'09PF#O'@8U2$3LM'2C04\\R;78P#M8#?,N-`QC0<`AM3-81HI;1O$P1V!@@(\/=-O,!9N<`*Y'!J/8C>ED0)'FM,O
MO=1F%Y(T=*.*@5%'6@2%,Z?]#&A`1&I7O=3@X=4P]DJF\=)X8`9^_==^;=>3
M\W9S;78_FZ6LDG6PHR2%7==7K21JH'L@H%`N&-(JD`##W,R:31]`=!3*,=B[
M@=+8X73$D\PA!R!1IWMW``)GT7AIH%1WN=CNK-`D0`,UL-0)7=;UC-`3K3$+
MC0<7G=N`8\\LL,N[C#UK\`4ZM(3.B]=K0"W&$2@?#'>-B`/Z9@9L0!I%_7:\
MC%10<3=.?7VD)L?)PMO5:$LK#DG<&XXA7#+=/]7;EHI3G,#=1*$*LK-Y)2HCG=K$R/AU$@<`6V$=<4
M;=$F,8:[W$MBSMLW&R[>+Q+`,&W2$R;A.IH2:(EN,";8;3$42(!D00+=$Z;H:$G.1R8-!D(-%S
M4`>C-MP1+>42G>'`C=$\$\PV+BRV(2S-P1Q"WBI$3MN&@0,]?L]?;LZN,N.V
MD1IKOM@QH`/X;88Q$.-OKL^Y"6,0#M"]'2@5W>60?>(Q<+$6R$Y?8;@`^I$B
M(GAT^[2+CA0;=U"./@G;
M6;=/X`1*XK.;#F@GX`+BP05KHP!/CNFS+@[email protected][HNH%<09W\WSM30?O?0>&
M>P=`M!L,71,V\'S,#0(E/73-''5\)RY6%T1'.@7C1`?3/=VNFA3?IV_.`15U
M<`:&$W42=AIE,3C/SLVP<@=>)!9GX'27?J3`S,THX1?O,P/_SAMT`(XD!(WT`)=BA(;01*;>1$"3Q)\<1,HL2M``4C#P(=+PMMS\TU'QM/T-K8_0:K/?8B+?.QX>E.\I"-M?8`BO7!_/*Q`05]YW2O?3?/HGMX
M$`:`HO)C0=5>Q/;!S,T_I_233>;-+!*N8S>"@UUU=^E^`@=TH-C-??=#D/=[
M3SS9]>Y0P0;<7/"*J/*F^G8^!"P;=VQ)(S.IR/=+5/HFT0*H;VROB#NF"NJP
M_Y&@?XHB?]+_O`O_E;U_E;E_KM=`;EFREEPMKGNY8136+Q+##_W:GSMWX/?+T6N[8?O)`24V)"XM$\PXML1LK4'?L+RQY\/X"1_Z)OV2&_^3`;C`>4*'G!3.1!_0<2++[#:*O^HFT848=
MS,X8NFC6S>WLAHLF`S+@$F$#($(!"K.B8R+(0*]99D\B#%`UGX8'8,`,N`%F
MX)I5"RB&DJ*=#)AVFB]JS+(=R`-[H`_\"9#L(&2)[-0$BL`/!`J0*2[$F\ET
M%T3190I--R#@\!N8`!?RC4@0.`6/-"F&([@66!G#:0JO[#)PP3$XRZX`EG`"
M/>H(`+V2X:JP`]`A5:8J*:B]NT'-OAEV$0/:X3<\EVIA-^H?22IW2D("I:HC
M)03*3AM<-/'A`@DUVU#N&A\"`R$K1_$D&V81*)":[email protected]!(GDL`:00R4\
M`\UB="03`K2J0,4R$@CV"%7-#%R$V9B#$:-N=;!5F:H(408\CE00',ACZP0!
M-;;_Y@`-Z5QS``X(OG66`&A`$Q`"NE"-,8O?(`;BCAA8$;DBJOP&-65#VH#%
MR#K(9^-TKCL'`H[`.,$.S+")5`<#8LN0@@O0`DMB)4C+*Y9+DL``\%4):)F6&S>#C(Q:7#P594.TJ&/E!4VG`+:4*?$AB:@';3&
M;\!CRB3NO!W$M8R``XYX=MK/AS1"M?+=7(=R.(@)<0;$!BHP+L2$>YACV?"&
M>8216!*?UQN$792+](D)=?BJ\B%WT`]OD$@TEC?0"[email protected]>I_42TD#I\
MB'&P#5#L#I0B5_@&%<1RT8"R2NV%NO_3"4\'@"`-S6)R4;<,D7*R#AG\BF`Q
M+(I%AA`$7\!05!!CL=U@!"^"O&F"FVR*U`"9`,HZDTR``37@X1D&&'`#
M8D`J2XM"P0NN)C"H%5P38#R,%$?SM*@9!*,T%(="5A_*8BF`/-4#J%2Z0DY8
MRL!M*8&'IEI6C7I9KN--Q:E+9ZX6E8""3_S)`R6H"L0$B("?>E`!2O)41F<%
MK:05V(@*"N!A38';R"9TAP(`5SO';1VMN-4"V)7!$AP(2UXM%>&XCPJ6NSI8
M"0M<^:NB9:C,%=*"5,VQ9J&OA'6GMA<5Z%CF2C=Z$>98LV[6[G*-Z8I1B4=S
M)7E*UWD\5!_*)K6AG$4D'%9Z^E.&"3&*Q;)X%M4-6$R"F(PR9;PF:/%<0EVD
M@H#A!MZ$S`02_J)^I`B"T9451ECV("NDN`E13\`XI8;B,]GPCV)44AEJ#%C&
M!I4:()0"$`(]"N8L(&:E&SX0]-F-8B)+>2P,":DP%@J@D7\J?%5'@94>G>.=
M0I%.8'G1(?-(']DC]J(5%C))*LDER22;I)-\DE`R2DK)*4DEJZ25O))8,DMJ
MR2W));NDE_R28#),BLDQ22;+I)D\DV@R3:K)-`
end


[next article]
From: [email protected] (Demian A. Johnston)
Newsgroups: alt.os.linux,comp.os.linux

Subject: Token Ring?
Message-ID:
Date: 15 Apr 92 17:45:15 GMT
Organization: Sophomore, Physics, Carnegie Mellon, Pittsburgh, PA
Lines: 22

Hi,
I have had a wonderful time getting Linux 0.95c+ up and running. I
first had 0.95a and then compiled my own kernel. (Gcc 1.40) And
currently and loving gnu-emacs, gcc, bash, nethack, and kermit.. Great
Job Linus et al...

My problem. I have a network connection from my PC to my campus
network and the INTERNET via Token Ring.. (This sux if I want to use
unix anywhere)
I have software drivers for MSDOS. But I would love to be able to use my
token ring from under linux.. I am pretty sure the was an IBMTOKEN
packet driver that worked with the MSDOS ka9q port. Is there one I can
use with Linux's KA9Q???

Thanks again, and congratualtions to everyone on a great OS...



Demian
---> [email protected]
---> [email protected]
---> [email protected]


[next article]
From: [email protected] (Joerg Pommnitz)
Newsgroups: alt.os.linux,comp.os.linux

Subject: Re: Taylor-uucp 1.03 beta - porting status
Message-ID:
Date: 16 Apr 92 06:48:34 GMT
References: <[email protected]>
Sender: [email protected] (Owner of all binaries)
Organization: tu-chemnitz
Lines: 6

Well, could be you might help me.
In your article about Taylor-uucp you wrote that you were able to fix
the "Memory fault problem".
It's possible I have the same problem when running compress (and some other
programs).
If you found a solution not specific to uucp, please make it available.


[next article]
From: [email protected] (David J. Hughes)
Newsgroups: aus.archives,alt.os.linux,comp.os.linux

Subject: Linux patches now available from kirk.bu.oz.au
Message-ID: <[email protected]>
Date: 7 Apr 92 02:36:01 GMT
Organization: ITS, Bond University
Lines: 13

Kirk now mirrors tsx-11's linux patches area on a daily basis.

Hostname : kirk.bu.oz.au
IP Address : 131.244.1.1
Path : /pub/OS/Linux/patches


bambi
___ David J. Hughes [email protected]
/ \ / / /
/ __/ __ __ ____/ / / __ Senior Network Programmer
/ \ / \ / \ / / / / / \ / Comms Development & Operation
\____/ \__// / \__/ \___/ / / / AUSTRALIA (+61 75 951450)


[next article]
From: [email protected] (Cornell Kinderknecht)
Newsgroups: alt.os.linux,comp.os.linux

Subject: linux support MFM hard disks?
Keywords: mfm linux disk drive hard junk
Message-ID:
Date: 16 Apr 92 20:49:03 GMT
Sender: [email protected] (-- cjk --)
Reply-To: [email protected]
Organization: NEC USA, Corporate Network Administration Division
Lines: 9
Nntp-Posting-Host: vivaldi.inoc.dl.nec.com

I wonder if linux supports MFM hard drives?

--- Cornell

--
| Cornell Kinderknecht | Email: [email protected] |
| CNAD/CCSL | UUCP: uunet!necbsd!cornell |
| NEC USA/NEC America, Inc. | Phone: 214-518-3509 |
| Irving, TX | Fax: 214-518-3552 |


[next article]
From: [email protected] (Mike Roberts)

Subject: Disk full question
Message-ID: <[email protected]>
Date: 13 Apr 92 16:00:37 GMT
Organization: Cray Research, Inc., Chippewa Falls, WI
Lines: 14

The other day I was uncompressing a file in /tmp
and after a while uncompress just hung. In another
console I did a "df" and found that I had completely
filled my root partition.

My question is: shouldn't the filesystem at least
send some sort of error code when a disk is full, or
is this a problem with compress/uncompress?

I'm using 0.95c, gcc2.1, and the shared utilities in
2.1shared-A.tar.Z.

Mike Roberts
[email protected]


[next article]
From: [email protected]

Subject: Re: Pascal compiler
Message-ID: <[email protected]>
Date: 6 Apr 92 00:21:18 GMT
References: <[email protected]>
Sender: [email protected] (The News Administrator)
Reply-To: [email protected]
Organization: Rose-Hulman Institute of Technology
Lines: 20
Nntp-Posting-Host: manager.nextwork.rose-hulman.edu

In article <[email protected]> [email protected]
(Drew Eckhardt) wrote about GNU's Pascal compiler.

I had compiled p2c and had it working back with 0.12 with no problems that
I remember.

(p2c is a free pascal to C translator. I believe that it is covered by
the GNU copyright.)

I only have 20 MB drive, so I can't keep things like that around very
long.


--
Joseph T. Batt - deranged CS undergrad...
-----------------------------------------------------
\ [email protected] = &(Joseph T. Batt); \
\ \
\ Real programmers start counting with zero. \
------------------------------------------------------


[next article]
From: [email protected]

Subject: Re: serial port in linux
Message-ID: <[email protected]>
Date: 6 Apr 92 00:25:41 GMT
References: <[email protected]>
Sender: [email protected] (The News Administrator)
Reply-To: [email protected]
Organization: Rose-Hulman Institute of Technology
Lines: 23
Nntp-Posting-Host: manager.nextwork.rose-hulman.edu

In article <[email protected]> [email protected] (Andy I. Liaw)
writes:
> In article <[email protected]> [email protected] (Joseph Laval
Tremblay) writes:
..other claims....
> I can set line /dev/ttys1, which is the device name for com1,
> thanks to those reply. But when I dial there's no reaction
> from the modem. What's wrong?
>
> Andy

me too... I havent' been able to use my modem. Mine configured to com3.

I have used lines 1 and 2 for links to a NeXT and a terminal, but I can't
get my modem to work.


--
Joseph T. Batt - deranged CS undergrad...
-----------------------------------------------------
\ [email protected] = &(Joseph T. Batt); \
\ \
\ Real programmers start counting with zero. \
------------------------------------------------------


[next article]
From: [email protected] (Joel M. Hoffman)

Subject: What must be done to add a system call???
Message-ID: <[email protected]>
Date: 6 Apr 92 01:10:32 GMT
Sender: [email protected] (USENET News system)
Organization: University of Maryland, College Park
Lines: 12
Nntp-Posting-Host: next.wam.umd.edu

I have another silly question: What do I have to do to add a system
call to Linux. I've written the code, and compiled it into the
kernel, but how do I tell the compiler that outside programs can use
it??

I notice, e.g., that the code for getuid() is in a block headed by
sys_getuid(), and that getuid() appears in unistd.h. A reference to
sys_setuid() appears in sys.h. Then user programs can use setuid().
How do I work similar magic for my own kernel functions??

-Joel
([email protected])


[next article]
From: (Fred)
Newsgroups: comp.os.linux,alt.os.linux

Subject: Compilers
Message-ID: <[email protected]>
Date: 6 Apr 92 01:56:43 GMT
Sender: [email protected]
Organization: Sydney University Computing Service, Sydney, NSW, Australia
Lines: 8
Nntp-Posting-Host: mgfish33.fisher.su.oz.au

What is going to happen with compilers under Linux? We have gcc1.40 with
problems in the libraries. Then we have gcc2.x that is broken as well.

The other thing that people out there may be forgetting is that there are
some of us that only have 2M ram, and cant afford to spring for more ram, and
the move to 2.x will leave us in the cold.

just some thoughts....


[next article]
From: [email protected] (Charles Hedrick)

Subject: Re: partitioning problem
Message-ID:
Date: 6 Apr 92 02:05:57 GMT
References: <[email protected]>
Organization: Rutgers Univ., New Brunswick, N.J.
Lines: 9

[email protected] (Andy I. Liaw) writes:

>How do I restrict access to the linux partition from DOS?

Use something like edpart and set the partition to a code other than
DOS. Linux doesn't care what you claim the partition is. It treats
everything other than extended partitions the same. However DOS
apparently only look at partitions that claim to be DOS. So call it a
Minix partition or System V release 3 or whatever.


[next article]
From: [email protected] (Drew Eckhardt)

Subject: Re: What must be done to add a system call???
Message-ID: <[email protected]>
Date: 6 Apr 92 02:19:31 GMT
References: <[email protected]>
Sender: [email protected] (The Daily Planet)
Organization: University of Colorado at Boulder
Lines: 40
Nntp-Posting-Host: kinglear.cs.colorado.edu

In article <[email protected]> [email protected] (Joel M. Hoffman)
writes:
>I have another silly question: What do I have to do to add a system
>call to Linux. I've written the code, and compiled it into the
>kernel, but how do I tell the compiler that outside programs can use
>it??
>
>I notice, e.g., that the code for getuid() is in a block headed by
>sys_getuid(), and that getuid() appears in unistd.h. A reference to
>sys_setuid() appears in sys.h. Then user programs can use setuid().
>How do I work similar magic for my own kernel functions??
>

1. The system call goes into the kernel code someplace.
Mostly, this is linux/kernel/sys.c, and is
traditionally named with something begining
sys_.

2. An entry is made in
fn_ptr sys_call_table[] in sys.h.

3. For consitency, it is declared extrn() in sys.h.

4. A __NR_name entry is #define'd in unistd.h. THIS IS
IMPORTANT. The syscall[numberofparameter]
macros defined in unistd.h macros
append a name to __NR_ to get the syscall number
to use.


5. An object code module calling this, a function is
made to call this syscall generally using
the syscall macros. The
'C' source for this should #define
__LIBRARY__ before including unistd.h.

6. The object code module is added to libc.a with
ar -r libc.a [name]


`


[next article]
From: [email protected] (Drew Eckhardt)
Newsgroups: comp.os.linux,alt.os.linux

Subject: Re: Compilers
Message-ID: <[email protected]>
Date: 6 Apr 92 02:25:53 GMT
References: <[email protected]>
Sender: [email protected] (The Daily Planet)
Organization: University of Colorado at Boulder
Lines: 21
Nntp-Posting-Host: kinglear.cs.colorado.edu

In article <[email protected]> (Fred) writes:

>What is going to happen with compilers under Linux? We have gcc1.40 with
>problems in the libraries. Then we have gcc2.x that is broken as well.

There is nothing stopping you from using the gcc2.1 libraries, shared,
or static.

>The other thing that people out there may be forgetting is that there are
>some of us that only have 2M ram, and cant afford to spring for more ram, and
>the move to 2.x will leave us in the cold.
>
>just some thoughts....

gcc2.1 will work in 2M of ram, the same way that previous versions ran
in 2M of ram : with some swapping. People have verified this, and
it's slower, but it still works.

GCC2.1 will become the "standard" Linux compiler. The prebeta release
appears more stable than any of the previous versions of gcc, and
has the additional advantages of compiling c++, and doing shared
libraries.


[next article]
From: [email protected] (Ajay Shah)
Newsgroups: comp.os.linux,alt.os.linux

Subject: Re: Compilers
Message-ID:
Date: 6 Apr 92 03:57:29 GMT
References: <[email protected]>
Sender: [email protected] (Ajay Shah)
Organization: University of Southern California, Los Angeles, CA
Lines: 11
NNTP-Posting-Host: alhena.usc.edu

>The other thing that people out there may be forgetting is that there are
>some of us that only have 2M ram, and cant afford to spring for more ram, and
>the move to 2.x will leave us in the cold.

Come on... it's $130 for a 4 Meg SIMM.

-ans.

--

Ajay Shah, (213)749-8133, [email protected]


[next article]
From: [email protected] (Richard Alan Brown)

Subject: Mounted file system policy and partition tables
Message-ID: <[email protected]>
Date: 6 Apr 92 04:02:54 GMT
Sender: [email protected]
Followup-To: [email protected]
Organization: Information Technology Services, University of Melbourne
Lines: 49

Regarding hard disk partitioning:

1. It seems that the preferred Linux partitioning is this:
/ /usr /home swap DOS...
i.e. at least 5 partitions! Now, I do see the need for mounting
parts of the file system in large multi-user systems, but I expect
most of us Linuxers will be using relatively small disks (<=100MB)
with one or two users, and would like to have as much free space as
possible available to us in in one partition.

It makes sense to me to have the swap space as a partition rather
than just as a file (maybe be this is just a wee bit safer in the
event of a catastrophe?), but otherwise I don't really see the point
in subdividing the disk in such a small system. In fact, for only
one or two users, maybe ye olde Unix structure of having user areas
in /usr should be used ?

I believe that having more than 4 partitions on a hard disk requires
'extended partitions'. There seems to be some doubt whether Linux will
work properly with them. Being a conservative kind of guy (I *hate*
restoring from floppies!), I'd quite like to avoid doing this. I'm
wondering if the recommended filesystem breakdown should be changed,
or am I missing something?

2. I've seen mentioned somewhere here that after editing the partition table,
I will need to edit the boot record or something so that DOS will
not try to write into the Linux partition. Is this correct or does
DOS get all of its information from the partition table at boot time ?
What do I need to edit ? (No mention in the install notes).

3. From INSTALL NOTES FOR LINUX v0.95a :

> Jim Winstead Jr. - March 17, 1992
>
> 4) Make sure you have a free (preferably primary) partition on your
> hard drive.

Why does it matter if the partition is primary or not ? Will DOS be
upset by this ?

4. While I'm here: Linus, will the patched 0.95a kernel binary be available
for FTP ?

rab
------------------------------------------------------------------------
Richard Brown | E-mail: [email protected]
School of Physics | Phone : +61 3 344 5081
University of Melbourne | Fax : +61 3 347 4783
Parkville Victoria AUSTRALIA 3052 | Telex : AA35185


[next article]
From: [email protected] (Paul Nulsen)

Subject: Re: HD timeout Errors (with .95a)
Keywords: IDE
Message-ID: <[email protected]>
Date: 6 Apr 92 05:43:04 GMT
References: <[email protected]> <1992Apr3.144129.15122@kl
aava.Helsinki.FI>
Organization: University Of Wollongong
Lines: 42

[email protected] (Linus Benedict Torvalds) writes:

>In article <[email protected]> [email protected]
l (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 am one of these people. My system boots smoothly off the 0.95a images, but
after I have installed the binaries onto my hard disk most of them refuse to
function. There are a few corrupt sectors on the disk, but I don't think
there are enough to account for this problem (I have tried installing into
different partitions with the same results).

I might add, fsck detects no problems, and I can cat the files (though they
may be corrupt). The timeouts only occur when I try to run a binary.

My questions:

1. Are these problems basically due to bad hardware? In other words, should
I be demanding that the supplier fix the machine under warranty? I have had
some problems with the disk under DOS, but put them down to a few marginal
sectors.

2. How does linux deal with bad sectors? I am no Unix guru, but I don't see
any attempt to check for bad sectors during the installation.

>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.

While I am about it, can I join the appeal for these patches to be put into
a new set of images. Until I can get a functional system I have no hope of
really looking into the cause of my problems.

Thanks
Paul Nulsen
[email protected]


[next article]
From: [email protected] (Michael Haardt)

Subject: line printer deamon for Linux, any suggestions?
Message-ID: <[email protected]>
Date: 5 Apr 92 18:12:59 GMT
Sender: [email protected] (Newsfiles Owner)
Reply-To: [email protected] (Michael Haardt)
Organization: Gandalf - a 386-20 machine
Lines: 38
Nntp-Posting-Host: kaa

I applied the 0.95a patches to the 0.95a kernel, corrected the out_b
instruction with exchanged parameters and changed the test to != 255.
I have one printer connected to lp0 and everything works ok.

Now I need a line printer deamon. I wrote a small deamon with lpd, lpr,
lprm and lpq a few months ago for MINIX. It is a simple deamon, only made
for one queue. I am not familiar with real UNIX printer deamons, I know
how to use lpr, lpq and lprm, that's all.

My deamon will be started by lpr and then runs until the queue is empty.
When there is nothing to print, no deamon runs and no resources are
wasted. It supports multiple kinds of paper, because I sometimes change
between single sheets and endless paper. It has no priority classes, no
net interface, no printcap, no multiple queues with job scheduling or
things like that.

- Would you prefer a standard system V deamon?
- Do you think a Linux specific deamon with special support for typical
PC problems (multiple kinds of paper etc.) may be better?
- I don't know how printcap works, I only know it is a termcap like
database for printer capabilities. Do you think it is useful, do you
use it and can you give me more informations about it?
- I thought about running the spool files through compress, the printer
is so slow, that uncompressing should not need much CPU power. Any
comments?
- My lpr does not know any filter options. Do you consider them to be
useful? Any suggestions, standards on how to implement them?
- My lpr does not change the incoming data in any way, no \n to \r\n
expansion, no tab expansion, no line break after a fixed number of
characters etc. Ok?

I set the \n -> \r\n menu entry on my printer, so I can print usual text
files without converting them.

I am waiting for your suggestions, what would you appreciate to see in a
line printer deamon package?

Michael


[next article]
From: [email protected] (Michael Haardt)

Subject: File system naming conventions
Message-ID: <[email protected]>
Date: 5 Apr 92 20:16:42 GMT
Sender: [email protected] (Newsfiles Owner)
Reply-To: [email protected] (Michael Haardt)
Organization: Gandalf - a 386-20 machine
Lines: 21
Nntp-Posting-Host: kaa
Return-Path:

- Where should the swap file/partition be? I saw /swap, but personally,
I would appreciate /dev/swap for various reasons.

- If there is more than one device which should be mounted temporarily,
/mnt is not enough. I recently saw /mnt/exch0, /mnt/exch1 ...

- I don't like the various links of gcc, cc, gas, as, gcc-as etc. gcc is
the system compiler, why not just using cc, as, ld and putting cc1 in
/usr/lib? Concerning cpp, I am not sure. I am used to see it in either
/lib or /usr/lib, but never in /usr/bin. But then again, I think it is
very different to e.g. cc1 in meaning of being an internal compiler pass.
Many programs use it as usual command, so /usr/bin might be better.

- The floppy devices: Why not using the MINIX naming conventions? /dev/at0
would be /dev/fd0-dshd5in, /dev/PS1 would be /dev/fd1-dshd3in. I can
post the complete MINIX floppy names, but I think not all of them are
implemented in Linux.

Any comments?

Michael


[next article]
From: [email protected] (E Elu)

Subject: UUCP
Message-ID: <[email protected]>
Date: 3 Apr 92 00:46:35 GMT
Sender: [email protected] (Usenet system)
Organization: Monash University
Lines: 12


I am looking for a reliable port of a uucp system to Linux so I can set
up a bbs. I have tried Taylor version 1.3 with no luck, and also a couple of
the minix systems still with no sucess in getting all the modules to port
across.

Any suggestions or diffs would be greatly appreciated.


Thanky in advance

Ernie Elu


[next article]
From: [email protected] (Sverker Wiberg)

Subject: Re: LINUX FAQ
Keywords: linux
Message-ID:
Date: 6 Apr 92 08:07:05 GMT
References: <[email protected]> @dartagnan.rutgers.edu> <[email protected]> <1992Apr1.135523@h
ammer.Prime.COM>
Sender: [email protected]
Organization: UDAC, Uppsala, Sweden
Lines: 33

[email protected] (Kevin Cummings) writes:

> - There is no "jmp short", the opcodes are "j" for a short jump and
> "jmp" for a long one.

> - "jmpi" is a jump with a segment:offset pair. I don't know how this
>is
> written in DOS assembly.

MASM and TASM talks about short, near and far jumps, where a
short jump is +/- 128 (could be +130/-126...) bytes, a near jump
is within the current code segment (reload ip only), and a far
jump is anywhere in address space (reload cs too).

Ordinarily, you don't specify what sort of jump you want, so
JMP foo
becames a short, near, or far jump depending on how far away
the label foo is (the shortest one that will reach, is picked).
If you want to defeat this feature you write one of
JMP short foo
JMP near foo
JMP far foo
which will assemble correctly, or complain if stretched too far.

Debuggers (DOS debug, Turbo Debugger) usually give the fully
qualified op-code to tell the user exactly what sort of jump it
is.

--
--------------------------------------------------------------------------------
--------------
#include Every interesting data structure
is recursive.
[email protected] (Sverker Wiberg) /Cons
--------------------------------------------------------------------------------
--------------


[next article]
From: [email protected] (Henning Spruth)

Subject: German 84-key keyboard support
Message-ID:
Date: 6 Apr 92 11:44:32 GMT
Sender: [email protected]
Lines: 15

When I first installed linux on my system, I had to patch the
(german) keyboard driver, because several important keys
(<, >, ~ and some others) were missing. In addition, the
cursor keys were only accessible by pressing them together with
a shift key (84-key keyboards do not have separate cursor
keys). I'd like to know if other users with 84-key keyboards
have similar problems or if there is a bug in the german key
tables.

Henning
--
------------------------------------------------------------------------
Henning Spruth [email protected]
Institute of Electronic Design Automation
Technical University of Munich, Germany


[next article]
From: [email protected] (Henning Spruth)

Subject: linux available from ftp.regent.e-technik.tu-muenchen.de
Message-ID:
Date: 6 Apr 92 12:00:25 GMT
Sender: [email protected]
Lines: 18

For ftp users in germany, linux is available by anonymous ftp
from ftp.regent.e-technik.tu-muenchen.de in directory
/pub/linux.

This directory contains a (small) subset of the files in
tsx-11.mit.edu and/or nic.funet.fi.

Henning

------------------------------------------------------------------------
Henning Spruth [email protected]
Institute of Electronic Design Automation
Technical University of Munich, Germany
--
------------------------------------------------------------------------
Henning Spruth [email protected]
Institute of Electronic Design Automation
Technical University of Munich, Germany


[next article]
From: [email protected] (gary a moyer)

Subject: where could one obtain ash?
Keywords: ash, shells
Message-ID: <[email protected]>
Date: 5 Apr 92 22:58:51 GMT
Sender: [email protected]
Lines: 4

I heard it mentioned that ash was now shipping on .95a root. Is it
posted separately somewhere?

Thanks. Gary Moyer


[next article]
From: [email protected] (Joel M. Hoffman)
Newsgroups: comp.os.linux,alt.os.linux

Subject: Re: Compilers
Message-ID: <[email protected]>
Date: 6 Apr 92 14:09:01 GMT
References: <[email protected]>
Sender: [email protected] (USENET News system)
Organization: University of Maryland, College Park
Lines: 7
Nntp-Posting-Host: next.wam.umd.edu

In article [email protected] (Ajay Shah) wri
tes:
>Come on... it's $130 for a 4 Meg SIMM.

Dell still charges $399 for their proprietary-format 4M SIMM. :-((((

-Joel
([email protected])


[next article]
From: [email protected] (Lance Pickup)

Subject: Re: What is linux?
Keywords: linux
Message-ID: <[email protected]>
Date: 6 Apr 92 13:29:35 GMT
References: <[email protected]> @dartagnan.rutgers.edu> <[email protected]> <1992Apr1.182702.8
[email protected]> <[email protected]> UI.EDU> <1992Ap
r03.132049.1
Sender: @watson.ibm.com
Reply-To: [email protected]
Organization: IBM Technology Products
Lines: 29

In article <[email protected]>, [email protected]
(Lance Pickup) writes:
|> If I can find any free time, I'll be
|> taking a stab at it, but I won't guarantee my success!

Hmmm...I may have to back down on my statement. I keep breaking through
what appears to be brick walls (afterwards they look like paper though),
only to find another brick wall in front of me!

This time it is the fact that DJGPP is adding 0x1020 to all my data
addresses. Not necessarily a problem for DJGPP because it expects
to load all its programs with GO32 which sets up a data segment
some magical way such that you need to add that number to all
the addresses! This is based on an examination of head.s only--
I have no idea what assumptions it's making in the C code!

Anyway, I am slowly becoming convinced that it will be a major
uphill battle to recompile Linux in DOS. I think the Minix
compiler is probably required.

I'll keep plugging away at it, but I'm not so optimistic any more.
--
...Lance

-----------------------------------------------------------------------
Lance Pickup VNET/IBM Internet: lpickup@btv
IBM/Vendor Systems Internet: [email protected]
Technology Products
Burlington, VT Phone: (802) 769-7104 (tie 446)
-----------------------------------------------------------------------


[next article]
From: [email protected] (Michael Hamilton)

Subject: Minor patch to swap.c for ps.
Message-ID: <[email protected]>
Date: 6 Apr 92 10:55:25 GMT
Organization: Wellington City Council, Wgtn, NZ
Lines: 82

Here's a patch to apply after Linus's recent patches, but first the
reason behind it:

>Date: Mon, 6 Apr 92 13:02:11 +0300
>From: [email protected] (Linus Torvalds)
>In-Reply-To: Michael Hamilton's message as of Apr 6, 21:56
>X-Mailer: Mail User's Shell (7.1.1 5/02/90)
>To: [email protected] (Michael Hamilton)
>Subject: Re: ps patches included in 0.95a patches?
>
>Michael Hamilton: "ps patches included in 0.95a patches?" (Apr 6, 21:56):
>> Hi,
>> I applied you're recent patches. You posting said the ps
>> patches were applied. I just noticed that swap.c doesn't appear to be
>> patched as described in [email protected] (I Reid) instructions
>> concerning the ps patches. I applied your patches to a virgin 0.95a
>> kernel, but perhaps I made a mistake somewhere. I couldn't find a
>> subtract on "->rss" in /usr/src/linux/mm/*.c, which worries me as well.
>
>You are right: it's not there. I've fixed it myself, but it wasn't in
>the patches (the ps-patches didn't patch cleanly into the mm, so I did
>them partly by hand, and left that out as I wanted to see what it did).
>The bug is trivial: it only results in some incorrect values for rss,
>which isn't used by anything but 'ps'. You might want to post the patch
>to the newsgroups: I won't post new patches until I have some other
>things fixed as well...
>
> Linus

The following patches seem to do the trick for me - they are the
result of following the ps-0.95 installation instructions posted by
[email protected] (I Reid). (I don't consider myself to be a
kernel hacker - so buyer beware.)

----- cut here -----
*** swap.c.orig Mon Apr 6 22:17:10 1992
--- swap.c Mon Apr 6 22:21:59 1992
***************
*** 156,160 ****
static int page_entry = -1;
int counter = VM_PAGES;
! int pg_table;

check_dir:
--- 156,160 ----
static int page_entry = -1;
int counter = VM_PAGES;
! int pg_table = 0, d_entry;

check_dir:
***************
*** 173,176 ****
--- 173,177 ----
goto check_dir;
}
+ d_entry = dir_entry;
pg_table &= 0xfffff000;
check_table:
***************
*** 184,189 ****
goto check_dir;
}
! if (try_to_swap_out(page_entry + (unsigned long *) pg_table))
return 1;
goto check_table;
no_swap:
--- 185,192 ----
goto check_dir;
}
! if (try_to_swap_out(page_entry + (unsigned long *) pg_table)) {
! --task[d_entry >> 4]->rss;
return 1;
+ }
goto check_table;
no_swap:

----- cut here -----
--
________________
Michael Hamilton, Computer Services Section, Wellington City Council, P.O. Box
2199, Wellington, New Zealand. Phone: (64) (4)801-3300 FAX: (64) (4)801-3020
Domain: [email protected] PSImail: PSI%0530147000090::HAMILTON


[next article]
From: [email protected] (Hien Luu)

Subject: Linux and DOS
Keywords: Partitions
Message-ID: <[email protected]>
Date: 5 Apr 92 05:11:02 GMT
References: <[email protected]>
Organization: Station Zebra Corp, Sunnyvale CA Ph:408-739-1520
Lines: 12


I would like to install Linux on my system, but the installation
instruction doesn't mention how to set the system up so one can have
DOS and Linux on the same system. Also, how do I switch between DOS
and Linux and vice versa, so would some one please explain and give

some hints...
The installation instruction did mention that Linux must be installed
on a primary partition. But in DOS, one can only have one primary
partition, so is there a trick to this?

Thanks in advance...


[next article]
From: [email protected] (Haohui Wang)

Subject: Re: serial port in linux - and Problem
Message-ID: <[email protected]>
Date: 5 Apr 92 19:47:22 GMT
References: <[email protected]>
Sender: [email protected]
Organization: UB
Lines: 20
Nntp-Posting-Host: lictor.acsu.buffalo.edu

In article <[email protected]> [email protected] (Andy I. Liaw) writes
:
>Can someone please tell me how to set line in kermit? My
>hayes compatible modem is set to COM1. What is the device
>name for it?
>
>andy

It's /dev/ttys1.

BTW - I have an internal modem configed at COM4. I tried kermit, pcomm,
but they wouldn't dial out, i can't even hear the tone.

When I issue "dial 6202" in kermit, it says will timeout in 39sec, but it
gives up in 12sec with a timeout error. Same with pcomm.

Can anybody help?

thanks

howie


[next article]
From: [email protected] (Drew Eckhardt)

Subject: Re: Linux and DOS
Keywords: Partitions
Message-ID: <[email protected]>
Date: 6 Apr 92 20:19:45 GMT
References: <[email protected]> <1992Apr5.051102.2249@Saig
on.COM>
Sender: [email protected] (The Daily Planet)
Organization: University of Colorado at Boulder
Lines: 34
Nntp-Posting-Host: kinglear.cs.colorado.edu

In article <[email protected]> [email protected] (Hien Luu) writes:
>
>I would like to install Linux on my system, but the installation
>instruction doesn't mention how to set the system up so one can have
>DOS and Linux on the same system. Also, how do I switch between DOS
>and Linux and vice versa, so would some one please explain and give
>
>some hints...

If you are booting with shoelace, with a /vmunix or whatever, you will
probably have installed winiboot, or a similar primary bootblock. At boot
up, this will prompt you for which partition you want to boot - and
will timeout to provide a default.

winiboot is similar to the OS/2 bootmanager.

If you are booting the kernel off of floppy (slow), simply insert the
bootimage floppy when you want Linux, DOS or nothing when you
want DOS to come up.

>The installation instruction did mention that Linux must be installed
>on a primary partition. But in DOS, one can only have one primary
>partition, so is there a trick to this?
>

This is incorrect. DOS will only support one "DOS primary partition".
Current versions of FDisk will create "one DOS primary partition, and
one DOS extended partition". Using an alternate partition editor
will allow you to create all four partitions. Mark the partition as
something other than DOS - 0x81 "Minix" is traditionally used because that
is the file system we are currently using.


>Thanks in advance...


[next article]
From: [email protected] (Erik Fichtner)

Subject: Linux Drivers and Kernel question.
Message-ID:
Date: 6 Apr 92 20:08:38 GMT
Organization: University of New Mexico, Albuquerque
Lines: 38

Is there an ethernet driver in the works? along with patches to ka9q to
make ethernet work?

How about multiterminal boards? I've got a Digichannel PC/xi 16port serial
board... I don't have docs, and I haven't the foggiest where to begin to
write a driver for it. Anyone else out there have neat hardware like this too?
..didn't think so. oh well.

is there a "Driver Hackers FAQ" out there somewhere to aide those people
who want to add their custom hardware?

Kernel question: I'm running 0.95. i still haven't gotten around to installing
0.95a, as it doesn't offer me much... however, all the kernel patches don't
work for 0.95. now we're up to 0.95c (in patch form only,..) is there a way
I can just download the root/boot image disks and do some quick and dirty
manipulation to grab the new kernel without having to rebuild the entire system?
..no, i don't have the kernel source online yet.. it's a real low priority..
I'm still looking for a complete working Unix enviornment before I go patching
away at the kernel. ( I came to Linux from SVR4, not dos.. I want that level
of functionality! 🙂

Oh... yet another question.. is anyone working on X windows for Linux? It does
not have to be X11r5 or anything... older r3 or r4 is plenty of functionality
at least to start out with.. somewhere i saw someone complaining that it
reqiured 300 megs to compile X and several days of time.... there's gotta be
*someone* out there with enough space to produce one good working binary..
X isn't something you need to have the source online for, IMHO... once it's
working, leave it alone.. (No, i'm obviously not an X junkie.. i just need
xv and some other graphics manipulation tools and that's it. virtual consoles
are faster than point-and-click)




--
Techs [email protected]
Disclaimer: any inaccuracies are caused by the phase of the moon
"...whaddya mean I ain't kind... Just not *your* kind!" - Megadeth


[next article]
From: [email protected] (Toshiyasu Morita)

Subject: mtools (mcopy) problems
Keywords: mtools mcopy problem help augh
Message-ID:
Date: 5 Apr 92 11:25:56 GMT
Organization: Netcom - Online Communication Services (408 241-9760 guest)
Lines: 14

I'm having some problems with mtools, specifically mcopy from tsx-11.mit.edu.
I've created the /etc/mtools file correctly (A /dev/at0, b /dev/PS1, c /dev/
hd4 -16) and mdel, mdir, mcat, etc. work fine except for mcopy.

mcopy does nothing at all - when I type mcopy c:/gcc-1.40/gccinfo.1 gccinfo.1
nothing happens. I can "mcat" the file and I can view it, but I can't
copy it to my Linux directory.

(BTW I'm still running Linux 0.12 if it's relevant.)

Thanks in advance...

Toshi Morita
[email protected], [email protected]


[next article]
From: [email protected] (Al Clark)

Subject: Re: adduser (again!)
Keywords: login, passwd, salt
Message-ID:
Date: 5 Apr 92 21:27:41 GMT
References: <[email protected]>
Organization: Self
Lines: 26

In article <[email protected]> [email protected] (Z
eyd M. Ben-Halim) writes:
>
>Sorry to be such a nag, but has anybody managed to use adduser correctly.
>From looking at the sources and those for login.c (poeigl .12) they do not
>agree on the salt to use in encryption and as a result never generate thae
>same result for a given password. From login.c salt is the first 2 char. from
>the users name. adduser uses a random salt each time it's envoked.
>I changed adduser to use uname[0] and uname[1] but no joy.
>So what gives?
>
>Zeyd
>

I believe that the adduser should use a random salt, so that you do not
normally get the same encrypted password for the same input password.

However, for login to work, you need to get the same encryped password for
the same input password, so you need to use the salt in the password file
for the user you are checking against. Adduser should be adding a completely
new account, so there is no existing password entry to check. The password
changing program should also choose a new salt for the new password.


--
Al - [email protected] - My opinions are my own.
*** Practice random kindness and senseless acts of beauty! ***


[next article]
From: [email protected] (Harold Tomlinson)

Subject: Re: line printer deamon for Linux, any suggestions?
Message-ID:
Date: 6 Apr 92 22:57:39 GMT
References: <[email protected]>
Sender: [email protected] (USENET News System)
Organization: York Computing Services
Lines: 159
In-Reply-To: [email protected]'s message of 5 Apr 92 18:
12:59 GMT


(Note:: The following is my opinion only and assumes that there is a
future for Linux that involves useful work. I don't have a copy of
Linux yet and can only share from experience with Ultrix, RiscOs, and
SunOs.)

🙂 Now I need a line printer deamon. I wrote a small deamon with lpd, lpr,
🙂 lprm and lpq a few months ago for MINIX. It is a simple deamon, only made
🙂 for one queue. I am not familiar with real UNIX printer deamons, I know
🙂 how to use lpr, lpq and lprm, that's all.

Have you heard of plpd? There is a set of programs that are in the
public domain which work slightly better than lpd/lpr/lpq/lpc/lprm/pac.

🙂 My deamon will be started by lpr and then runs until the queue is empty.
🙂 When there is nothing to print, no deamon runs and no resources are
🙂 wasted. It supports multiple kinds of paper, because I sometimes change
🙂 between single sheets and endless paper. It has no priority classes, no
🙂 net interface, no printcap, no multiple queues with job scheduling or
🙂 things like that.

These are nice features. Do you think it will save that much to
have the daemon die as opposed to having swap out util needed? What
about the overhead of startup? *nix "lpd" is not a big resource user.
When started, it reads the /etc/printcap file, then checks the printer
queues listed there (usually /usr/spool/). If it runs
out of stuff to print, it goes to sleep for a while. This usually
gets it swapped out.

🙂 - Would you prefer a standard system V deamon?

Personally, I think you should look at the plpd set. Perhaps it
will be a small-effort port.

🙂 - Do you think a Linux specific deamon with special support for typical
🙂 PC problems (multiple kinds of paper etc.) may be better?


We handle multiple forms with multiple queues, not a great way, but
it works. The other option is to use the "-C" option and have the
output-filter (see printcap(5)) check for that then send the setup
string to the printer.

🙂 - I don't know how printcap works, I only know it is a termcap like
🙂 database for printer capabilities. Do you think it is useful, do you
🙂 use it and can you give me more informations about it?

It's not that difficult. The manuals with the plpd set helps you to
understand it better. Yes, it is useful. Very. It allows you to
configure the following options:
(NOTE: IMHO: U=useful, N=not useful, D=doesnt matter, V=vital)

af where to keep track of who printed and how much (D)
br set baud rate on serial printer port (V)
cf special filter (D)
df special filter (TeX) (D)
dn binary for printer daemon. (daemon name) (D)
du userid for daemon (D)
fc clear tty flags. ie reset serial line (U)
ff string to use for form feed, ie CR vs CRLF (U)
fo form-feed on open. send a preceeding ff (U)
fs set tty flags. ie reset serial line (U)
gf special filter (D)
ic non-standard ioctrl (D)
if input filter - format and print header page (U)
lf where to log errors (U)
lo name of lock file (D)
lp device for printer. /dev/tty01 or whatever (V)
mc maximum copies (U)
mx maximum file size, ensure that /tmp has room (U)
nc remove control characters (U)
nf special filter (D)
of output filter, format document (U)
pl page length (U)
pw page width (U)
sb short banner, one line only (U)
sc supress copies (U)
sd spool directory (V)
sf suppress form feeds (U)
sh suppress banner page (U)
st printer status file name (U)

There are other options which are even less useful or may not be
relivant to a PC-*nix.

(I can send you an anotated sample printcap if you want to
understand it.)

🙂 - I thought about running the spool files through compress, the printer
🙂 is so slow, that uncompressing should not need much CPU power. Any
🙂 comments?

Usually not benificial, most files are so small they are not worth
compressing. If they are large enough to warrent compression, you
would be spending a lot of time compressing only to uncompress before
printing. There is also an option in the standard "-s" option to
lpr which uses symlinks instead of copying the file to the spool
directory.

🙂 - My lpr does not know any filter options. Do you consider them to be
🙂 useful? Any suggestions, standards on how to implement them?

Again, check plpd. Why reinvent the sports car? Filters are vital.
For example, if you have a postscript printer, you have to use the
"of" in the printcap to specify a filter that will convert the header
page to postscript. The "of" (output filter) also sends setup
information to the printer if you so desire. The the rest of the
document should be sent through the "if" (input filter) which can
check to see if it is postscript and convert if needed.

If you don't use filters, you are assuming the world has the same
printer as you.

🙂 - My lpr does not change the incoming data in any way, no \n to \r\n
🙂 expansion, no tab expansion, no line break after a fixed number of
🙂 characters etc. Ok?

Yes. lpr should not touch the file. It should only copy it to the
spool directory (if required) and go away. lpd checks the spool
directories every once in a while to see if there are any files. lpr
also creates a "control file" ...

paralandra> l /usr/spool/laser2/
total 11
1 drwxrwxrwx 2 root 512 Apr 6 17:31 ./
1 -rw-rw---- 1 daemon 103 Apr 6 17:31 cfA075paralandra
5 -rw-rw---- 1 haroldt 4832 Apr 6 17:31 dfA075paralandra
1 -rw-r----- 1 root 4 Apr 6 17:31 .seq
1 -rwxr--r-- 1 root 4 Apr 2 14:49 lock*
1 -rw-rw-rw- 1 root 26 Mar 2 16:14 laser2_status
1 drwxr-xr-x 18 daemon 512 Jan 15 10:53 ../
0 -rw-rw-rw- 1 root 0 May 15 1991 laser2_log
0 -rw-rw-rw- 1 root 0 May 15 1991 laser2_acct

The file beginning with "cf" contains information needed by lpd. lpd
looks for this file, reads it, and prints to the printer associated
with this dir (from the printcap file). If there are changes to be
made, it will be by the "of" and/or "if" specified in the printcap.

🙂 I set the \n -> \r\n menu entry on my printer, so I can print usual text
🙂 files without converting them.

🙂 I am waiting for your suggestions, what would you appreciate to see in a
🙂 line printer deamon package?
🙂
🙂 Michael


Again, I'd strongly suggest finding out about plpd. It seems
confusing at first but is well worth knowing. If you want the
information I can dig it up and send it to you.

I don't remember where it came from, nor do I know if there is a
newer version available, but, if nothing else, this information will
help you to understand the lpd suite. Please do, it would help.

--
# Harold Tomlinson ## [email protected] #
# Computing & Communications Services ## (416)736-5257-33802 #
# YORK UNIVERSITY, Ont, CANADA ## ########################### #


[next article]
From: [email protected] (Harold Tomlinson)

Subject: Re: File system naming conventions
Message-ID:
Date: 6 Apr 92 23:13:59 GMT
References: <[email protected]>
Sender: [email protected] (USENET News System)
Organization: York Computing Services
Lines: 62
In-Reply-To: [email protected]'s message of 5 Apr 92 20:
16:42 GMT


In article <[email protected]> [email protected]
ik.rwth-aachen.de (Michael Haardt) writes:

🙂
🙂 - Where should the swap file/partition be? I saw /swap, but personally,
🙂 I would appreciate /dev/swap for various reasons.

I'm not certain yet how you are partitioning the disk. I'd agree
with you on the /dev/swap but again, look to the future. Will you
allow for a /dev/swap0, /dev/swap1...? Are you using an fstab to
determine the partition mounts and can it (as with some versions of
*nix) include an additional swap partition?

Are you using a swap "file" or "partition"? This makes a big
differance. If you use a file, then I'd stick to /dev/swap. If a
partition, then why mount it?

🙂 - If there is more than one device which should be mounted temporarily,
🙂 /mnt is not enough. I recently saw /mnt/exch0, /mnt/exch1 ...

If more than one device is being mounted temporarily, it is up to
the user as to where to put it. Usually, on Paralandra (DecStation
2100) I create a /mnt0,/mnt1... if I need more than one temporary mount.

🙂 - I don't like the various links of gcc, cc, gas, as, gcc-as etc. gcc is
🙂 the system compiler, why not just using cc, as, ld and putting cc1 in
🙂 /usr/lib? Concerning cpp, I am not sure. I am used to see it in either
🙂 /lib or /usr/lib, but never in /usr/bin. But then again, I think it is
🙂 very different to e.g. cc1 in meaning of being an internal compiler pass
.
🙂 Many programs use it as usual command, so /usr/bin might be better.

No. At least, not by convention. Things in /usr/bin, if I'm not
mistaken, are things that users run. Things like diff, du, make, and
sleep. Things in /usr/lib and things that objects (in /usr/bin) and
*not users* run. Things like lpd, atrun, cron...

Things in /bin are both for users and for the system. Not that it
matters, except if you want to stick to the conventional *nix. At
boot time only the root partition is mounted, so only /bin is
available. In system installation, cc is needed so it appears in
/bin. (So do other things that you need, for example, if you boot
"single user mode".)

🙂 - The floppy devices: Why not using the MINIX naming conventions? /dev/at
0
🙂 would be /dev/fd0-dshd5in, /dev/PS1 would be /dev/fd1-dshd3in. I can
🙂 post the complete MINIX floppy names, but I think not all of them are
🙂 implemented in Linux.
🙂
🙂 Any comments?
🙂
🙂 Michael
🙂

Sounds good to me. If we can tell at a glance what the device is,
that's nice. Under Ultrix, all disks are "rz". rz0c could be my
floppy or my hard drive. How it's handled and mounted is determined
from the /etc/fstab. I like having a differant name better.

--
# Harold Tomlinson ## [email protected] #
# Computing & Communications Services ## (416)736-5257-33802 #
# YORK UNIVERSITY, Ont, CANADA ## ########################### #


[next article]
From: [email protected] (Mike Ching)
Newsgroups: comp.os.linux,alt.os.linux

Subject: directory tree
Message-ID: <[email protected]>
Date: 6 Apr 92 17:20:38 GMT
Sender: [email protected] (NetNews)
Organization: Advanced Micro Devices; Sunnyvale, CA
Lines: 8
Nntp-Posting-Host: brahms


I am just getting Linux installed and would like to use a "standard"
directory tree if possible so that I don't have to edit makefiles
any more than necessary. The installation script creates src and bin
directories in several places and it's not clear what usually goes
in them. Any suggestions?

Mike Ching


[next article]
From: [email protected] (Duty Programmer)

Subject: Re: Linux Drivers and Kernel question.
Message-ID: <[email protected]>
Date: 7 Apr 92 00:04:33 GMT
References:
Sender: [email protected]
Reply-To: [email protected]
Organization: School of Physics, University of Melbourne
Lines: 22

In article , [email protected] (Erik Fichtner) writes:
|> *someone* out there with enough space to produce one good working binary..
|> X isn't something you need to have the source online for, IMHO... once it's
|> working, leave it alone.. (No, i'm obviously not an X junkie.. i just need
|> xv and some other graphics manipulation tools and that's it. virtual consoles
|> are faster than point-and-click)
|>

Yes, but there's just no substitute to having everything visible on the screen
at the same time. Certainly there is no need for people to compile X individuall
y.

What I've been wondering is this: I suspect many of us would love to have X runn
ing
on isolated PCs: is it possible to cut out all the X networking stuff ? Would th
is
reduce the size of the server by very much ? Then we could have 2 X binaries to
ftp ...

rab
--
------------------------------------------------------------------------
Richard Brown | E-mail: [email protected]
School of Physics | Phone : +61 3 344 5081
University of Melbourne | Fax : +61 3 347 4783
Parkville Victoria AUSTRALIA 3052 | Telex : AA35185


[next article]
From: [email protected] (Bob Davis)
Newsgroups: comp.os.linux,comp.sys.ibm.pc.hardware

Subject: Help: Straighten Me Out on SCSI Controller Cards (Host Adapter?)
Message-ID: <[email protected]>
Date: 7 Apr 92 02:44:52 GMT
References: <[email protected]> <[email protected]> <1992
[email protected]>
Sender: [email protected] (News stuff)
Reply-To: [email protected] (Bob Davis)
Organization: Advanced Technology Dept., Harris ESS, Melbourne, FL
Lines: 28


I am *so* confused about SCSI -- all aspects of it. And I can't
seem to locate answers. First, what do I get for the $250 for an Adaptec 1542
SCSI controller that I don't get with the $150 controllers , or with the $35
Seagate ST02?
I have an ST02 but I cannot get it to format a drive. Do I need
a special piece of low level formatting software? I've tried using DOS
DEBUG and g=C800:5 to run the BIOS format routine in the ST02 card ROM
but it seems never to recognize that a hard drive is connected. I've
programmed the Drive address to 000 for logical unit 0, but the BIOS
routine keeps insisting that "unit 0 is not connected to this controller"
when I tell it to format unit 0.
If you can recommend a file on the net, or a book, or any material
for me to read that might make someone who is intimately familiar with
LL formatting, partitioning, and FORMATting MFM and RLL hard drives feel
comfortable with doing the analogous things to a SCSI drive, I'd like
to hear about it. (How do I tell the system about Cylinders, Heads, and
Sectors on a SCSI drive? How do I set up CMOS to tell it what kind of
drive is connected?) I am lost...

Thanks,

_____________________________________________________________________________
Bob Davis, UofALA alum \\ INTERNET: [email protected] | _ _ |
Harris Corporation, ESS \\ UUCP: ...!uunet!x102a!trantor!sonny |_| |_| | |
Advanced Technology Dept.\\ AETHER: K4VNO |==============|_/\/\/\|_|
PO Box 37, MS 3A/1912 \\ VOICE: (407) 727-5886 | I SPEAK ONLY | |_| |_| |
Melbourne, FL 32902 USA \\ FAX: (407) 729-3363 | FOR MYSELF. |_________|


[next article]
From: [email protected] (Linus Benedict Torvalds)
Newsgroups: comp.os.linux,alt.os.linux

Subject: Re: Second 0.95a alpha-patch
Summary: include-file changes
Message-ID: <[email protected]>
Date: 5 Apr 92 17:44:30 GMT
References: <[email protected]>
Organization: University of Helsinki
Lines: 51

I forgot to mention some of the changes this alpha-patch brings to the
user: the kernel include-files have been slightly changed in a couple of
cases, which can result in unexpected behaviour...

a.out.h is now the latest GNU a.out.h, and it seems to have slightly
different magic-number handling than the original 386-minix version I
used for older versions. Without recompiling the kernel with the new
a.out.h, programs linked with the new binutils will not run (unable to
execute binary file errors). This just means that even if you don't use
the patch in any other capacity, you had better upgrade to the new
a.out.h, or newer programs won't necessarily run. Right now there are
no such binaries available, but when gcc2 gets more used, they will show
up.

The other change is that limits.h and sys/dirent.h are now part of the
kernel include-files: they were needed for the readdir() system call.
Normally this wouldn't change anything, but there is also a slight
change in limits.h - NAME_MAX is now defined to be 255 so that linux
will eventually handle filesystems with longer names than 14 chars.
This means that the old direntry-routines in the library won't compile
correctly, as they depended on NAME_MAX being the size of a directory
name. I hope the gcc-2 library won't have this problem, and that we can
move over to the more general readdir-function without undue growing
pains.

The a.out.h change was made just to minimize the differences between the
linux headers and the library headers - but the second change is pretty
fundamental. If you are porting software with the old libraries, I'd
suggest keeping to the old limits.h in /usr/include - that way nothing
should break until we get non-minix filesystems. Adventurous people
might want to test out the new kernel functions that will be supported
even with new filesystems.

In case anyone is wondering why the NAME_MAX change is needed, it's due
to the fact that the old /library/ readdir only handled a 14-char
library entry. When the VFS code is enhanced to allow different
filesystems, you no longer can depend on this, and the library routine
wouldn't know what type of directory it's supposed to read - so the code
has to be moved into the kernel which knows about these things. The new
readdir() will work correctly independently of the underlying filesystem
(so that you can freely mix different filesystems without needing to
bother about it).

I'm sorry all this is certain to cause it's share of confusion (using
the old library with the new NAME_MAX and vice versa), but there wasn't
any graceful way to handle it all - unless I would have anticipated
these problems from the start, which I didn't... You can console
yourselves with the thought that linux should be able to handle longer
filenames and >64MB partitions within a couple of releases.

Linus


[next article]
From: [email protected] (Drew Eckhardt)
Newsgroups: comp.os.linux,comp.sys.ibm.pc.hardware

Subject: Re: Help: Straighten Me Out on SCSI Controller Cards (Host Adapter?)
Message-ID: <[email protected]>
Date: 7 Apr 92 05:58:24 GMT
References: <[email protected]> <[email protected]
du> <[email protected]>
Sender: [email protected] (The Daily Planet)
Organization: University of Colorado at Boulder
Lines: 73
Nntp-Posting-Host: caesar.cs.colorado.edu

In article <[email protected]> [email protected]
td.com (Bob Davis) writes:
>
> I am *so* confused about SCSI -- all aspects of it. And I can't
>seem to locate answers. First, what do I get for the $250 for an Adaptec 1542
>SCSI controller that I don't get with the $150 controllers , or with the $35
>Seagate ST02?

The 1542 :

Bus mastering - the controller will take over the bus and handle
transfers itself.

Other better SCSI hosts :

The host has inteligence on it, and may have some amount of buffering,
where as the Seagate is (Trust me on this)... very dumb. The card
handles the ACK / REQ handshaking during asyncrhonous transfer,
but that's about it. You have to manually check what phase
of the transfer you're in, weather or not your target is
still online, etc. This takes a LOT of time, and degrades
performance.

The seagate hosts are not very well suited to a multi-tasking
environment : they need the host CPU's attention 100% of the
time during a scsi command. (They can disconnect and generate interrupts on
a reconnect, elimanating some overhead, but some braindamaged
drives don't support disconnect / reconnect, and you still have
the actual data transfer)

This also means that no other
tasks will be scheduled during an read / write operation from
seek to message out from the target. On a better host,
it will get back to you when it's done or time to
actually transfer data.

> I have an ST02 but I cannot get it to format a drive. Do I need
>a special piece of low level formatting software? I've tried using DOS
>DEBUG and g=C800:5 to run the BIOS format routine in the ST02 card ROM
>but it seems never to recognize that a hard drive is connected. I've
>programmed the Drive address to 000 for logical unit 0, but the BIOS
>routine keeps insisting that "unit 0 is not connected to this controller"
>when I tell it to format unit 0.

Disk mangler will format the drive, the BIOS calls should do it.

> If you can recommend a file on the net, or a book, or any material
>for me to read that might make someone who is intimately familiar with
>LL formatting, partitioning, and FORMATting MFM and RLL hard drives feel
>comfortable with doing the analogous things to a SCSI drive, I'd like
>to hear about it. (How do I tell the system about Cylinders, Heads, and
>Sectors on a SCSI drive? How do I set up CMOS to tell it what kind of

You don't. The SCSI host doesn't know about cylinders and the
like - the SCSI device appears as an error free block device,
with logical block numbers from 0 to whatever (bad ones
can be mapped out).

In fact, many of the better drives use ZBR, and have zones of
different densities (say 70 sectors on the outer tracks,
less on the inner tracks), and X sectors per track, Y
cylinders per drive are reserved for spares. This means
"normal" disk geometries do not apply.


Note that the SCSI BIOS will make the drive "appear" to have
a normal BIOS interface, with cylinders, heads, etc. This
is just a facade, and can be pure fiction.

>drive is connected?) I am lost...

You don't tell CMOS a drive is connected. If the SCSI disk is your only
disk, tell it you don't have a hard disk. The SCSI BIOS will then
handle it.


[next article]
From: [email protected] (Nicolai Wadstr|m)

Subject: Has anyone thought about threads?
Message-ID: <[email protected]>
Date: 6 Apr 92 22:29:32 GMT
Sender: [email protected] (Lord of the News)
Organization: Lysator Computer Club, Linkoping University, Sweden
Lines: 19

Has anyone given any thoughts on implementing threads in linux?
I know it's not a standard UNIX feature, but it's were nice to
write programs with threads, it's often possible to gain a lot
of performance in using threads.
(I've done a lot of OS/2 programming and just love the idea of
threads!)
The shared libraray support (when implemented), will it be like
OS/2 DLL's? I mean a .lib or .dll file that is loaded upon runtime
of the program?

Regards,
nicolai, [email protected]


--
Nicolai Wadstr|m # Internet: [email protected]
#
# BBS: +46-40-124554 (down)
# Voice: +46-40-975535


[next article]
From: [email protected] (Drew Eckhardt)

Subject: Re: Linux Drivers and Kernel question.
Message-ID: <[email protected]>
Date: 7 Apr 92 06:02:52 GMT
References: <[email protected]>
Sender: [email protected] (The Daily Planet)
Organization: University of Colorado at Boulder
Lines: 15
Nntp-Posting-Host: caesar.cs.colorado.edu

In article <[email protected]> [email protected] writes:
>
>What I've been wondering is this: I suspect many of us would love to have X run
ning
>on isolated PCs: is it possible to cut out all the X networking stuff ? Would t
his
>reduce the size of the server by very much ? Then we could have 2 X binaries to
ftp ...
>

Nope. X clients talk to the server through sockets - it doesn't matter
weather the socket is unix domain or something else, the same code has
to be there.

The biggest bloat in 'X' comes from libraries statically linked into
each X file.

However, we now use shared libraries, which eliminates this problem.


[next article]
From: [email protected] (Brett McCoy)

Subject: Re: Binaries considered harmful
Message-ID:
Date: 7 Apr 92 05:57:16 GMT
References: <[email protected]>
Reply-To: [email protected]
Organization: Kansas State University
Lines: 30
NNTP-Posting-Host: maverick.ksu.ksu.edu
In-reply-to: Paul Allen's message of Tue, 7 Apr 1992 05:40:01 GMT

In article <[email protected]> Paul Allen m> writes:

I may be missing something really obvious here, but I just don't see
the need for all these binaries. I'd rather just have sources. With
the exception of a few things that simply couldn't be compiled, Minix
has always been a source-only system. (You got the binaries when you
bought Minix, and sources/diffs from the net kept you current from then
on.) I see no reason that Linux needs to work differently.

I, for the most part, agree with this. What I'd really like to see is
a full distribution of the standard UNIX utilities in binary form, and
the sources, or diffs the the GNU sources, for these programs right
along side them. Most of the time you grab a binary of something, and
then can't find sources or diffs, and if you do they always seem to be
for some other version. It's nice to be able to bring up a working
system without having to spend a couple days recompiling everything,
but I think I would prefer that to having something akin to DOS
(binaries only).

The great thing about Linux is that it is available in source form, so
you can play with it to your hearts content. Why is everyone else so
reluctant to include their diffs and/or sources with the binaries they
upload. I for one am reluctant to just run any jo blow's binary. How
long with it be before trojan horse or virus programs start to appear.
I know that most people blindly compile and run the sources, but if
things crash, and you go looking through the source and find trojan
code, you have proof and know where it came from. With only the
binaries it is very hard to figure out what is going on.

++Brett;


[next article]
From: [email protected] (Hongjiu Lu -- Graduate Student)
Newsgroups: comp.os.linux,alt.os.linux

Subject: Re: Second 0.95a alpha-patch
Message-ID: <[email protected]>
Date: 7 Apr 92 05:56:13 GMT
References: <[email protected]> <1992Apr5.174430.17408@kl
aava.Helsinki.FI>
Sender: [email protected] (USENET News System)
Organization: Washington State University
Lines: 64

In article <[email protected]> [email protected]
I (Linus Benedict Torvalds) writes:
>I forgot to mention some of the changes this alpha-patch brings to the
>user: the kernel include-files have been slightly changed in a couple of
>cases, which can result in unexpected behaviour...
>
>a.out.h is now the latest GNU a.out.h, and it seems to have slightly
>different magic-number handling than the original 386-minix version I
>used for older versions. Without recompiling the kernel with the new
>a.out.h, programs linked with the new binutils will not run (unable to
>execute binary file errors). This just means that even if you don't use
>the patch in any other capacity, you had better upgrade to the new
>a.out.h, or newer programs won't necessarily run. Right now there are
>no such binaries available, but when gcc2 gets more used, they will show

Ld in binutils.tar.Z with gcc 2.1 beta will produce such binaries.

>up.
>
>The other change is that limits.h and sys/dirent.h are now part of the
>kernel include-files: they were needed for the readdir() system call.
>Normally this wouldn't change anything, but there is also a slight
>change in limits.h - NAME_MAX is now defined to be 255 so that linux
>will eventually handle filesystems with longer names than 14 chars.
>This means that the old direntry-routines in the library won't compile
>correctly, as they depended on NAME_MAX being the size of a directory
>name. I hope the gcc-2 library won't have this problem, and that we can

I am changing it to sys call of readdir. But telldir and seekdir are
missing.

>move over to the more general readdir-function without undue growing
>pains.
>
>The a.out.h change was made just to minimize the differences between the
>linux headers and the library headers - but the second change is pretty
>fundamental. If you are porting software with the old libraries, I'd
>suggest keeping to the old limits.h in /usr/include - that way nothing
>should break until we get non-minix filesystems. Adventurous people
>might want to test out the new kernel functions that will be supported
>even with new filesystems.
>
>In case anyone is wondering why the NAME_MAX change is needed, it's due
>to the fact that the old /library/ readdir only handled a 14-char
>library entry. When the VFS code is enhanced to allow different
>filesystems, you no longer can depend on this, and the library routine
>wouldn't know what type of directory it's supposed to read - so the code
>has to be moved into the kernel which knows about these things. The new
>readdir() will work correctly independently of the underlying filesystem
>(so that you can freely mix different filesystems without needing to
>bother about it).
>
>I'm sorry all this is certain to cause it's share of confusion (using
>the old library with the new NAME_MAX and vice versa), but there wasn't
>any graceful way to handle it all - unless I would have anticipated
>these problems from the start, which I didn't... You can console
>yourselves with the thought that linux should be able to handle longer
>filenames and >64MB partitions within a couple of releases.
>
> Linus

Keep a copy of .a for each program you compile. You can link it against
any libraries (for now no telldir and seekdir).

H.J.


[next article]
From: [email protected] (Michael Haardt)
Newsgroups: comp.os.linux,alt.os.linux

Subject: Re: Compilers
Message-ID: <[email protected]>
Date: 6 Apr 92 21:27:02 GMT
References: <[email protected]>
Sender: [email protected] (Newsfiles Owner)
Reply-To: [email protected] (Michael Haardt)
Organization: Gandalf - a 386-20 machine
Lines: 26
Nntp-Posting-Host: kaa

From article <[email protected]>, by gandalf! (Fred):
> What is going to happen with compilers under Linux? We have gcc1.40 with
> problems in the libraries. Then we have gcc2.x that is broken as well.
And we will get 2.1 with libraries. Be patient for a while, all known
problems will be exchanged with new and more interesting ones 🙂

> The other thing that people out there may be forgetting is that there are
> some of us that only have 2M ram, and cant afford to spring for more ram, and
> the move to 2.x will leave us in the cold.
Right, I have 2 MB and I can't afford to buy a new motherboard, mine is toooo
old and this special custum memory expansion cards are disappeared from world.
(I need Micronics/Compaq compatible sandwich boards, any offers?)

So I will have to live some some swapping. Remember, it still works. I
could also buy a cheap 16 bit memory expansion card, but haven't checked
that yet. Or I could invest some work trying to implement a better
paging strategy.

Where are the problems? Beta means beta == brandnew and interesting,
may need additional work. Use Emacs without cursor keys, return or
backspace key (use Control J and Control H) and some bugs in memory
management on a keyboard with totally different layout than labeled for
a while to develop a better version, and you wouldn't complain about
these *minor* problems with compilers 🙂

Michael


[next article]
From: [email protected] (Drew Eckhardt)

Subject: Re: Has anyone thought about threads?
Message-ID: <[email protected]>
Date: 7 Apr 92 08:06:18 GMT
References: <[email protected]>
Sender: [email protected] (The Daily Planet)
Organization: University of Colorado at Boulder
Lines: 25
Nntp-Posting-Host: caesar.cs.colorado.edu

In article <[email protected]> [email protected] (Nicolai Wadstr|m) write
s:
>Has anyone given any thoughts on implementing threads in linux?
>I know it's not a standard UNIX feature, but it's were nice to
>write programs with threads, it's often possible to gain a lot
>of performance in using threads.
>(I've done a lot of OS/2 programming and just love the idea of
>threads!)

fork()ing is the same thing as creating a heavy weight thread.
lightweight threads can be implemented with minimal modifications.

Interprocess communication can be accomplished with pipes,
or signals.

Hopefully, sysV shared memory and semaphores will be
implemented too.


>The shared libraray support (when implemented), will it be like
>OS/2 DLL's? I mean a .lib or .dll file that is loaded upon runtime
>of the program?
>

Yes. Note that Linux uses shared, statically linked libraries.
With some header field munging, and running ld(1) from user space, it would
be possible to do dynamically linked shared libraries.


[next article]
From: [email protected] (A. V. Le Blanc)

Subject: Re: gcc problems
Message-ID: <[email protected]>
Date: 6 Apr 92 15:14:55 GMT
References:
Followup-To: [email protected]
Organization: Computing Centre, University of Manchester
Lines: 22

In article [email protected] (Alex R.N. W
etmore) writes:
>Linux seems to be missing quite a few
>of the process utils (kill, who, etc). I assume that this is just
>because of the betaness.

The bash shell actually contains a builtin kill command. There are
also the following utilities, which might be called 'process' or related
utilities, which have been included in various packages, and
which can be compiled and run under 095a plus patches:

doshell start a subprocess attached to a different virtual console
free report usage of memory and swap space
kill signals other processes
login lets you login and can be used instead of 'su -'
logname returns the name of the owner of the current process
nice starts a background job with low priority
nohup starts a background job that should survive parent's logout
ps process status reporting utility
top usage monitor

-- Owen
[email protected]


[next article]
From: [email protected] (A. V. Le Blanc)

Subject: Re: Problem with gcc2
Message-ID: <[email protected]>
Date: 6 Apr 92 15:34:21 GMT
References: <1992Apr3.061149.12337@
mnemosyne.cs.du.edu> <[email protected]>
Followup-To: [email protected]
Organization: Computing Centre, University of Manchester
Lines: 45


In article <[email protected]> [email protected] (J
oseph Knapka) writes:
>In article <[email protected]> [email protected]
(Zeyd M. Ben-Halim) writes:
>>[email protected] (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.
>>
>>>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 " leads to the error-message:
>>>In file incluced from gcc2test.c:1:
>>>/usr/gcc2/include/stdio.h:99: parse error before '('

This problem is frequently caused by inconsistencies in the include
files, particularly since we have had several versions of the C
compiler with several include file collections. Imagine that one
include file contains a definition of a macro 'file', while another
contains the word 'file' as the name of a field of a structure.
Either file will work properly, but any program which includes both
must contain the second (with the structure) before the first (the one
with the macro). This problem actually occurred.

Another problem has occurred in which a macro, e.g., free(a) is
defined in one header file, while the prototype free() occurs in
stdlib.h or unistd.h. Further problems have occurred when
include files depend on each other in a manner which is not clear.
I think the people getting these things out have done a marvellous
job.

The whole include file mess should actually get straightened out
only when an official 1.0 release of Linux comes out, and perhaps
not even then.

>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.

Of course, if an include file is actually corrupted, we would
expect the compiler to find syntax errors in it.

-- Owen
[email protected]


[next article]
From: [email protected] (A. V. Le Blanc)

Subject: Re: Swapfile vs. Swap partition
Keywords: Swap space
Message-ID: <[email protected]>
Date: 6 Apr 92 15:46:56 GMT
References: <[email protected]>
Followup-To: [email protected]
Organization: Computing Centre, University of Manchester
Lines: 28

In article <[email protected]> [email protected] (Tim Rhodes) writes:
>I am planning on installing linux this weekend. I'll have to repartition my
>hard drive which I plan to do with regular DOS tools.

Be warned that many DOS partitioning tools waste a bit of space in
every partition they allocate. If you're squeezing the last bytes out
of a disk, don't use an MS-DOS fdisk.

>My question is which is
>preferable, using a partition for swap space or or a swapfile?

The two are apparently about equivalent, I understand, as far as efficiency.
A swap file is much easier to resize if your needs should change.

>For my setup, how many partitions should I define?

You need at least one, the root partition. Most boot methods require
that this be a primary partition (i.e., 1--4). You may create as many
others as you like, but you'll not need more than one swap partition.

>Also in using a separate partition, the
>RELNOTEs remark is not clear if I need to make a file system on a swap
>partition or if the mkswap does it all.

mkswap does it all.

-- Owen
[email protected]


[next article]
From: [email protected] (Charles Hedrick)

Subject: we need a new gdb (I think)
Message-ID:
Date: 7 Apr 92 07:28:07 GMT
Organization: Rutgers Univ., New Brunswick, N.J.
Lines: 7

gdb does not work under 0.95c. I get a general protection fault when
doing about anything. I'm using the old ld. In fact I get it even
with images built under gcc 1.40. It's the kernel that matters: 0.95b
works, but not 0.95c. I conjecture that some critical header file
changed, e.g. the change in the process description in sched.h.
Could whoever built gdb please build it again with the new
header files, or tell me how to do it.


[next article]
From: [email protected] (Jon Tombs)

Subject: Sorry Linus
Message-ID: <[email protected]>
Date: 7 Apr 92 10:45:40 GMT
Sender: [email protected]
Organization: Robotics Research Group, Engineering Science Dept, Oxford, UK.
Lines: 31
Originator: [email protected]

The patch for linux 0.95c Linus probably doesn't want to see...

Apply to /usr/src/linux/include/linux/config.h

*** config.h^ Tue Apr 7 00:09:54 1992
--- config.h Tue Apr 7 00:09:54 1992
***************
*** 16,22 ****
#define UTS_SYSNAME "Linux"
#define UTS_NODENAME "(none)" /* set by sethostname() */
#define UTS_RELEASE "0" /* patchlevel */
! #define UTS_VERSION "0.95a"
#define UTS_MACHINE "i386" /* hardware type */

/* Don't touch these, unless you really know what your doing. */
--- 16,22 ----
#define UTS_SYSNAME "Linux"
#define UTS_NODENAME "(none)" /* set by sethostname() */
#define UTS_RELEASE "0" /* patchlevel */
! #define UTS_VERSION "0.95c"
#define UTS_MACHINE "i386" /* hardware type */

/* Don't touch these, unless you really know what your doing. */


-------
Or should we be updating patch level?

--
Jon
"I got in today just as the sun was crashing"


[next article]
From: [email protected] (Werner Almesberger)

Subject: Re: mtools (mcopy) problems
Keywords: mtools mcopy problem help augh
Message-ID: <[email protected]>
Date: 7 Apr 92 12:54:06 GMT
References:
Sender: [email protected] (USENET News System)
Organization: Swiss Federal Institute of Technology (ETH), Zurich, CH
Lines: 24

In article [email protected] (Toshiyasu Morita) writes:
> mcopy does nothing at all - when I type mcopy c:/gcc-1.40/gccinfo.1 gccinfo.1
> nothing happens. I can "mcat" the file and I can view it, but I can't
> copy it to my Linux directory.

If "nothing happens" means that mcopy returns without doing anything,
you've probably forgotten to make links for mread and mwrite. mcopy
execs mread or mwrite to do the transfers, but it doesn't warn when
the exec fails. Fix: chdir to the directory where mtools live and do
a ln mtools mread; ln mtools mwrite (or, if you're not using the
merged version, compile mread and mwrite and put them in the approp-
riate directory.)

If "nothing happens" means that mcopy hangs and has to be killed with
^C, you've probably compiled mcopy.c without -DMERGED. mcopy sets
argv[0] of the execed mread/mwrite to "mcopy" if -DMERGED is omitted,
which breaks the merged mtools. Fix: re-compile mcopy.c with -DMERGED

- Werner
--
_________________________________________________________________________
/ Werner Almesberger, ETH Zuerich, CH [email protected] /
/ IFW A44 Tel. +41 1 254 7213 [email protected] /
/_BITNET:_ALMESBER@CZHETH5A__HEPNET/CHADNET:_[20579::]57414::ALMESBERGER_/


[next article]
From: [email protected] (Craig Burley)

Subject: Re: Help Mr. Fortran
Message-ID:
Date: 6 Apr 92 18:52:36 GMT
References: <[email protected]>

<[email protected]>
<[email protected]>
<[email protected]>
<1992Apr03.132
Sender: [email protected]
Organization: Free Software Foundation 545 Tech Square Cambridge, MA 02139
Lines: 38

In-reply-to: [email protected]'s message of 3 Apr 92 14:20:14 GMT

In article <[email protected]> [email protected] (James L Nance)
writes:

I remember reading a post here not too long ago from the person who was writi
ng
gnu FORTRAN. I was wondering if this person, whose name I can not recall, ha
s
considered porting FORTRAN to Linux. I would like to see if I can get Spice2
g6
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.

Hi, that was me. If (99% likely) and when (within a month or so) I do get
a 486 box and start getting Linux up and running, I indeed would like to use
it as my main Fortran-compiler-development machine. BUT, that might take a
while (especially if I have to do driver development to get my peripherals
working) and in the meantime I'll likely be modeming in to a Sun4 to do the
Fortran work.

Furthermore, GNU Fortran hasn't even reached alpha test yet, though from a
technical perspective it's pretty close (testing and some backend changes to
get COMPLEX support working still needed).

I suggest you ftp to research.att.com, login as netlib (I think; maybe it's
anonymous ftp, but I think they changed that), get f2c from ~dist/f2c or
some such place, and use that to compile your Fortran code for use on Linux.
f2c is pretty good; it's free (public domain); and when GNU Fortran comes
out (as alpha), it'll be inter-module compatible with f2c-compiled programs.
(f2c comes with it's own I/O and intrinsic libraries, so it ought to be quite
easy to get it running under Linux.)

If nobody seems to have gotten f2c ported to Linux by the time I get Linux
working on my machine, I'll do it myself and let people know what changes,
if any, need to be made. I doubt it'll be much; I've gotten f2c running on
two 68k machines (HP and NeXT) and a Sun4 with a trivial amount of work.
(On the other hand, I haven't run a whole lot of Fortran code through it
anywhere, yet.)
--

James Craig Burley, Software Craftsperson [email protected]
Member of the League for Programming Freedom (LPF)


[next article]
From: berger@atropa (Mike Berger)
Newsgroups: comp.os.linux,comp.sys.ibm.pc.hardware

Subject: Re: Help: Straighten Me Out on SCSI Controller Cards (Host Adapter?)
Message-ID: <[email protected]>
Date: 7 Apr 92 16:06:52 GMT
References: <[email protected]> <[email protected]> <1992
[email protected]> <[email protected]>
Sender: [email protected] (News)
Organization: University of Illinois at Urbana
Lines: 24

[email protected] (Bob Davis) writes:
> I am *so* confused about SCSI -- all aspects of it. And I can't
>seem to locate answers. First, what do I get for the $250 for an Adaptec 1542
>SCSI controller that I don't get with the $150 controllers , or with the $35
>Seagate ST02?
*----
The extra money buys substantially higher performance. The Seagate is
the bottom-of-the-line. It uses the CPU heavily, and doesn't support
non-Seagate drives too extensively. I don't know what kind of software
is available for tape or CD-ROM drivers.

The Adaptec 1542B is on the opposite end of the spectrum. It's a
bus-master disk controller, and has very high performance if your
motherboard can handle it. Software support is excellent.

At the higher end are EISA type controllers, and below the 1542B
are non bus-master controllers that are a lot smarter and faster
than the ST-01/02.

--
Mike Berger
Department of Statistics, University of Illinois
AT&TNET 217-244-6067
Internet [email protected]


[next article]
From: [email protected] (Jon Tombs)

Subject: Byte Unix Benchmarks.
Keywords: benchmark.tar.Z
Message-ID: <[email protected]>
Date: 7 Apr 92 16:35:51 GMT
Sender: [email protected]
Organization: Robotics Research Group, Engineering Science Dept, Oxford, UK.
Lines: 65
Originator: [email protected]


I have uploaded to tsx-11, and I will upload to nic (if the link will stay
up for long enought), the Byte Unix Benchmarks that came through in
comp.unix.sources a while back. These required /usr/bin/time and a gawk with
trig functions so I've included them in the upload (gawk-2.13.2 and
gnu time-1.3). I expect it to appear as:

/pub/linux/sources/test_suites/benchmark.tar.Z
on tsx-11

In the mean time any _UK_ user can pick it up from ftp.robots.ox.ac.uk
in the same place.

As an idea of its output here the summary reports for my 486 and a Sparc1+

----
Sparc1+
Arithmetic Test (type = double) 2541.7 1494.0 0.6
Dhrystone 2 without register variables 22366.3 13077.0 0.6
Execl Throughput Test 16.5 10.7 0.6
File Copy (30 seconds) 179.0 83.0 0.5
Pipe-based Context Switching Test 1318.5 947.1 0.7
Shell scripts (8 concurrent) 4.0 1.7 0.4
=========
SUM of 6 items 3.4
AVERAGE 0.6


----
486 33c
Arithmetic Test (type = double) 2541.7 1891.6 0.7
Dhrystone 2 without register variables 22366.3 19281.7 0.9
Execl Throughput Test 16.5 89.2 5.4
File Copy (30 seconds) 179.0 44.0 0.2
Pipe-based Context Switching Test 1318.5 3738.0 2.8
Shell scripts (8 concurrent) 4.0 5.0 1.3
=========
SUM of 6 items 11.3
AVERAGE 1.9


Nice work Linus! READMEs etc are in the archive. mail me directly if you
have any problems.

--
Jon
"I got in today just as the sun was crashing"




















[next article]
From: [email protected] (Keith L. Fries)

Subject: Linux & Ultra 12F ESDI?
Message-ID: <[email protected]>
Date: 7 Apr 92 15:34:02 GMT
Organization: Hewlett Packard Company, Spokane, Wa.
Lines: 12

From the FAQ:

> IDE and MFM seem to work with no problem. It works, also, for some
> ESDI drive (Joincom controller with Magtron drive after you have
> commented out the "unexpected hd interrupt"-message from hd.c).

Has anyone gotten linux to run with an Ultrastor Ultra 12F ESDI
controller?

Keith Fries
Hewlett Packard, Spokane Div.
[email protected]


[next article]
From: [email protected] (Paul Fox)
Newsgroups: comp.os.linux,alt.os.linux

Subject: Re: vile 3.13 + gcc 2.1 anyone?
Message-ID: <[email protected]>
Date: 7 Apr 92 18:08:53 GMT
References: <[email protected]>
Followup-To: comp.os.linux
Organization: Cayman Systems Inc., Cambridge Ma
Lines: 14

[email protected] (Hongjiu Lu -- Graduate Student) writes:
: How good is vile? 3.13 is out. Is anyone doing the porting? If you are not
: using gcc 2.1, please do.

I'm not doing anything linux-like right now -- but I recommend version 3.14
of vile, due to a (possible) file corruption bug in earlier versions of vile
(3.8 through 3.13). (It may not show up in linux -- it depends on the
behaviour of realloc(), when asked to shorten a buffer).

paul fox, [email protected]
(vile is available from ftp.cayman.com, in pub/vile
--
paul fox, [email protected], (617)494-1999
Cayman Systems, 26 Landsdowne St., Cambridge, MA 02139


[next article]
From: [email protected] (David.L.)

Subject: bash ?
Message-ID: <[email protected]>
Date: 7 Apr 92 19:26:26 GMT
Organization: Security Pacific Automation Co., LA, CA
Lines: 8

Hi

Where do I get am image or sources for bash ? do I need to get the earlier root
image ? (.12 , .11 ) ? I'm using 0.95a.

--
----> [email protected] OR [email protected]
David L.


[next article]
From: [email protected] (Bill Flowers)

Subject: Re: Since I haven't seen an FAQ...
Message-ID:
Date: 7 Apr 92 20:17:03 GMT
References: <[email protected]> artagnan.rutgers.edu>
Organization: Quantum Software Systems, Ltd.
Lines: 13

In article [email protected]
utgers.edu (Charles Hedrick) writes:
>[email protected] (Sergio L Aponte) writes:
>The other major thing is that the kernel is small and is written
>"close to the iron". There's some philosophical similarity to Minix.
>And of course the file system is from Minix, which is probably closer
>to the Berkeley fast file system than System V (though it's still got
>the short file names). However Minix is intended to have a version 7
>flavor. Its author is attempting to resist attempts to add Berkeley
>features. A number of people claim that Linux feels fast, but since I
>haven't run any other Unix variant on my system, I have no good
>comparison. Certainly I have no performance complaints.

Has anyone run IOZONE on Linux? With what results?


[next article]
From: [email protected] (Britt Park)

Subject: libtermcap.a bugs?
Summary: Are there known bugs in libtermcap?
Keywords: termcap libtermcap
Message-ID: <[email protected]>
Date: 7 Apr 92 22:45:43 GMT
References: <[email protected]>
Sender: [email protected] (Mr News)
Organization: DSG, Stanford University, CA 94305, USA
Lines: 8


I was trying to compile my much beloved version of uEmacs 3.8, (I've made
several changes to it and am very fond of it.) and found that it compiles
but doesn't work under GCC 1.40. Does the GCC 1.40 libtermcap have a known
bug in it? I've traced the problem to that vicinity. Thanks.

Britt
[email protected]


[next article]
From: [email protected] (Britt Park)

Subject: libtermcap.a bugs?
Summary: Are there known bugs in libtermcap?
Keywords: termcap libtermcap
Message-ID: <[email protected]>
Date: 7 Apr 92 22:52:11 GMT
References: <[email protected]>
Sender: [email protected] (Mr News)
Organization: DSG, Stanford University, CA 94305, USA
Lines: 8


I was trying to compile my much beloved version of uEmacs 3.8, (I've made
several changes to it and am very fond of it.) and found that it compiles
but doesn't work under GCC 1.40. Does the GCC 1.40 libtermcap have a known
bug in it? I've traced the problem to that vicinity. Thanks.

Britt
[email protected]


[next article]
From: [email protected] (Jonathan O'Neal)
Newsgroups: comp.os.linux,comp.sys.ibm.pc.hardware

Subject: Re: ST-506 and IDE drives on the same machine ?
Message-ID: <[email protected]>
Date: 6 Apr 92 13:05:01 GMT
References: <[email protected]>
Organization: SCS/Compute, Inc.
Lines: 22

In article <[email protected]>,
[email protected] (Richard Lewis) writes:
> I have a Viglen 386LS (2M RAM, 44M ST-506 Drive, Western Digital
> Controller). I have heard that it is possible to add a new IDE drive
> and controller while maintaining the original ST-506. Does anyone know
> if this is so and how can I do it ?

I have an IDE "paddle card" controller as my primary controller and a
Seagate ST-22M as my secondary controller. The BIOS on the ST-22M works
very nicely with the IDE card as long as (1) the ST-22M is set up using the
secondary address (0x170), and (2) the second CMOS drive type is set to 0
(not installed). I run DOS this way regularly; I recently put SCO Unix on
the MFM drive, set the first CMOS Drive type to 0 and the second CMOS type
to match the MFM drive, and was able to boot unix (just by changing CMOS).
I paid about $54 for the ST-22M, which also has a floppy controller; the
ST-21M (without floppy controller) should cost less.

> Richard.
--
Jonathan O'Neal [email protected] ...!uunet!emory!scscom!jko
SCS/Compute, Inc. WD4DAR Any opinions expressed are all
Norcross, GA 30093 (404) 368-1040 mine, not those of SCS/Compute.


[next article]
From: [email protected] (Stephen Harris)
Newsgroups: comp.os.linux,alt.os.linux

Subject: Large disk partition
Message-ID: <[email protected]>
Date: 7 Apr 92 20:22:08 GMT
Organization: Spuddy's Public Usenet Domain
Lines: 27

Is it possible for Linux to use a disk partition greater than 64Mb ?
I have a 650Mb disk at 1283x16x63. DOS takes up the first 1023 clinders
(the most it can have), leaving me 128Mb as a second partition. But when I
try mkfs it moans for greater than 64Mb 🙁

So I now have 64Mb of unused disk (and unusable! Stupid idea of BIOS boot to
have each partition starting before cylinder 1024. means my last partition
MUST be at least 128Mb).

Alternatively, could I make this into an extended partiton, with 2x64Mb
in that, or even 1x10Mb (root+tmp), 1x12Mb(swap=3xRAM),1x64 (/usr), 1x42 (think
of something!).

If so, how? I've read the begineers guide, and the FAQ (Mar 29th) mentions
hda5 as extended partition with a warning, but I couldn't find any info
on how to create this extended partition (NOTE: Dos fdisk won't do it because
it is above the 1024 cylinder limit), nor how to make this as root partition.

Help please! I am on 0.95c (yippeee! I successfully managed to get Linus's
patches to work! Well the kernel boots 🙂 ).

Thanks in advance.
--
Stephen Harris
[email protected] ...!uknet!axion!spuddy!sweh

* Meeeeow ! Call Spud the Cat on > +44 203 638780 < for free Usenet access *


[next article]
From: [email protected] (Ari Lemmke)

Subject: You're Linux FTP site admin?
Message-ID:
Date: 7 Apr 92 18:55:22 GMT
Sender: [email protected] (Usenet pseudouser id)
Reply-To: [email protected]
Organization: Helsinki University of Technology, Finland
Lines: 10
Nntp-Posting-Host: laphroaig.cs.hut.fi


Please join to the Linux Activists mailing list channel FTP.

More information by sending mail to

[email protected]

Reply should work.

arl


[next article]
From: [email protected] (System test)

Subject: gcc 2.1 woes
Message-ID: <[email protected]>
Date: 8 Apr 92 04:50:59 GMT
Sender: [email protected]
Organization: Stanford University, California
Lines: 12

I just installed gcc 2.1 and much to my chagrin can't get it to work.
I followed the instructions: removed all vestiges of 1.4, detarred, and
ran the script. But when I try to compile a simple test program I
get an error from ld that crt0.o is missing from /usr/lib/shared. And
indeed that link is missing. If I run gcc with the -static switch an
executable is created but it will not run.

Does anyone have a guess what I might have done wrong?

Don Taber
[email protected]
[email protected]


[next article]
From: [email protected] (Roger Books)

Subject: Re: compile kernel with gcc2.1
Keywords: linux, kernel , gcc2.1
Message-ID: <[email protected]>
Date: 2 Apr 92 19:36:26 GMT
References:
Reply-To: [email protected]
Organization: F.S.U. Dept of Physics
Lines: 16
News-Software: VAX/VMS VNEWS 1.3-4

In article , [email protected] (Jiansheng Zhao
) writes...
>Ok, I grabed gcc-as, gcc-ld etc from gccbin to compile the kernel. It went
>smoothly, until the Image building stage: error message:
>tools/build.c 28 : linux/fs.h : No such file or dirctory
>However, as I checked, there is linux/fs.h in /include. I cann't see why
>gcc2.1 wouldn't keep going.

I ran into the same problem. To solve it I made /usr/include/linux
subdirectory and linked all the stuff in the /usr/src/linux/include into it.

An odd one I ran into was when compiling ps095 it wouldn't accept some of
the include files, saying undeclared __(name).h in _text. I ended up
putting a full path name to the that particular include file.

Roger
[email protected]


[next article]
From: [email protected] (Roger Books)

Subject: Re: Help, can't compile 0.95a!
Message-ID: <[email protected]>
Date: 3 Apr 92 03:40:20 GMT
References: <[email protected]>
Reply-To: [email protected]
Organization: F.S.U. Dept of Physics
Lines: 17
News-Software: VAX/VMS VNEWS 1.3-4

In article , [email protected]
csu.edu (Mike Dodds) writes...
>[email protected] (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?),
>>and I get and error "cc1 got return code 16" when compiling fork.c!
>>Does this anything to to with the old gcc compiler? Or do I need new library
>>for 0.95a?
>Well,

Worked fine for me. I'm using a 486 with 4MB of ram and 4MB of swap space.
I also saw mentioned somewhere that Linus was converting to gcc 2.0 (now 2.1).
I am using 2.1

Roger

[email protected]


[next article]
From: [email protected] (Networking Research)

Subject: Relative speed of Linux
Message-ID: <[email protected]>
Date: 8 Apr 92 08:34:40 GMT
References: <[email protected]> artagnan.rutgers.edu>
Organization: Saint Cloud State University
Lines: 26
Nntp-Posting-Host: eeyore.stcloud.msus.edu

In article [email protected]
utgers.edu (Charles Hedrick) writes:
>features. A number of people claim that Linux feels fast, but since I
>haven't run any other Unix variant on my system, I have no good

Well, on a 386/25, no cache, Linux *feels* as fast as my Mach386 486/33C,
in serous processing the 486 blows away the 386 of course, but Linux has
a distinctly snappier OS response than Mach386. Same is true for SCO Open
DeathTop. I can definately report that on a 386/25, Linux outruns DeathTop
significantly. I'd guess that in the sense of operating system spped, Linux
will perform much better than any of the "big name" Unixes available for
the {3,4}86. They're not written right down on the hardware, and most of them
are bloated SVR4 beasts anyhow. (I do like Mach386 though, it's Mach+BSD4.3)

Keep adding features to Linux, surely, but keep the core small and stream-
lined. If things can stay small while adding networking and such, Linux
has the potential for being a real alternative for those who want extra
performance and don't require a megalithic company behind their OS.

Just as a sidenote. I just upgraded to 0.95a (I've got the patches to c,
just havn't doen them yet) from 0.12, and while I was impressed with 0.12,
I'm *very* impressed with 0.95a! My thanks to everybody who's been
contributing to the effort, and my profound respect to Linus for getting this
to this point in just a year!

-Mark Holden
[email protected]


[next article]
From: [email protected] (Hongjiu Lu -- Graduate Student)

Subject: Re: gcc 2.1 woes
Message-ID: <[email protected]>
Date: 8 Apr 92 06:42:51 GMT
References: <[email protected]>
Sender: [email protected] (USENET News System)
Organization: Washington State University
Lines: 27

In article <[email protected]> [email protected]
ord.EDU (System test) writes:
>I just installed gcc 2.1 and much to my chagrin can't get it to work.
>I followed the instructions: removed all vestiges of 1.4, detarred, and
>ran the script. But when I try to compile a simple test program I
>get an error from ld that crt0.o is missing from /usr/lib/shared. And

Very strange. Is inst2.x broken? At least, I can install it. Please do
a

sh -v inst2.x

There should be symbolic link from /usr/shared/lib/crt0.o to
/usr/lib/shared/crt0.o.

>indeed that link is missing. If I run gcc with the -static switch an
>executable is created but it will not run.
>

Are you using new ld and old kernel? If you are, read FAQ.

>Does anyone have a guess what I might have done wrong?
>
>Don Taber
>[email protected]
>[email protected]

H.J.


[next article]
From: [email protected] (Networking Research)

Subject: Re: Byte Unix Benchmarks.
Keywords: benchmark.tar.Z
Message-ID: <[email protected]>
Date: 8 Apr 92 10:38:27 GMT
References: <[email protected]>
Organization: Saint Cloud State University
Lines: 50
Nntp-Posting-Host: eeyore.stcloud.msus.edu

In article <[email protected]> [email protected] (Jon Tombs) writes:
[Info on the Byte Unix Benchmarks available for Linux]

I though it might be interesting to compare numbers with that of
a different Unix running on the same hardware, so I ran the benchmarks on
my 486/33c running Mach386.

[Sparc results deleted]

>486 33c [Linux]
>Arithmetic Test (type = double) 2541.7 1891.6 0.7
>Dhrystone 2 without register variables 22366.3 19281.7 0.9
>Execl Throughput Test 16.5 89.2 5.4
>File Copy (30 seconds) 179.0 44.0 0.2
>Pipe-based Context Switching Test 1318.5 3738.0 2.8
>Shell scripts (8 concurrent) 4.0 5.0 1.3
> =========
> SUM of 6 items 11.3
> AVERAGE 1.9


486/33C Running Mach386
Arithmetic Test (type = double) 2541.7 1968.8 0.8
Dhrystone 2 without register variables 22366.3 23723.6 1.1
Execl Throughput Test 16.5 176.2 10.7
File Copy (30 seconds) 179.0 155.0 0.9
Pipe-based Context Switching Test 1318.5 1558.5 1.2
Shell scripts (8 concurrent) 4.0 6.0 1.5
=========
SUM of 6 items 16.2
AVERAGE 2.7

Looking at this, the results would seem to indicate that Mach386 is faster,
contradicting something I myself said earlier. However, these factors can
be influenced by things like hard disk speed (maybe I have a faster one?),
caching, and such. But the thing I have never liked about the Byte benchmarks
is that they're too susceptable to being skewed by one factor. For instance,
the Execl throughput test in the above example throws the overall factor off
by a lot.

Something that I find telling is the Pipe based context switching, where Linux
is nearly 2 1/2 times faster than Mach 386. I'm willing to bet this is why
Linux feels faster.

Anybody else out ther got a 486/33c running SCO ODT, or one of the SVR4s
(i.e. Dell, Esix), or 386BSD (Jolitz), or BSD/386, or . . . who'd be willing
to contribute numbers from those OS's?

-Mark Holden
[email protected]


[next article]
From: [email protected] (Linus Benedict Torvalds)

Subject: linux-0.95c known bugs
Summary: well, it was alpha...
Message-ID: <[email protected]>
Date: 8 Apr 92 09:03:21 GMT
Organization: University of Helsinki
Lines: 58

Well, only one known bug so far, but a couple of problems. I thought I'd
mention them before anyone else does, and we'll call them "features" :^)

The BUG:

when using the readdir() system call, linux incorrectly doesn't let go
of the last buffer used for reading: this results in the buffer being
marked as used (and if you use readdir() heavily, the counter will
eventually wrap around, which might result in incorrect marking as
"not used"). This bug happily isn't easy to find: no current binary
uses the readdir() system call unless you have gotten your hands on
the new VFS gcc-2.1 release. Thanks to Remy Card for finding this
one.

Not too bad a bug though: the fix is very easy. Add a 'brelse(bh);'
in the minix_readdir() function, like this:

if (i) {
put_fs_long(de->inode,&dirent->d_ino);
put_fs_byte(0,i+dirent->d_name);
put_fs_word(i,&dirent->d_reclen);
return i;

should really be :

if (i) {
put_fs_long(de->inode,&dirent->d_ino);
put_fs_byte(0,i+dirent->d_name);
put_fs_word(i,&dirent->d_reclen);
+ brelse (bh);
return i;

That is, just add the brelse before the early return. Sorry for the
lack of real cdiffs - I'm not at home right now, and the above was
taken directly from the bug-report mail.

The problems:

I've had one report that the floppy-driver in versions 0.95x breaks
when accessing drive nr 2. It doesn't on my machine, but I'd
appreciate it if people would test it out, and mail me about any
problems. So far, only one report, but that's one too many.

0.95c doesn't correctly keep track of the 'rss' field in the
task-structure. A fix was already posted (and nothing breaks even if
you don't apply the fix - ps just gives slightly incorrect output)

And the expected troubles with the change of 'a.out.h' - the old gdb
doesn't recognize new executables etc.

As soon as I get my own sources cleaned up, I'll send out a new binary
for the 0.95c+ kernel to the ftp-sites. I've gotten a few mails from
people unable to recompile everything - either because of lack of
diskspace or some other problem. Tomorrow I'll put the new kernel image
on nic.funet.fi and tsx-11 - it's basically the 0.95c kernel + the above
bugfix + the lp-patches (somewhat edited by me).

Linus


[next article]
From: [email protected] (Haohui Wang)

Subject: Re: gcc 2.1 woes
Message-ID: <[email protected]>
Date: 8 Apr 92 07:27:30 GMT
References: <[email protected]>
Sender: [email protected]
Organization: UB
Lines: 21
Nntp-Posting-Host: lictor.acsu.buffalo.edu

I downloaded the GCC2.1 from tsx-11.mit.edu in the directory:
/pub/linux/incoming/GCCv2

Then moved gcc1.40 to another partition and installed 2.1, but it seems that
the cc1 came with it dies out on:

main()
{
printf("hello");
}

It says:"parse error before '('"

If I use the cc1 from 1.40 it compiles. But if I want to use gcc2.1 to link
it, ie - gcc junk.o, ld says "'---' unknown option" and prints a list of options
.

Did I miss something?

thanks

howie


[next article]
From: [email protected] (Bailey Brown)

Subject: ksh in linux?
Keywords: linux ksh
Message-ID: <[email protected]>
Date: 6 Apr 92 19:30:31 GMT
References: <[email protected]> @dartagnan.rutgers.edu> <[email protected]> <1992Apr1.182702.8
[email protected]> <[email protected]>
Organization: Intergraph Corp. Huntsville, AL
Lines: 7

Does linux have ksh (Korn Shell)? And if so, does it's ksh implementation
have the command line editing (vi mode)?

------------
Bailey Brown "Above all else, confusion reigns."
Intergraph Corporation
[email protected] Procol Harum


[next article]
From: [email protected] (Linus Benedict Torvalds)
Newsgroups: comp.os.linux,alt.os.linux

Subject: Re: Large disk partition
Message-ID: <[email protected]>
Date: 8 Apr 92 12:20:49 GMT
References: <[email protected]>
Organization: University of Helsinki
Lines: 40

In article <[email protected]> [email protected] (Stephen Harris)
writes:
>Is it possible for Linux to use a disk partition greater than 64Mb ?
>I have a 650Mb disk at 1283x16x63. DOS takes up the first 1023 clinders
>(the most it can have), leaving me 128Mb as a second partition. But when I
>try mkfs it moans for greater than 64Mb 🙁

Yes, the 64MB limit is still there in the filesystem: it seems the VFS
functions start to get complete soon, but I haven't written any actual
code for any other fs-type. I'll be adding an "extended-minix" type
soon - it will conform to the native-order-long minix filesystem type,
which allows bigger disks, but still has the 14-character name limit (in
case someone wonders why keep to minix types: it's easy to add, as I can
use most of the old routines anyway, and some misguided persons use
linux and minix side by side 🙂

>Alternatively, could I make this into an extended partiton, with 2x64Mb
>in that, or even 1x10Mb (root+tmp), 1x12Mb(swap=3xRAM),1x64 (/usr), 1x42 (think
>of something!).

As you report you have 0.95c running, it should be safe to use extended
partitions, but the old fdisk probably has problems with them still. At
least I haven't heard any reports of problems with extended partitions
yet, although I haven't heard any success-stories either.

There are other fdisks than the linux or dos ones - pfdisk etc is
reported to work. I haven't got any personal experience (I should be so
lucky to have the 1024-cylinder problem :). Anyway, linux 0.95c should
report any partitions it finds at bootup (along with some data on size
and position), and you should check the numbers before using extended
partitions the first time.

If the bootup-information seems to be ok, try out the extended
partitions (but back up just in case...).

>Help please! I am on 0.95c (yippeee! I successfully managed to get Linus's
>patches to work! Well the kernel boots 🙂 ).

Is everybody this surprised when one of my patches happen to work? :^)

Linus


[next article]
From: [email protected] (Todd Radel)

Subject: Shoelace booting problems
Summary: "Kernel panic" on booting
Message-ID: <[email protected]>
Date: 8 Apr 92 14:59:05 GMT
Organization: University of Delaware
Lines: 33

Hello, all! I spent most of yesterday backing up my hard disk, repartitioning,
restoring my DOS partition, then (finally!) making /dev/hda2 my Linux
partition. I installed Shoelace so that I could boot either partition. All
seemed to go OK--the DOS partition is still bootable, and if I select "2" from
the menu, Linux starts to boot. But what I get on the screen is this:

Load Image
Press to see SVGA-modes available or any other key to continue.

8 virtual consoles
4 pty's
Partition table ok.
Kernel panic: Cannot mount root
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

Argh! What am I doing wrong? I used the bootimage-0.95a as my /vmunix file.
I also tried patching it, per instructions in the FAQ, to use /dev/hda2 as
the root. No dice.

Anybody know what it is I'm doing wrong? Heellllp..... I really like Linux
so far. It's good to have a real operating system (e.g. Unix) running.
If you could respond by e-mail rather than posting I'd appreciate it.

Thanks in advance,
Todd Radel
Linux neophyte


--
* Todd Radel Sysop, Dickinson Domain BBS *
| CIS/English FidoNet: 1:150/160; VirtualNET @3023 |
| University of Delaware REGISTERED Alpha Colony VI! |
* "You don't need to be crazy to be a sysop, but it helps!" *


[next article]
From: [email protected] (Ryan Mitchell)

Subject: Newuser help?
Message-ID: <[email protected]>
Date: 8 Apr 92 16:08:04 GMT
Organization: Purdue University Computing Center
Lines: 19

This may be a really dumb question, so please forgive my ignorance...

I just installed linux on my HD yesterday and now I am interested
in getting gcc and some other programs to use with it. How do I
go about getting these to my Linux partition? It's easy enough
to transfer to DOS, but how can Linux access the DOS to move it
over? It seems that kermit or some other comm program would be
necessary in Linux?

Please e-mail, as I'm still wading through a mess of unread
articles.

Thanks,
Ryan
--
===============================/ /=============================*
========/ Ryan /======/ / [email protected] |
======/ Mitchell /======/ / [email protected] |
=========================/ /===================================*


[next article]
From: [email protected] (Todd Radel)

Subject: Re: Shoelace booting problems
Message-ID: <[email protected]>
Date: 8 Apr 92 16:29:22 GMT
References: <[email protected]>
Organization: University of Delaware
Lines: 17

Oops, stupid me. I accidentally wrote out the unpatched version of the
kernel to the tar file -- i.e., it was looking for the root filesystem
on the floppy disk. Sorry for the stupid question.

I might add that I now have GCC 2.1, VI, Kermit, tcsh, and many other
toys to play with. Linux does indeed kick mucho ass!

Three cheers to Linus for a great job!

-- Todd


--
* Todd Radel Sysop, Dickinson Domain BBS *
| CIS/English FidoNet: 1:150/160; VirtualNET @3023 |
| University of Delaware REGISTERED Alpha Colony VI! |
* "You don't need to be crazy to be a sysop, but it helps!" *


[next article]
From: [email protected] (Kevin Cummings)

Subject: Re: bash ? And Root disk software.
Message-ID: <[email protected]>
Date: 7 Apr 92 20:07:22 GMT
References: <[email protected]>
Sender: cummings@hammer (Kevin Cummings)
Organization: Prime Computer R&D
Lines: 50

In article <[email protected]>, [email protected]
om (David.L.) writes:
> Where do I get am image or sources for bash ? do I need to get the earlier roo
t
> image ? (.12 , .11 ) ? I'm using 0.95a.

My hard disk with 0.95 installed over 0.12 crashed. After getting a new
drive, I found myself having to install from scratch. So I went and
got the latest root disk 0.95a. When I found that BASH was not on
it, I mounted the 0.95 root disk and copied both BASH (/bin/sh) and the
PASSWD executable from it to my new system disk (/usr/local/bin/bash,
and /bin/passwd).

In a related issue, When I formatted my new 200MB IDE drive under DOS,
I created one DOS primary (bootable) partition, and put the rest
of my DOS logical drives in the DOS extended partition. This leaves
me two primary partitions for LINUX, which I configured using the
pfdisk on the 0.95a root disk. However, the infamous extended partition
bug kept both the 0.95 and 0.95a fdisk from correctly reading my
partition table on /dev/hda. This resulted in my inability to
know for sure which logical linux devices I needed to run mkfs and
mkswap on! I had to get the fdisk from the 0.12 root disk, and
create a /dev/hd0 link to /dev/hda in order to find out which
was /dev/hd3 (/dev/hda3) and which was /dev/hd4 (/dev/hda4).

patches to the kernal are NO GOOD if you cannot be sure which
devices to make in order to put your system on hard disk!

Based on this, someone had better put together a BOOTIMAGE and
a ROOTIMAGE for 0.95c if we want users to be able to install
LINUX in systems that do not already have LINUX partitions!
You just can't expect non-hacker users to have to have 3 rootdisks
around to get all the necessary software in one place (never mind
the changing of the device names too. Sigh).

When I get done re-creating my system, I'll be happy to post a list
of the software I had to download, in addition to what was on the
0.95a root disk, and install in order to make my system useable
to me.

=================================================================
Kevin J. Cummings Prime Computer Inc.
20 Briarwood Road 500 Old Connecticut Path
Framingham, Mass. Framingham, Mass.

InterNet: [email protected]
UUCP: uunet!primerd.Prime.COM!cummings

Std. Disclaimer: "Mr. McKittrick, after careful consideration,
I've come to the conclusion that your new
defense system SUCKS..." -- War Games
=================================================================


[next article]
From: [email protected] (Alan B Clegg)

Subject: Unfortunate, but true....
Message-ID: <[email protected]>
Date: 8 Apr 92 18:04:09 GMT
Sender: [email protected]
Organization: Concert Network -- Internet Operations Group
Lines: 12

I will *NOT* be producing the "ABC Release of Linux .95a" before I go on
vacation. I am currently over-whelmed by work, and won't have free-time
before leaving on Saturday morning.

In the off-chance that I do get some time, it will be devoted to getting
my scsi disk back working again, as that is where the sources that I have
so-far compiled are living... 8-(

-abc
--
[email protected] Alan Clegg - Network Programmer
KD4JML (just my luck!) MCNC -- Center for Communications


[next article]
From: [email protected] (Herbert M. Zinn)
Newsgroups: comp.os.linux,alt.os.linux

Subject: lots o problems, many to choose from
Summary: GCC2.1 ld errs, kern2.1patch, serial ports, bad juju!
Message-ID: <[email protected]>
Date: 8 Apr 92 16:36:04 GMT
Sender: [email protected] (Mr. News)
Organization: The Cabal of the Infinite bounded Spoo
Lines: 31


Hi, this is my first post here so if this has already been dealt with
in a FAQ somewhere please let me know. I recently grabbed GCC2.1 beta, the
shared libs, and the kernel patches from tsx-11 and am having problems with
the libs. When I run the install.sh in /usr/lib/shared/obj, it invokes the
2.1 GCC with options -s -o on the make.a file. this (i assume) then invokes
the linker, which promptly bombs with "unknown option: ---" followed by a
usage message. Ive tried invoking it by hand and it always seems to generate
this error message. I'm using the ld that came in binutils.tar.Z, but I noticed
its only a couple hundred bytes different, could I perhaps have grabbed the
wrong one?
Another GCC2.1 problem: when I install the 2.1kern patches (aside from
the somewhat odd path [/mnt/linux/... for the kernel sources] that it puts in
every makefile) it generates parse errors on a lot of the linux kernel headers.
This may have been covered elsewhere; are the /usr/src/linux/include
files duplicated in the old gcc1.4 /usr/include? are they in gcc2.1 includes?
Finally, until I get around to digging through the famous ralph brown
interrupt list, has anyone else found the need to use more than 2 serial ports
at once? I don't know enough about the coding of comm routines myself, but,
since sharing IRQ's is basically nonfunctional for concurrent IO in DOS,
I have my third serial port (3e8) set to IRQ5. Anyone have patches || diffs
|| help in general with changing the serial drivers to use it? (Digging through
the includes didn't seem to help 😉 )
Thanks in advance for any help, responses, email, flames, et cetera,
and a great big THANKS! to Linus and all the other code wizards for putting
together a great package!
Herb
_____________________________________________________________________________
|Herb M. Zinn INET: [email protected] USPS: 6211 N. Artesian Ave. |
|"Standing is harder than moving." Chgo., IL. 60659-2828 |
|_______________-Moshe Feldenkrais___________ISDN:_312-274-8149_____________|


[next article]
From: [email protected] (Dennis Petrocelli)
Newsgroups: comp.os.linux,alt.os.linux

Subject: Scheme, Lisp for Linux?
Summary: Ported either?
Keywords: scheme lisp linux-ports
Message-ID: <[email protected]>
Date: 6 Apr 92 07:53:22 GMT
Sender: [email protected]
Reply-To: [email protected] (Dennis Petrocelli)
Followup-To: comp.os.linux
Organization: Boston University
Lines: 14


Is anyone working on porting any version of scheme or lisp to
linux? I would be VERY appreciative of anyone's attempts to do so. I
have successfully compiled a C implementation of Scheme, siod-2.7
(Scheme in One Defun, from world.std.com), with gcc-1.4 on linux .95a.
I am hoping that someone might have tried to port a more substantial
version, such as scheme 7.1, from altdorf.ai.mit.edu.

Dennis Petrocelli
[email protected]

BTW, way to go LINUS, et al !!!!!!!!!!!!!!!!!

OS/2 - half the OS at twice the price!


[next article]
From: [email protected] (P Thomson)

Subject: floppies Help !
Message-ID: <[email protected]>
Date: 8 Apr 92 20:05:36 GMT
Organization: Edinburgh University
Lines: 14



Hi , can someone help here please. I got hold of the images for boot and
root and tried to run Linux from a 1.44Mb floppy.

The program starts to load ie Loading..........
prompts for a svga setting then bombs out. This happens no matter what
mode I try to set ( pressing space has the same effect)

System config 386DX 20Mz
91Mb Segate/Imperus Hard drive (SCSI)


Thanks in advance Peter.


[next article]
From: [email protected] (Steve McCalla)

Subject: PS/2 Systems and Linux. was: What is linux?
Keywords: linux
Message-ID: <[email protected]>
Date: 8 Apr 92 18:58:32 GMT
References: <[email protected]> <1992Apr1.182702.8464@colo
rado.edu> <[email protected]>
Sender: [email protected]
Reply-To: [email protected] (Steve McCalla)
Lines: 25
Nntp-Posting-Host: joego

The following is a recollection of news items posted when Minix was ported to
the PS/2 back in the dark ages, so take it with a grain of salt.

The main (only?) difference between microchannel machines and ISA bus machines
is the way interrupts happen. The microchannel has level-triggered interrupts
as opposed to edge-triggered interrupts on the ISA bus. This means that
multiple devices may share a single interrupt line. When an interrupt comes
in on a shared line, the devices sharing that line must be polled to find out
what caused the interrupt. That device must be serviced before sending the EOI
out, otherwise you end up getting another interrupt from the same device.
With edge-triggered interrupts, you can only have a single device per interrupt
line so this is never a problem; you can simply send the EOI and then
handle the
device that caused the interrupt. Naturally, different devices must be serviced
in different ways; I seem to recall a SENSE command being sent to the floppy
device in Minix as being sufficient for having it remove the intterupt request.
I don't know what the commands are for other devices. I believe they found a
way to set up Minix such that it would run on either bus, but I don't have
the details. (They = Bruce Evans, Steve Ackerman, and a host of others).
Hope these ramblings help.

- stevem
"#include my opinions are my own, etc"
--
"It's pushbutton automation. First you push 'v', then you push 'i'...:)


[next article]
From: [email protected] (David Buckley)

Subject: Job Ctrl && $$->programmers (How?)
Keywords: jobs, money, bash, happy
Message-ID: <[email protected]>
Date: 7 Apr 92 08:24:28 GMT
Sender: [email protected]
Reply-To: [email protected] (David Buckley)
Organization: Computer Science Department, Boston University, Boston, MA, USA
Lines: 21


Is there some version of jsh, or some implementation of the
functions fg, bg, %[n] etc... I didn't find them on the version of
bash I am using. Are there several versions of bash? How do we get
access to job ctrl? I know these exist from reading previous
postings, are they hiding or should I chmod u+rwx mybrain.out?

Also, is there some list of developers-port|ers so that
it might be easyer to direct financial contributions to this venture?
I haven't seen any postings with regards to how and to who money
should be donated, and I think it would be interesting to be able
to direct donations for specific implementations (our unix favourites),
beyond Linux itself. Such a posting (I feel) would be greately
appreciated.

I am very impressed with the work being done!
(As a newbie-linuxite I would love to contribute...)


David Yates Buckley | Disclaimer: Recursive lunacy is
[email protected], @csa.bu.edu | employed in the most trivial tasks...


[next article]
From: [email protected] (Maad Al-Dabbagh)

Subject: Is Linux real-time ?
Keywords: real-time os
Message-ID:
Date: 8 Apr 92 20:50:19 GMT
Sender: [email protected] (Maad Al-Dabbagh)
Reply-To: [email protected] (Maad Al-Dabbagh)
Organization: University of Southern California, Los Angeles, CA
Lines: 6
NNTP-Posting-Host: sal-sun118.usc.edu


Is linux a real-time os ? please advise. I am working on an os called
Lynx OS , and it is supposed to be rea-time. I am wondring if it is the
same os or not.

maad


[next article]
From: [email protected] (Mike Ching)

Subject: Re: Linux Drivers and Kernel question.
Message-ID: <[email protected]>
Date: 8 Apr 92 17:34:23 GMT
References:
Sender: [email protected] (NetNews)
Organization: Advanced Micro Devices; Sunnyvale, CA
Lines: 10
Nntp-Posting-Host: brahms

In article [email protected] (Erik Fichtner) writes:
>Is there an ethernet driver in the works? along with patches to ka9q to
>make ethernet work?
>
Does a hardware independant approach like the DOS packet drivers make
sense for Linux? Any hope that the existing packet drivers can be used
with a protected mode widget like the one for Windows?

Mike Ching
(not a kernal hacker yet)


[next article]
From: [email protected] (Drew Eckhardt)

Subject: Re: floppies Help !
Message-ID: <[email protected]>
Date: 9 Apr 92 00:18:27 GMT
References: <[email protected]>
Sender: [email protected] (The Daily Planet)
Organization: University of Colorado at Boulder
Lines: 27
Nntp-Posting-Host: romeo.cs.colorado.edu

In article <[email protected]> [email protected] (P Thomson) writes:
>
>
>Hi , can someone help here please. I got hold of the images for boot and
>root and tried to run Linux from a 1.44Mb floppy.
>
>The program starts to load ie Loading..........
>prompts for a svga setting then bombs out. This happens no matter what
>mode I try to set ( pressing space has the same effect)
>
>System config 386DX 20Mz
> 91Mb Segate/Imperus Hard drive (SCSI)
>
>
> Thanks in advance Peter.

Several things :
1. The "normal" distribution kernel does not support SCSI.
2. The SCSI kernel probably tries to mount root off of the first
partition on the first "normal" hard disk,
and will panic when root can't be mounted.
3. The scsi kernel ONLY supports seagate and Ultrastor 14F
hosts adapters. Adaptec support is finished,
and will be released RSN.

4. "bombs out" doesn't mean to much. What kernel panic is created,
or does it just stop?


[next article]
From: [email protected] (gary a moyer)

Subject: problems compiling kernel -- fatal signal 6
Keywords: kernel
Message-ID: <[email protected]>
Date: 7 Apr 92 23:50:06 GMT
Sender: [email protected]
Lines: 9

With a little luck and insight from some users, the mystery
surrounding the fatal signal 6 dilemna has been solved.
Evidently the makefile in linux/kernel contains the line
GCC_OPT = -fcombine-regs commented out due to problems (or
possibly a fix) for persons using gcc2.0. This, however,
makes gcc1.4 choke and die. Just remove the comment delimiter
and blamo -- it compiles.

[email protected]


[next article]
From: [email protected] (gary a moyer)

Subject: shoelace -- errors at boot from partition
Keywords: shoelace
Message-ID: <[email protected]>
Date: 7 Apr 92 23:53:04 GMT
Sender: [email protected]
Lines: 8

I recently got linux to boot from the second partition via
shoelace. It boots fine and everything, however, I get some
interesting warning messages. Anyone else had this problem?

BTW -- kudos for the port to Linux, that floppy grinding away
at bootup was s l o w......

[email protected]


[next article]
From: [email protected] (Brian Mcbee)

Subject: Seagate SCSI *almost* working...
Message-ID: <[email protected]>
Date: 9 Apr 92 04:27:03 GMT
Sender: [email protected]
Organization: CS Dept. Oregon State University, Corvallis, Oregon.
Lines: 32
Nntp-Posting-Host: nyssa.cs.orst.edu

I've been using linux now for a couple of months, and I decided to
try and get an old seagate st-277n I had laying around working
with it. I got an ST-01 controller, and ftp'd the linux95a
scsi patches from headreset.colorado.edu. Hacked them into
my Linux .95c system. Had to make one small change to
seagate.c, as my signature "SEAGATE SCSI BIOS REVISION"
is at offset 17 rather than the 16 indicated in the file.
Upon bootup, the system does find the host adapter, and 1
scsi disk (yay!). Where do I go from here? I get several
messages here, but I don't know if they are normal or not.
and how do I partition the drive? Here is what I see:

...
1 host adapters detected
Detected scsi disk at host 0, ID 0, lun 0
Detected 1 disks, 0 tapes total
Warning: SCSI device at host 0, id 0, lun 0 failed READ CAPACITY
status=5345c
Sense not available
Block size assumed to be 512 bytes, disk size 1 GB
Cannot read partition table for sd 0

Is any of the above normal for the seagate? Or is it totally
hosed? Or is it just a matter of getting the partition info
on there? Any help appreciated.


--
Brian McBee [email protected] or [email protected]
"The strongest reason for the people to retain the right to keep and bear
arms is, as a last resort, to protect themselves against tyranny in
government." Thomas Jefferson


[next article]
From: [email protected] (Drew Eckhardt)

Subject: Re: Seagate SCSI *almost* working...
Message-ID: <[email protected]>
Date: 9 Apr 92 07:16:00 GMT
References: <[email protected]>
Sender: [email protected] (The Daily Planet)
Organization: University of Colorado at Boulder
Lines: 51
Nntp-Posting-Host: romeo.cs.colorado.edu

In article <[email protected]> [email protected] (Brian
Mcbee) writes:
>I've been using linux now for a couple of months, and I decided to
>try and get an old seagate st-277n I had laying around working
>with it. I got an ST-01 controller, and ftp'd the linux95a
>scsi patches from headreset.colorado.edu. Hacked them into
>my Linux .95c system. Had to make one small change to
>seagate.c, as my signature "SEAGATE SCSI BIOS REVISION"
>is at offset 17 rather than the 16 indicated in the file.
>Upon bootup, the system does find the host adapter, and 1
>scsi disk (yay!). Where do I go from here? I get several
>messages here, but I don't know if they are normal or not.
>and how do I partition the drive? Here is what I see:
>
>...
>1 host adapters detected
>Detected scsi disk at host 0, ID 0, lun 0
>Detected 1 disks, 0 tapes total
>Warning: SCSI device at host 0, id 0, lun 0 failed READ CAPACITY
>status=5345c
>Sense not available
>Block size assumed to be 512 bytes, disk size 1 GB
>Cannot read partition table for sd 0
>
>Is any of the above normal for the seagate? Or is it totally

It detected your host adapter, disk drive, etc. Not being able to
read capacity is normal for SOME drives, most namely older Seagates
with brain dead firmware.

>hosed? Or is it just a matter of getting the partition info

And, it's also hosed. If compiled with gcc2.x, the seagate SCSI
driver will become unhappy. It has to do with how gcc decided
to push bx and si onto the stack, changing where the len and
data variables really were and where the assembler code thought
they were.

>on there? Any help appreciated.

A new release of the SCSI drivers will be made "RSN" - hopefully
tonight. This should correct the stack frame problem caused by
gcc2, add adaptec support, and a few other things.

It will be put up on

headrest.woz.colorado.edu

You might want to join the SCSI mailing list -
mail to

[email protected]


[next article]
From: [email protected] (stephen.p.hill)

Subject: backup problems
Keywords: cpio tar zip etc
Message-ID: <[email protected]>
Date: 9 Apr 92 10:37:37 GMT
Organization: AT&T
Lines: 51


Has anyone had much luck in backing up their systems? I
have heard so many things about problems with different
versions of tar ( and I have no clue which version I
ended up with - I got at least 3 versions in the files
I ftp'e ) that I didn't want to use it for my backup.

I got cpio, zip and zoo. I never learned to use zoo, so
I tried zip first. It works for text files, but when
I try to zip /bin, it only does 3 or 4 files before
printing "EINVAL - error making temp file".

cpio seemed to work fine, so I made 6 cpio files that would
each fit (compressed) onto a 1.4 M floppy and used dd
to write them onto a raw dos-formatted floppy. I also
un-cpio'd them into /tmp for a confidence check.

I booted Linux from floppy, re-partitioned my HD (needed
more DOS space), installed the basic Liunux to HD, and tried
to restore from my backups. I read the compressed cpio files
in from the floppies, uncompressed them and attempted to
un-cpio them. It would give me the first few files from
any archive fine, but then it would tell me that it was
"skipping xxx bytes of garbage" for each file after that.

Well, my heart nearly stopped because I had spent way too
many hours downloading that stuff at 2400 baud. I tried
the cpio again with the exact same results. I tried it
again and accidentally left off the "u" flag. This made
it skip the files that it had been able to extract before,
and it was able to extract another 5 or 6 files before it
started "skipping xxx bytes of garbage" again. Repeating
this a few times allowed me to recover all the files.

I then did another "fsck". The filesystem had been fine
before I started the restore, but it now gave me several
pages of errors. It was very late and I don't quite remember
what they were but it was something about a zone being
marked in use but wasn't really used. I ran "fsck -a"
and it fixed the errors.

Anyone got any ideas what went wrong? It looks to me
like there is a bug in the gnu cpio port. Does anyone

have a magic method for backing up their system?

BTW, I am using linux 0.95a on a AMD 386-40 with SIS chipset
and Diamond SS VGA. Works great!


Stephen Hill
[email protected]


[next article]
From: [email protected] (Linus Benedict Torvalds)

Subject: Re: Job Ctrl && $$->programmers (How?)
Keywords: jobs, money, bash, happy
Message-ID: <[email protected]>
Date: 9 Apr 92 14:49:03 GMT
References: <[email protected]>
Organization: University of Helsinki
Lines: 73

In article <[email protected]> [email protected] (David Buckley) writes:
>
> Is there some version of jsh, or some implementation of the
>functions fg, bg, %[n] etc... I didn't find them on the version of
>bash I am using. Are there several versions of bash? How do we get
>access to job ctrl? I know these exist from reading previous
>postings, are they hiding or should I chmod u+rwx mybrain.out?

Yes, there are several versions of bash existing - some of them are
/very/ old. I don't know if the original bash.Z is still somehwre to be
found, but it was not too long ago, and that is the version I used under
0.01, long before linux had job control. That version certainly doesn't
support job control.

Versions of bash that support job control (including the builtins fg,
bg, kill, jobs etc) are available on older root-disks, notably the 0.95
(not "a") rootdisk. I don't know where to find a "standalone" version
of bash - I know somebody ported bash 1.12 (the one on the rootdisk is
1.11 patch-level 1, although I think "help" returns 0.00.1).

You might want to check out the abc-release when it gets out - right now
the binaries under linux are a major mess (different versions under the
same name etc...), but this will eventually be sorted out, I hope.

> Also, is there some list of developers-port|ers so that
>it might be easyer to direct financial contributions to this venture?
>I haven't seen any postings with regards to how and to who money
>should be donated, and I think it would be interesting to be able
>to direct donations for specific implementations (our unix favourites),
>beyond Linux itself. Such a posting (I feel) would be greately
>appreciated.

I have never felt anybody needed to pay for linux - I didn't want to
make it share-ware, and I never wanted to make it commercial. The
biggest reason I started on linux at all was that I had no good unix and
no money - I wanted a totally free unix (not "GNU-free", but "no-money"
free - the "GNU-free" came later). I still have no money, but at least
I've got the unix 🙂

In fact, the earliest copyrights expressly forbid /any/ money at all
changing hands due to linux - the above was the reason. As the system
actually became useable, I had to change my mind - I don't mind people
making money off linux, as long as it essentially is available for free
as well, which it is.

I still feel strongly about the money-thing though - I don't want to
encumber the "official" version with any kind of pleas for money. If I
had something like "if you enjoy this package, please send $10 to xxx"
it might scare away people from then sending other feedback (bugreports
etc), as they didn't send the money... As it is, linux has no special
"registered users" which get special treatment by me - if you are
porting something important, I might give your troubles more thought
than the average user, but I try to answer every problem that gets
reported (even if it's just a "sorry, can't help this one" or "it's a
known problem, I don't know the cure", "hardware dependant").

One thing I would appreciate is people sending me postcards from all
over the world: I don't want to be a new Craig Shergold, but it might be
fun to see where linux is used :^). Postcards (with least name and
location written in for reference) could be sent to:

Linus Torvalds
Pietarinkatu 2 A 2
00140 Helsinki
Finland

and maybe I'll wallpaper my room with them :). Or maybe not.

Linus

PS. If you have contacts with hardware suppliers, and would like to get
things like multi-serial cards/tape streamers etc supported... I just
might be persuaded to try it out 🙂


[next article]
From: [email protected] (Joseph Knapka)

Subject: gcc-2.1 header files
Message-ID: <[email protected]>
Date: 7 Apr 92 14:55:23 GMT
Organization: University of Georgia, Athens
Lines: 20

H.J. asked me to forward this exchange to the newsgroup:

I said:
>> Kernel build of .95c went fine, but when compiling ps, I get
>> "undefined symbol __ctype referenced from text segment" while linking
>> ps.o. This is with very latest gcc-2.1, libraries, and binutils from
>> banjo. I linked xxx/src/linux/lib/ctype.o and got it to compile, and
>> (seems that) everything works. Did I do the right thing?
>>
>

H.J. said:
>No. Please get my 2.1ps.tar.Z, 2.1kernel.tar.Z and apply Linus's second
>patch first. When compile any application, do not use ANY header files
>from ./linux/include.

Joseph

P.S.: but I'd already applied .95c, 2.1kernel, and ps patches. Still
waiting for a reply...


[next article]
From: [email protected] (Joseph Knapka)

Subject: __ctype undefined symbol error
Message-ID: <[email protected]>
Date: 7 Apr 92 17:55:22 GMT
References: <[email protected]>
Organization: University of Georgia, Athens
Lines: 27

Hi, all. I was having trouble compiling ps after applying Linus' and
H.J.'s patches to the .95a kernel. Turned out I hadn't rooted out all
traces of gcc-1.40 --- that part of the FAQ is to be taken very seriously!
Here's H.J.'s message:

__ctype is from the header files of gcc 1.40. Please backup and remove
any files related to the previous compilers before you install gcc 2.1.
There is no __ctype anywhere in the header files for gcc 2.1.

BTW, they are in my FAQ. PLEASE READ IT WORD BY WORD AND FOLLOW THE
INSTRUCTIONS.

Could you please forward my msg to news groups?

H.J.
--
School of EECS Internet: [email protected]
Washington State University BITNET: [email protected]
Pullman, WA 99164 Phone: (509) 335-6470 (O)
USA (509) 334-6315 (H)

Thanks for your help, H.J. -- the fault was entirely mine.

For those who are installing 2.1 for the first time, "Remove all other
compilers" means "rm +recursive /usr/include /usr/lib /usr/local/lib
(and anything else that looks suspicious)." Took me a while to catch
on to that 🙂


[next article]
From: [email protected] (Joseph Knapka)

Subject: in gcc2.1
Message-ID: <[email protected]>
Date: 8 Apr 92 13:39:22 GMT
Organization: University of Georgia, Athens
Lines: 10

Has anyone had trouble using assert under gcc-2.1? The problem I'm
having (and it only takes a trivial bit of code to demonstrate) is
that while #define-ing NDEBUG turns off assertions as advertised, it
seems that subsequently #undef-ig NDEBUG and re-#include-ing
*fails* to turn them back on. Am I hallucinating, or is
there really a problem?

Thanks for any assistance,

Joseph


[next article]
From: [email protected] (Dennis Flaherty)

Subject: Patch to preset the SVGA mode
Message-ID: <[email protected]>
Date: 9 Apr 92 16:49:27 GMT
Organization: Miles, Inc. Diagnostics Divis.
Lines: 92


If you are tired of hitting RETURN and selecting the
same SVGA mode every time you boot Linux, apply this
patch and rebuild the kernel. It allows you to preset
the mode in the Makefile, so that the setup program
proceeds as if you had hit RETURN and your favorite number.

Dennis


*** Makefile~ Mon Apr 6 15:41:45 1992
--- Makefile Thu Apr 9 10:39:19 1992
***************
*** 44,49 ****
--- 44,53 ----
AS86 =as86 -0 -a
LD86 =ld86 -0

+ # If you want to preset the SVGA mode, uncomment the next line
+ # and set SVGA_MODE to whatever number you want. I use #2.
+ #SVGA_MODE= -DSVGA_MODE=\#2
+
AS =as
LD =ld
#LDFLAGS =-s -x -M
***************
*** 124,130 ****
$(LD86) -s -o boot/setup boot/setup.o

boot/setup.s: boot/setup.S include/linux/config.h
! $(CPP) -traditional boot/setup.S -o boot/setup.s

boot/bootsect.s: boot/bootsect.S include/linux/config.h
$(CPP) -traditional boot/bootsect.S -o boot/bootsect.s
--- 128,134 ----
$(LD86) -s -o boot/setup boot/setup.o

boot/setup.s: boot/setup.S include/linux/config.h
! $(CPP) -traditional $(SVGA_MODE) boot/setup.S -o boot/setup.s

boot/bootsect.s: boot/bootsect.S include/linux/config.h
$(CPP) -traditional boot/bootsect.S -o boot/bootsect.s
*** boot/setup.S~ Mon Apr 6 13:32:48 1992
--- boot/setup.S Thu Apr 9 10:36:55 1992
***************
*** 245,250 ****
--- 245,251 ----
mov es,ax
lea si,msg1
call prtstr
+ #ifndef SVGA_MODE
flush: in al,#0x60 ! Flush the keyboard buffer
cmp al,#0x82
jb nokey
***************
*** 259,264 ****
--- 260,266 ----
mov ax,#0x5019
pop ds
ret
+ #endif
svga: cld
lea si,idati ! Check ATI 'clues'
mov di,#0x31
***************
*** 497,502 ****
--- 499,507 ----
call prtstr
pop si
add cl,#0x80
+ #ifdef SVGA_MODE
+ nonum: mov al,SVGA_MODE ! Preset SVGA mode
+ #else
nonum: call getkey
cmp al,#0x82
jb nonum
***************
*** 508,513 ****
--- 513,519 ----
zero: sub al,#0x0a
nozero: sub al,#0x80
dec al
+ #endif
xor ah,ah
add di,ax
inc di

--
Dennis T. Flaherty [email protected]
Miles, Inc. [email protected]
Diagnostics Division
My doctor says Mylanta, but my boss says Alka-Seltzer!


[next article]
From: [email protected] (Todd Radel)

Subject: Wish list...
Summary: Things I'd like to have for Linux
Message-ID: <[email protected]>
Date: 9 Apr 92 17:42:25 GMT
References: <[email protected]>
Organization: University of Delaware
Lines: 17

Does anyone have any of the following compiled for Linux?

- System information (e.g. du, df, users, uptime, etc.)
- Mail system (mail, rmail, sendmail, etc.)
- Any type of BBS or Muck software

Also, how close is the port of X-Windows coming to reality? If someone gets
X working, I'll format my DOS partition!

Cheers,
Todd

--
* Todd Radel Sysop, Dickinson Domain BBS *
| CIS/English FidoNet: 1:150/160; VirtualNET @3023 |
| University of Delaware REGISTERED Alpha Colony VI! |
* "You don't need to be crazy to be a sysop, but it helps!" *


[next article]
From: [email protected] (David Engel)

Subject: Re: in gcc2.1
Message-ID: <[email protected]>
Date: 9 Apr 92 17:58:59 GMT
References: <[email protected]>
Organization: Optical Data Systems, Inc.
Lines: 16

[email protected] (Joseph Knapka) writes:
: Has anyone had trouble using assert under gcc-2.1? The problem I'm
: having (and it only takes a trivial bit of code to demonstrate) is
: that while #define-ing NDEBUG turns off assertions as advertised, it
: seems that subsequently #undef-ig NDEBUG and re-#include-ing
: *fails* to turn them back on. Am I hallucinating, or is
: there really a problem?

You're right! You need to delete the #ifndef _ASSERT_H/#define _ASSERT_H
and associated #endif stuff from include/assert.h.

David
--
David Engel Optical Data Systems, Inc.
[email protected] 1101 E. Arapaho Road
(214) 234-6400 Richardson, TX 75081


[next article]
From: [email protected] (Joel M. Hoffman)

Subject: Another bug? [was Re: linux-0.95c known bugs]
Message-ID: <[email protected]>
Date: 9 Apr 92 20:02:44 GMT
References: <[email protected]>
Sender: [email protected] (USENET News system)
Organization: University of Maryland, College Park
Lines: 14
Nntp-Posting-Host: next.wam.umd.edu

In article <[email protected]> [email protected]
(Linus Benedict Torvalds) writes:
>Well, only one known bug so far, but a couple of problems. I thought I'd
>mention them before anyone else does, and we'll call them "features" :^)

More than once, I've gotten a message about mis-linked processes. I
usually get it when I'm compiling from one VC and using Kermit
interactively on another, but most of the time I'm compiling and using
Kermit....

I'm using a patched .95a kernel, GCC 1.40, and the kermit binary from
tsx (running on a Dell 325D -- '386DX with 2M).

-Joel
([email protected])


[next article]
From: [email protected] (Hongjiu Lu)

Subject: Re: gcc-2.1 header files
Message-ID: <[email protected]>
Date: 9 Apr 92 18:54:02 GMT
References: <[email protected]>
Sender: [email protected] (USENET News System)
Organization: Washington State University
Lines: 83

In article <[email protected]>, [email protected] (
Joseph Knapka) writes:
|> H.J. asked me to forward this exchange to the newsgroup:
|>
|> I said:
|> >> Kernel build of .95c went fine, but when compiling ps, I get
|> >> "undefined symbol __ctype referenced from text segment" while linking
|> >> ps.o. This is with very latest gcc-2.1, libraries, and binutils from
|> >> banjo. I linked xxx/src/linux/lib/ctype.o and got it to compile, and
|> >> (seems that) everything works. Did I do the right thing?
|> >>
|> >
|>
|> H.J. said:
|> >No. Please get my 2.1ps.tar.Z, 2.1kernel.tar.Z and apply Linus's second
|> >patch first. When compile any application, do not use ANY header files
|> >from ./linux/include.

That is not totally right. Here again.

I am enclosing a program, conv. You run it under ./linux. That program will ensu
re only one set of header files is needed for compiling appplications and
kernel.

Before you run it, you have to have gcc 2.1 beta, mv, cp and sed
installed on your system.

Before you do anything, please backup gcc 2.1 and kernel sources. Here
goes.

Step 1: Apply Linus's second patch and then my 2.1kernel.tar.Z.

Step 2: Read conv and understand what it is doing. Then go to the
toplevel of kernel source tree and do

conv

Step 3: Edit the toplevel Makefile and add -DPRE_GCC_2 to CFLAGS.
You have to get rid of "-nostdinc -Ixxxx" stuff since
./include has been moved to /usr/include.

Step 4: Type

make

If you have any problem, please let me know.

H.J.
------------CUT HERE------
#!/bin/sh
# you have to run this script under ./linux
INCS="a.out.h const.h asm linux"

# move ctype.h and string.h in kernel source, which are different from
# gcc 2.1 to ./include/linux.
mv ./include/ctype.h ./include/string.h ./include/linux

# copy those header files which don't exist in gcc 2.1 to /usr/include
for h in $INCS
do
cp -r ./include/$h /usr/include/$h
done

# move ./include to ./include.kernel
mv ./include ./include.kernel

# change
#include
#include
# to
#include
#include
#
DIRS="boot fs init kernel lib mm"
for d in $DIRS
do
for f in $d/*.c $d/*.h $d/*/*.c $d/*/*.h
do
if [ -f $f ]
then
sed 's,,,' $f | sed 's,,
,' > 1
mv 1 $f
fi
done
done


[next article]
From: [email protected] (Stephen Harris)
Newsgroups: comp.os.linux,alt.os.linux

Subject: Re: Large disk partition
Message-ID: <[email protected]>
Date: 9 Apr 92 15:47:26 GMT
References: <[email protected]> <[email protected]
lsinki.FI>
Organization: Spuddy's Public Usenet Domain
Lines: 18

In article <[email protected]> [email protected]
I (Linus Benedict Torvalds) writes:
>There are other fdisks than the linux or dos ones - pfdisk etc is

Obviously, I have pfdisk (from the 95a root disk), but I didn't think this
could create extended partitions. I'll recheck though. Thanks.

>Is everybody this surprised when one of my patches happen to work? :^)

Dunno about other people, but I was amazed that I had managed to put enough
stuff together in such a way that the patch procedure worked for me!
This is my first experience at beta testing, so everything I do correct is
amazing 🙂 Dunno why people make so much fuss about problems compiling
the kernel with gcc 1.40 🙂
--
Stephen Harris
[email protected] ...!uknet!axion!spuddy!sweh

* Meeeeow ! Call Spud the Cat on > +44 203 638780 < for free Usenet access *


[next article]
From: [email protected] (Karl R. Buck)

Subject: Where are expr and mkdir binaries?
Message-ID:
Date: 10 Apr 92 01:02:13 GMT
Organization: Kansas State University
Lines: 3
NNTP-Posting-Host: matt.ksu.ksu.edu

Subject says it all. Please email and I'll followup. Thanks.
--
Karl Buck, [email protected] 🙂


[next article]
From: [email protected] (S3679988)

Subject: 0.95a
Message-ID: <[email protected]>
Date: 10 Apr 92 02:04:40 GMT
Sender: [email protected] (USENET News System)
Organization: University of Massachusetts at Amherst
Lines: 9
Nntp-Posting-Host: deimos.ucc.umass.edu
X-News-Reader: VMS NEWS 1.20

I have recently been trying to upgrade from linux 0.12 to linux 0.95a
(yes, i have taken my time, being a student does that...). Anyhow,
when i boot 0.95, i get hd timeout errors (this is from an unmodified
distribution floppy, that is supposed to go to the floppy device.
This all worked just fine with 0.12, and i am completely baffled as
to why it doesn't work. My system is a 486/33, with a seagate st-225
MFM harddrive.

-- craig hagan


[next article]
From: [email protected] (Ross Biro)
Newsgroups: comp.os.linux,alt.os.linux

Subject: WD8003 ethernet driver for linux
Message-ID: <[email protected]>
Date: 10 Apr 92 02:47:01 GMT
Sender: [email protected] (News Service)
Organization: DSG, Stanford University, CA 94305, USA
Lines: 15


I have just ported the 386bsd wd8003 ethernet driver to linux;
I have also patched ka9q to work with it. Right now it is buggy, but
it appears to be useable (I am using it to type this.). I should have
diffs and such ready to go in a few days. More if I upgrade to 95c
in the meantime. I would like to hear from anyone with this
hardware who is willing to help test/debug.

Has anyone actually been working on sockets or tcp/ip in the
kernel, or just arguing about it? I wondering if I should start
trying to migrate some of the ethernet code into the kernel, or just
leave it like it is until someone comes up with a good generic
system.

Ross Biro [email protected]


[next article]
From: [email protected] (Andy I. Liaw)

Subject: Re: Patch to preset the SVGA mode
Message-ID: <[email protected]>
Date: 10 Apr 92 03:10:19 GMT
References: <[email protected]>
Organization: Texas A&M University, College Station
Lines: 17

In article <[email protected]> [email protected] (Dennis Flaherty)
writes:
>
>If you are tired of hitting RETURN and selecting the
>same SVGA mode every time you boot Linux, apply this
>patch and rebuild the kernel. It allows you to preset
>the mode in the Makefile, so that the setup program
>proceeds as if you had hit RETURN and your favorite number.
>
>Dennis
>
Could you please also post the same thing for presetting
the mode to VGA (or EGA)? Maybe it's trivial, but I'm
not a C programmer and have no idea how this is done.

Thanks!

andy [email protected]


[next article]
From: [email protected] (A. V. Le Blanc)

Subject: Re: documentation for shoelace
Message-ID: <[email protected]>
Date: 9 Apr 92 10:05:11 GMT
References: <[email protected]>
Followup-To: [email protected]
Organization: Computing Centre, University of Manchester
Lines: 67


In article <[email protected]> [email protected]
(The WOZ) writes:
>1) does shoelace _need_ to be in the root directory? Can I remove
> it later?

Yes, shoelace does not work unless it is /shoelace. You can never remove
it. It is something like a /vmunix file, in that its name and location
are hard-coded into the bootstrap loader. When shoelace starts to
execute it searches for /etc/config, another hard-coded file name.
A sample /etc/config comes with it.

>2) The readme and config files refer to files in /etc/system (mm,
> kernal, fs, and init) The 0.95a dist does not have a /etc/system Do I
> need to get these files from somewhere?

These are a hangover from minix. They should normally be commented
out in your /etc/config file. Instead of /etc/system you should
use the Image file produced by a make in the linux directory. I
keep this in /etc/Image, with the previous copy in /etc/Image.old.

>4) What is the file /usr/src/linux/linux/Image ? (Please tell me I'm
> not going to have to recompile the kernal or something...)

You must indeed recompile the kernel, that is, unless you get a
compiled kernel from somewhere. The interim version of Linux
which I am about to release in Manchester contains a precompiled
kernel (0.95c).

>3) What is the file bootlace used for? It's not mentioned in the
> readme file.

bootlace contains the code which is (sometimes) written by laceup
into a file system's boot block.

>5) When is the file disktab used?

disktab is used when configuring a floppy to boot using shoelace.

>6) How exactly does winiboot work? What does it do?

winiboot is the part of shoelace that goes into the hard disk's boot
sector, as opposed to the file system's.

When you install shoelace, you give the command

laceup -w 1 /dev/hda

This puts the winiboot file into the hard disk's boot sector and
sets a default of partition 1 as the one which will boot if you
don't type anything within a certain period of time. Then you
give the command

laceup /dev/hda1 wini

This installs the bootlace code into the boot block of the linux
partition. This code is executed by the code in sector 0, and
this code finds /shoelace and executes it. /shoelace finds
/etc/config and uses its instructions to find /Image. Once
shoelace is executed, you can interrupt by typing RETURN, and
it allows you to boot from some file other than the configured
one.

Shoelace is pretty out-of-date and designed for a completely
different system. We need someone to write a replacement.

-- Owen
[email protected]


[next article]
From: [email protected] (A. V. Le Blanc)

Subject: Re: ksh in linux?
Message-ID: <[email protected]>
Date: 9 Apr 92 17:19:49 GMT
References: <[email protected]> <[email protected]
bm.com> <[email protected]>
Followup-To: [email protected]
Organization: Computing Centre, University of Manchester
Lines: 15

In article <[email protected]> [email protected] (Bailey Brow
n) writes:
>Does linux have ksh (Korn Shell)? And if so, does it's ksh implementation
>have the command line editing (vi mode)?

I don't know about a ksh yet, since the public domain one is said
to be less than ideal. But you CAN use vi to edit command lines in
bash. Simply create a file .inputrc in your home directory containing

set editing-mode vi

and it should work like vi: ESC enters edit mode, and most of the
usual vi commands work in the usual manner, including / and ?.

-- Owen
[email protected]


[next article]
From: [email protected] (Rick Kelly)
Newsgroups: comp.os.linux,alt.os.linux

Subject: Re: Compilers
Message-ID: <[email protected]>
Date: 10 Apr 92 01:39:46 GMT
References: <[email protected]>
Reply-To: [email protected] (Rick Kelly)
Organization: The Man With Ten Cats
Lines: 13

In article [email protected] (Ajay Shah) wri
tes:
>>The other thing that people out there may be forgetting is that there are
>>some of us that only have 2M ram, and cant afford to spring for more ram, and
>>the move to 2.x will leave us in the cold.
>
>Come on... it's $130 for a 4 Meg SIMM.

But remember that some motherboards require that memory is updated four
chips at a time. Now you're up to $520.

--

Rick Kelly [email protected] unixland!rmkhome!rmk [email protected]


[next article]
From: [email protected] (Scott R. Grabowski)

Subject: More hard drive questions....
Message-ID:
Date: 10 Apr 92 04:17:54 GMT
Organization: Sophomore, Physics, Carnegie Mellon, Pittsburgh, PA
Lines: 51

Hello everyone,

I have been experimenting with Linux for about a month now from floppy,
and am considering installing it on my hard drive sometime over summer,
once I'm back home. So far, I have been VERY impressed with the speed
of Linux. Very nicely written. Also, as an aside, my 486-33 runs fine,
with the exception below, at both 8MHz and 33MHz. I know some people
were having problems at high speeds.... It's a TPE motherboard. Real
nice.

Anyways, back to my problem. If I use fdisk to attempt to read the
partition data off of my hard drive, it doesn't report anything, just
returns to the prompt. Using pfdisk doesn't work either. My questions
are:

1. Does Linux support 2 drive controllers in 1 machine? I have 2
controllers in mine (1 MFM, 1 RLL), and it runs fine under DOS (albeit
with a device driver to read the second controller)

2. Does Linux read the hard drive data from CMOS? I have to leave my
CMOS hard drive entries blank in order for the system to boot. (Maybe
I'll try an AT controller board one of these days...)

3. Is there anything else that I need to know to make a HD work?

Thanks.

And a HUGE thanks to Linus, and all the people porting software over to
Linux. Can't wait to see X, with or without network support! Any ideas
how long before at least a late alpha or beta release is available?
(Size?? 😎

FYI - running bootimage .95c+ (from tsx-11)
rootimage .95a ( "" )

4 MB memory
/dev/at0
/dev/PS1

1st hard drive: Miniscribe 10 MB with Omni controller (primary bootable)
2nd drive: Seagate 60MB (2 30MB partitions) with a WD RLL controller.

Oh, one other thing before I forget, is it possible to change the screen
size once the kernel has been loaded? It would be nice to be able to
switch between 80x25, and 132x44 with a setterm option. Just curious.

-Scott

Scott Grabowski + "I don't pretend to understand women's
[email protected] + little quirks, just one thing I know for
Physics is Phun! + sure, chicks dig jerks." - great song!


[next article]
From: [email protected] (Nicolai Wadstr|m)

Subject: Re: Has anyone thought about threads?
Message-ID: <[email protected]>
Date: 9 Apr 92 18:55:05 GMT
References: <[email protected]> <[email protected]>
Sender: [email protected] (Lord of the News)
Organization: Lysator Computer Club, Linkoping University, Sweden
Lines: 51

[email protected] (Drew Eckhardt) writes:

>In article <[email protected]> [email protected] (Nicolai Wadstr|m) writ
es:
>>Has anyone given any thoughts on implementing threads in linux?
>>I know it's not a standard UNIX feature, but it's were nice to
>>write programs with threads, it's often possible to gain a lot
>>of performance in using threads.
>>(I've done a lot of OS/2 programming and just love the idea of
>>threads!)

>fork()ing is the same thing as creating a heavy weight thread.
>lightweight threads can be implemented with minimal modifications.

Ok, fork()ing isn't as straight forward to use, but accomplishes roughly
the same thing, altough that approach has more overhead than if you
would have an createthread() call (ex. createthread( stack, funcaddr ); ).
If I'm right (I don't know that much about UNIX) fork makes another copy
of the program, with a seperate data segment, I'm I right?
With threads you still have access all the program data.

>Interprocess communication can be accomplished with pipes,
>or signals.

>Hopefully, sysV shared memory and semaphores will be
>implemented too.

Yeah, that would also be nice! But how about queues? If you
are transfering bigger chunks between programs queues are much
better than pipes!
(I know shared memory could be used for that, but queues are actually
only a way to handle shared memory).

>Yes. Note that Linux uses shared, statically linked libraries.
>With some header field munging, and running ld(1) from user space, it would
>be possible to do dynamically linked shared libraries.

Ok, also a very nice feature!
That's also a very good way implement API's, in OS/2 the API's implemented
in that way.



/ nicolai



--
Nicolai Wadstr|m # Internet: [email protected]
#
# BBS: +46-40-124554 (down)
# Voice: +46-40-975535


[next article]
From: [email protected] (Daniel A. Martin)

Subject: Where are the sources for xcomm/xc ?
Message-ID: <[email protected]>
Date: 10 Apr 92 06:36:42 GMT
Organization: Mississippi State University
Lines: 12

I would appreciate very much if whoever put xcomm.tar.Z at tsx would
also put the sources for it up...or if anyone else knows where I can
get these sources please let me know by mail...I was able to find
the sources for xcomm, but these were from 1989 and the README file
in xcomm.tar.Z at tsx shows that it came from a very recent release...
I would appreciate any help in this matter...

BTW...on a related not...when I tried to execute the new kermit (5A)
I got a segmentation fault...I am using the new kernel that Linus just
released....
--
Danny Martin : [email protected]


[next article]
From: [email protected]

Subject: Re: modem login
Message-ID: <[email protected]>
Date: 9 Apr 92 18:32:34 GMT
References: <[email protected]>
Organization: Oxford University VAXcluster
Lines: 60

In article <[email protected]>, [email protected] writes:

> ...

> Anyway, How _does_ one go about login via modem?

> I would ideally like to login to linux via modem(2400 on ttys2) and
> initiate a kermit session to my xt(2400 on ttys1 null modem), however,
> I can't seem to login via my modem. I have used getty 19200 ttys1 &
> to login directly from the xt, but getty 2400 ttys2 & always gives
> problems(i think the RING message is interfering with normal login)

> ...

> Jim Gifford
> [email protected]

The RING messages will definately mess up your login. We had a similar problem
with our dialin modems (to a Gandalf network) at Oxford. We solved this by
connecting a terminal to the modem & typeing the following :-

AT&F
AT&C1
AT&D3
AT&K1
ATQ1
AT&W
ATZ

now, for the explinations ...

AT&F = Factory default the modem (ie reset 🙂
AT&C1 = V24 signals follow true state (ie DCD comes on when there is a
carrier)
AT&D3 = Reset modem when DTR is droped. (Reset the modem when we are
disconnected)
AT&K1 = XON/XOFF flow control
ATQ1 = Turn off modem messages (you don't see any responces to commands after
this, but the commands are executed)
AT&W = Store the active configuration
ATZ = use stored configuration

This may vary from hayes implementation to implementation, so check your modem
manual; what can I say, it works for us !!! We push the answer button in on the
modem, you could use ATS0=1

Hope this helps,

Tony


_
Tony Kew, Oxford University Computing Services, (_)xford, England, U.K.
====================================================|=========================
Janet:- [email protected] --+-- Disclaimer :-
Internet:- tony%[email protected] | Who is the more foolish?
Earn/Bitnet:- tony%vax.oxford.ac.uk@ukacrl | The fool, or the
Phone:- +44 (0)865 273268 Fax:- +44 (0)865 273275 | man who follows him?
==================================================\=|=/=======================
\|/


[next article]
From: [email protected] (Ed Carp)

Subject: Re: Swapfile vs. Swap partition
Keywords: Swap space
Message-ID: <[email protected]>
Date: 10 Apr 92 06:58:25 GMT
References: <[email protected]> <[email protected]>
Organization: Bleeding Edge Technologies - 'Sliding Down The Razor Blade Of Life
...'
Lines: 24


----- News saved at 10 Apr 92 06:56:15 GMT
>>My question is which is
>>preferable, using a partition for swap space or or a swapfile?

>The two are apparently about equivalent, I understand, as far as efficiency.
>A swap file is much easier to resize if your needs should change.

Not exactly true. When you use a swapfile for swap, you are subject to the
vagaraties (and overhead) of whatever filesystem you are using, not to
mention the possible fragmentation. A partition is almost always more
efficient in terms of swapping -- but isn't so easy to resize (especially
upwards!) if that becomes necessary.

The reason most OS vendors offer that option is because customers demanded
it -- and didn't necessarily mind (or care about) the inefficiency.

Oh, BTW -- please make sure your Folloup-To: lines contains a valid
newsgroup, instead of an email address.
--
Ed Carp N7EKG/6 [email protected] [email protected]
Cupertino, CA 408/252-5947

-- Absolutely unabashed Gates McFadden groupie! --


[next article]
From: [email protected] (Rogier Wolff)

Subject: Re: __ctype undefined symbol error
Message-ID: <[email protected]>
Date: 10 Apr 92 08:24:08 GMT
References: <[email protected]> <[email protected]
du>
Sender: [email protected] (UseNet News System)
Organization: Delft University of Technology, Dept. of Electrical Engineering
Lines: 51
Nntp-Posting-Host: neuron.et.tudelft.nl

[email protected] (Joseph Knapka) writes:

>Hi, all. I was having trouble compiling ps after applying Linus' and
>H.J.'s patches to the .95a kernel. Turned out I hadn't rooted out all
>traces of gcc-1.40 --- that part of the FAQ is to be taken very seriously!
>Here's H.J.'s message:

> __ctype is from the header files of gcc 1.40. Please backup and remove
> any files related to the previous compilers before you install gcc 2.1.
> There is no __ctype anywhere in the header files for gcc 2.1.
>
> BTW, they are in my FAQ. PLEASE READ IT WORD BY WORD AND FOLLOW THE
> INSTRUCTIONS.
>
> Could you please forward my msg to news groups?
>
> H.J.
> --
> School of EECS Internet: [email protected]
> Washington State University BITNET: [email protected]
> Pullman, WA 99164 Phone: (509) 335-6470 (O)
> USA (509) 334-6315 (H)

>Thanks for your help, H.J. -- the fault was entirely mine.

>For those who are installing 2.1 for the first time, "Remove all other
>compilers" means "rm +recursive /usr/include /usr/lib /usr/local/lib
>(and anything else that looks suspicious)." Took me a while to catch
>on to that 🙂

However the __ctype that ps manages to find is the one that comes with
linux kernel sources.

ps needs a lot of include files from the system, to be able to
read kmem and so forth. However, ctype is also in the kernel
include directory, while the intended ctype is in the /usr/include
directory.

I changed
#include
to
#include "/usr/include/ctype.h"
and all managed to compile.

It still crashes though.....

Roger
--
If the opposite of "pro" is "con", what is the opposite of "progress"?
(stolen from kadokev@iitvax ==? [email protected])
EMail: [email protected] ** Tel +31-15-783644 or +31-15-142371


[next article]
From: [email protected] (Doug Evans)

Subject: Getting started
Message-ID: <[email protected]>
Date: 9 Apr 92 23:01:58 GMT
Followup-To: poster
Organization: Edmonton, Alberta
Lines: 14

I'd like to get linux and have several questions. I have set Followup-To
poster because I don't receive this newsgroup (yet!).

1) Does it come as source, like 386bsd? If so, what are the archive sites?
How big is the source?

2) Is there an archive site for alt.os.linux and/or comp.os.linux?

1e6 thanks in advance!
--
Doug Evans |
..!{canada,uunet}!sspiff!dje | Vancouver 11, Calgary 0
[email protected] (prefered!) | There is no joy in Cowtown.
[email protected] | But how 'bout those Canucks, eh?


[next article]
From: [email protected] (Alfred Terwindt)

Subject: big archives
Message-ID: <[email protected]>
Date: 10 Apr 92 11:28:04 GMT
Sender: [email protected]
Organization: University of Twente, Dept. of Computer Science
Lines: 16
Nntp-Posting-Host: maas

Hello Linuxers,

Linux runs, great to have a *real* OS running on my PC 🙂

I'm fairly unexperienced with the un*x world, so I wonder if there is good
method to save a software package (Linux binaries/sources or root (!) ) on
floppies. If not, I plan to write something like what I call a 'mdar' witch
stands for a Multiple Disk ARchive. So, if this a good idea, and I'm not
reinventing the wheel, I like to hear some suggestions, standards etc. I
can't keep up with the group news, so please email.

Alfred.

[email protected]
--
rn: .signature: Alfred: No such user or process


[next article]
From: [email protected] (Farhan H Garib)
Newsgroups: comp.os.linux,alt.os.linux

Subject: kermit5a
Message-ID: <[email protected]>
Date: 10 Apr 92 10:56:32 GMT
Sender: [email protected] (The Network News)
Organization: Columbia University
Lines: 7
Nntp-Posting-Host: cunixb.cc.columbia.edu

I downloaded kermit5a from tsx-11.mit.edu today. when I tried
to run it I got a segmentation fault.

Did the file get corrupted in transfer or is there a problem
with the file?

-Greg


[next article]
From: [email protected] (Joseph Knapka)

Subject: Writing a device driver
Message-ID: <[email protected]>
Date: 9 Apr 92 23:34:11 GMT
References: <[email protected]>
Organization: University of Georgia, Athens
Lines: 11

Hello, all...

I'd like to write a device driver for my Trident VGA card, so I could
implement something like Sun's pixrect library under Linux. Could
someone versed in such arcane endeavors perhaps email me a reference
or two to some good literature on device-driver writing? How do people
learn this stuff, anyhow? 🙂

Thanks in advance for any pointers,

Joseph


[next article]
From: [email protected] (david.a.cuka)
Newsgroups: comp.os.linux,alt.os.linux

Subject: Re: Large disk partition
Message-ID: <[email protected]>
Date: 10 Apr 92 13:51:24 GMT
References: <[email protected]> <[email protected]
lsinki.FI> <[email protected]>
Followup-To: comp.os.linux
Organization: AT&T
Lines: 35

In article <[email protected]> [email protected] (Stephen Harris)
writes:
>In article <[email protected]> [email protected].
FI (Linus Benedict Torvalds) writes:
>>There are other fdisks than the linux or dos ones - pfdisk etc is
>
>Obviously, I have pfdisk (from the 95a root disk), but I didn't think this
>could create extended partitions. I'll recheck though. Thanks.

I just repartitioned my HD last night. I used pfdisk to set up the
initial partitions for both DOS and linux. I used the following
configuration with pfdisk:
1 6 0 76 # 40Mb
2 5 77 153 # 40Mb
3 129 154 230 # 40Mb
4 129 231 307 # 40Mb

Something was not quite right because DOS formatted the first partition
to 75Mb iso 40Mb. To solve this, I deleted the 2 DOS partitions with
DOS 'fdisk' and recreated them. This seemed to work and the C: drive
formatted to the correct size. I then created a logical drive in the
extended partition and it also formatted correctly. LINUX happily made
file systems on both /dev/hda3 and /dev/hda4.

In the end, pfdisk works great for linux partitions and DOS fdisk worked
great for the DOS partitions. One interesting quirk with DOS; I had
to reinstall DOS from my original DOS 5.0 upgrade; even after format/v/s
the C: driver would not boot. Wierd.

>--
> Stephen Harris
> [email protected] ...!uknet!axion!spuddy!sweh
>
> * Meeeeow ! Call Spud the Cat on > +44 203 638780 < for free Usenet access *


David Cuka


[next article]
From: [email protected] (System test)

Subject: kernel building problems
Message-ID: <[email protected]>
Date: 10 Apr 92 15:38:57 GMT
Sender: [email protected]
Organization: Stanford University, California
Lines: 16

I've been going crazy trying to get either 0.95c or 0.95c+ to
compile with gcc 2.1. Which of these is the "second patch" to
which H.J.'s patch is to be applied?

In any event, the makefiles don't seem to work correctly for me and
I don't know how to fix them. The include files aren't found
when the various directories are stepped into.

I screwed around until I got most everything to compile but can't
get build to link. ld complains about an invalid offset for _write.
Does anyone know what might cause this.

Thanks

Don Taber
[email protected]


[next article]
From: [email protected] (P Thomson)

Subject: Compress/Uncompress (Were ?)
Message-ID: <[email protected]>
Date: 10 Apr 92 12:26:37 GMT
Organization: Edinburgh University
Lines: 5


Hi Quick Q. Can any body tell me were I can find Compress /Uncompress for
linux.

Thanks in Advance Peter.


[next article]
From: [email protected] (David Engel)

Subject: Re: kernel building problems
Message-ID: <[email protected]>
Date: 10 Apr 92 16:29:06 GMT
References: <[email protected]>
Organization: Optical Data Systems, Inc.
Lines: 33

[email protected] (System test) writes:
: I've been going crazy trying to get either 0.95c or 0.95c+ to
: compile with gcc 2.1. Which of these is the "second patch" to
: which H.J.'s patch is to be applied?

It refers to 0.95c. Unfortunately, Linus changed the Makefile around
some in 0.95c+ causing H.J.'s patch and fix for combining the kernel
and gcc 2.1 headers to fail.

: In any event, the makefiles don't seem to work correctly for me and
: I don't know how to fix them. The include files aren't found
: when the various directories are stepped into.
:
: I screwed around until I got most everything to compile but can't
: get build to link. ld complains about an invalid offset for _write.
: Does anyone know what might cause this.

I got 0.95c+ to compile with gcc 2.1 by mucking with the Makefiles and
adding some symlinks in /usr/include to /usr/src/linux/include. The
main thing to make sure of when compiling the kernel is that tools/build
should search /usr/include first and /usr/src/linux/include second, and
be linked with the gcc 2.1 libraries. Everything else in the kernel
distribution (I'm not including ps) should be self-contained.

Aside to Linus: Would you mind describing how you have your system set
up? It might help aleviate problems like this if we knew which compiler
and system libraries you are using.

David
--
David Engel Optical Data Systems, Inc.
[email protected] 1101 E. Arapaho Road
(214) 234-6400 Richardson, TX 75081


[next article]
From: [email protected] (Drew Eckhardt)

Subject: Linux SCSI support
Message-ID: <[email protected]>
Date: 10 Apr 92 16:48:33 GMT
Sender: [email protected] (The Daily Planet)
Organization: University of Colorado at Boulder
Lines: 140
Nntp-Posting-Host: hamlet.cs.colorado.edu

The new version (.95c+ plus SCSI) is available. It supports
Adaptec, Seagate, and Ultrastor host adapters.

You can obtain it in the files

headrest.woz.colorado.edu :
/pub/linux/scsi.95c+.tar.
/pub/linux/vmunix.Z (binary)

The Quick FAQ :

Q : What SCSI hosts are supported?

A : The Adaptec 1542 (The EISA bus master 17xx may work),
Seagate ST01/ST02, (These are
repackaged Future Domain 88x series boards.
If you have an 8 bit future domain controller,
you may be able to modify the seagate driver
to detect your card to. If any one does this,
please send diffs.), and Ultrastor 14F are supported.


Q : What about SCSI tapes ?

A : The stubs are sort of there. Basically, the SCSI tape
driver needs to be changed from a block to character
device, a read/write function made that makes
appropriate calls to the lowlevel
scsi_do_cmd function, and enough
ioctl's calling the same function to
implement a normal mt.

My guess is these will be about as complex as the sd.c
driver, or ~500 lines.

Q : what about extended partitions?

A : not yet.

Q : I get a message saying READ CAPACITY FAILED.
What does this mean?

A : In some cases, very old, brain dead drives (esp Seagate)
don't support read capacity. This is a warning -
the only real implication is that if you try to
read past the end of the drive, on the
raw device, it will have to timeout, instead of
simply checking bounds and failing outright.

In other cases, something just gets flakey.
I haven't had a chance to track this down,
as until I patched .95c+, I didn't
have this problem.

Q : What are the major / minor numbers for SCSI drives?

A : Right now, it is theoretically possible to have
21 SCSI devices in the system. With SCSI-
MFM / ESDI / SMD adapters this can be even
more drives, as each id may have several drives on it
each set up as a LUN of that id. Needless to say,
if we have allowed for 16 partitions per SCSI drive,
we're pushing the limits on minor numbers. Therefore,
a yucky dynamic numbering scheme is used.

Block device major 8 is used for SCSI drives. I
personally like /dev/sd[n][p], where n is the drive
number, and p partition letter.

Minors are assigned in increments of 16 to SCSI disks
as they are found, scaning from host 0, ID 0 to host
n, ID `7, excluding the host ID.

A minor where minor mod 16 = 0 is the whole drive,
where minor mod 16 is between 1 and 4, that partition,
and extended partitions are not yet implemented.

Example : I have four SCSI disks, set up as follows
Seagate ST02, ID=0
Seagate ST02, ID = 5
Ultrastor 14, ID = 0
Adaptec 1542, ID = 0

The first disk on the seagate at ID 0 will become
minors 0-15 inclusive, the second at ID5 16-31
inclusive, the disk on the Ultrastor 32-47, on the
Adaptec 48-63.


Q : The distribution kernel panics, saying that it can't mount
root.

A : This is because my system mounts root from /dev/hd1 (old
numbering) or /dev/hda1 (new numbering), dev b 03 01.
If you have no minix filesystem there, it will die.

If you are booting from floppy, change the word at
offset 508 in the bootsector to 0000, or the
appropriate value

Q : How do I reduce kernel bloat and eliminate the drivers I
don't want?

A : Simply #undef CONFIG_DISTRIBUTION in
include/linux/config.h, and define the
macros for the SCSI hosts you want
enabled.

Q : What about distributions, information, etc

A : Subscribe to the SCSI mailing list,

[email protected]

and the latest release is maintained on
headrest.woz.colorado.edu under /pub/scsi

Q : What about bugs?

A : It works on MY hardware. It works on Tommy Thorn's
Adaptec system, Dave Gentzel's Ultrastor, and
quite a few other systems.

Bugs that are there will mostly be very hardware
specific, and nasty to track down. SCSI should
be basically error free - consequently, the
error code has not been heavily tested, and
I'm sure there are bugs left in there.

If you have found a bug, please mail it
to the mailing list with specifics of
your hardware. Other people may have
the same problem, a solution, etc.

Chances are I will provide you with a debugging
version of the kernel, which will dump out important
SCSI information and let me see what exactly is
going on.

Happy hacking...



[next article]
From: [email protected] (Karl R. Buck)

Subject: Re: kernel building problems
Message-ID:
Date: 10 Apr 92 09:48:02 GMT
References: <[email protected]>
Organization: Kansas State University
Lines: 45
NNTP-Posting-Host: matt.ksu.ksu.edu

[email protected] (System test) writes:

>I've been going crazy trying to get either 0.95c or 0.95c+ to
>compile with gcc 2.1. Which of these is the "second patch" to
>which H.J.'s patch is to be applied?

Yes, I'm having problems too. I recently posted asking specifically
what the "second patch" was and if it was the 0.95c patches that was
the latest (at that time). I received one helpful email message from
some helpful person saying that the "second" patch was included in
this diff. I assumed (perhaps wrongly) that the latest 0.95c+ sources
also include this "second patch."

>In any event, the makefiles don't seem to work correctly for me and
>I don't know how to fix them. The include files aren't found
>when the various directories are stepped into.

What exactly are the errors?

Now, I'm (trying) to compile a virgin 0.95c+ kernel with GCC 2.1
without much luck. I had to fix a couple of minor things in the
Makefile and completely (I hope) got rid of the last bits of gcc 1.40.

>I screwed around until I got most everything to compile but can't
>get build to link. ld complains about an invalid offset for _write.
>Does anyone know what might cause this.

Well, I'm having a similar problem. I get several warnings but finally
get the error:

ld -M boot/head.o init/main.o kernel/kernel.o mm/mm.o fs/fs.o
fs/minix/minix.o kernel/blk_drv/blk_drv.a kernel/chr_drv/chr_drv.a
kernel/math/math.a lib/lib.a -o /tmp/system > /tmp/System.map
ld: malformed input file (not rel or archive) kernel/blk_drv/blk_drv.a(`
.SYMDEF)
make: *** [/tmp/system] Error 1

I thought It would be that I was using the wrong ld or a.out.h, but I
have tried the recommended combination with no luck. I also got this
message when trying to compile the 0.95c sources.

I'd be interested in hearing from anyone who has successfully compiled
the 0.95c+ kernel with 2.1. Thanks!
--
Karl Buck, [email protected] 🙂


[next article]
From: [email protected] (Linus Benedict Torvalds)

Subject: Re: Another bug?
Message-ID: <[email protected]>
Date: 10 Apr 92 20:39:16 GMT
References: <[email protected]> <1992Apr9.200244.28901@wam
.umd.edu>
Organization: University of Helsinki
Lines: 24

In article <[email protected]> [email protected] (Joel M. Hoffman
) writes:
>In article <[email protected]> [email protected]
I (Linus Benedict Torvalds) writes:
>>Well, only one known bug so far, but a couple of problems. I thought I'd
>>mention them before anyone else does, and we'll call them "features" :^)
>
>More than once, I've gotten a message about mis-linked processes. I
>usually get it when I'm compiling from one VC and using Kermit
>interactively on another, but most of the time I'm compiling and using
>Kermit....

The mis-linked process problem seems to be linked with swapping: I've
never seen it, and most (all?) that have reported it have only 2M ram -
it seems to be a race-condition in the exit code, that breaks while
swapping heavily. I think that running gcc and kermit on a 2M system
should be enough to swap quite a bit...

I have fixed some of the exit-code in 0.95c+ - I'd be very interested to
hear if this problem is gone with the new kernel. The exit-code was
"interesting" - I think tytso optimized the sibling linking a bit too
much, and it broke when paging meant that the kernel sleeps in some new
circumstances. I might be totally off base - apologies to tytso if I
am.

Linus


[next article]
From: [email protected] (Dale Ulan)

Subject: Compiling 0.95c+
Message-ID: <[email protected]>
Date: 10 Apr 92 21:04:18 GMT
Sender: [email protected]
Reply-To: [email protected] (Dale Ulan)
Organization: University of Alberta Electrical Engineering
Lines: 49
Nntp-Posting-Host: bode.ee.ualberta.ca


To Linus: Great job/ great idea!


Now... the problem.

I am having a (very) hard time getting 0.95c+ to compile.

It compiles with about five or six 'incompatible pointer' warnings,
and gets to disk no problem.

After booting, this is what appears:

8 virtual consoles
4 pty's
lp_init: lp1 exists (0)
stack segment:0000
EIP: 0008:0000D752
EFLAGS: 00010202
fs: 0010
base 00000000 limit 000A0000
pid: 0, process nr: 0
8b 45 d0 66 c7 04 cd e0 a4 00

Kernel panic: Trying to free up swapper memory space
In swapper task- not syncing

Here are the details of compiling:

software emulation flag is still set 'as was' in the .95c+ Makefile.

rootdev /dev/hda2

-fcombine-regs -fstrength-reduce uncommented so it will compile under 1.40.

2 megs RAM. 1.2 meg /dev/floppy0
(doesn't seem to complain that there isn't enough memory: it just makes
the disk drive dance a lot).

I've tried figuring out what's wrong, and I know it at least gets to
move_to_user_mode(); in main.c, but I got lost tracing it after that
(using printk("Got here...\n\r");)

The .95c+ binary works, but, of course, ps needs the system alone and not the
image.. to build /etc/psdatabase

-------
Dale Ulan [email protected]
Computer engineering undergrad


[next article]
From: [email protected] (Rob Coleman)

Subject: Re: Re: kermit5a
Message-ID: <[email protected]>
Date: 10 Apr 92 21:31:24 GMT
Organization: Georgia Institute of Technology
Lines: 20

In <[email protected]> [email protected] (
Farhan H Garib) writes:

>I downloaded kermit5a from tsx-11.mit.edu today. when I tried
>to run it I got a segmentation fault.

>Did the file get corrupted in transfer or is there a problem
>with the file?

>-Greg

I had the same problem with the kermit that was uploaded to tsx, so I got
the sources from somewhere else, and it compiled fine (using the sysvr3
option). I used archie to find the ftp site, so I don't remember where I
got it from; just search for "kermit5a". If there are enough requests,
I'll put up a good binary..
--
Rob Coleman ][ Science, like nature, must also be tamed.
[email protected] ][ With a view towards its presevation..
Georgia Tech, P.O. Box 30449 ][
Atlanta, GA 30332 ][ -Rush/Natural Science


[next article]
From: [email protected] (Drew Eckhardt)

Subject: Re: Has anyone thought about threads?
Message-ID: <[email protected]>
Date: 10 Apr 92 21:31:23 GMT
References: <[email protected]> <[email protected]> <1211@lysa
tor.liu.se>
Sender: [email protected] (The Daily Planet)
Organization: University of Colorado at Boulder
Lines: 51
Nntp-Posting-Host: romeo.cs.colorado.edu

In article <[email protected]> [email protected] (Nicolai Wadstr|m) write
s:
>[email protected] (Drew Eckhardt) writes:
>
>>In article <[email protected]> [email protected] (Nicolai Wadstr|m) wri
tes:
>>>Has anyone given any thoughts on implementing threads in linux?
>>>I know it's not a standard UNIX feature, but it's were nice to
>>>write programs with threads, it's often possible to gain a lot
>>>of performance in using threads.
>>>(I've done a lot of OS/2 programming and just love the idea of
>>>threads!)
>
>>fork()ing is the same thing as creating a heavy weight thread.
>>lightweight threads can be implemented with minimal modifications.
>
>Ok, fork()ing isn't as straight forward to use, but accomplishes roughly
>the same thing, altough that approach has more overhead than if you
>would have an createthread() call (ex. createthread( stack, funcaddr ); ).
>If I'm right (I don't know that much about UNIX) fork makes another copy
>of the program, with a seperate data segment, I'm I right?

Yes, depending on your definition of "copy". In a modern Unix such as
Linux, fork() uses COW (copy on write) pages. The child gets a
new task_struct, and has the page tables copied so that they reference
the same pages that are in the parent's address space. However, the
pages are not physically copied until the child actually tries to write
to one, and gets its own copy made.

The overhead is pretty close to a createthread() or whatever...

>With threads you still have access all the program data.
>
>>Interprocess communication can be accomplished with pipes,
>>or signals.
>
>>Hopefully, sysV shared memory and semaphores will be
>>implemented too.
>
>Yeah, that would also be nice! But how about queues? If you
>are transfering bigger chunks between programs queues are much
>better than pipes!
>(I know shared memory could be used for that, but queues are actually
>only a way to handle shared memory).
>

Agreed. You generally use a pipe for redirection I/O to chain
programs together, like within the Unix shell model. Pipes are not
ideal for interprocess communication.

Other than Unix domain sockets and pipes, I'm not sure that Unix
has a "queue", although this could be implemented with shared
memory and semaphores.


[next article]
From: [email protected] (M. Saggaf)

Subject: Re: Linux SCSI support
Keywords: scsi, future domain, linux
Message-ID: <[email protected]>
Date: 10 Apr 92 22:33:01 GMT
References: <[email protected]>
Sender: [email protected] (News system)
Organization: Massachvsetts Institvte of Technology
Lines: 8
Nntp-Posting-Host: erl.mit.edu

Is there any way Linux can work with a SCSI Future Domain 1670 host
adapter? I'm a new user of Linux. Version .95a could not recognize my
hard drive. I also have an RLL drive connected to an Adaptec 8-bit
(XT) controller (in my 386 computer). Version .95a failed to recognize
that one as well. Is there any way I can get Linux to work with any of
these? Thanks.

-- M. Saggaf ([email protected])


[next article]
From: [email protected] (William A Jones)

Subject: Modems and com port access
Message-ID: <[email protected]>
Date: 10 Apr 92 23:35:41 GMT
Sender: [email protected]
Organization: The Ohio State University
Lines: 23
Nntp-Posting-Host: bottom.magnus.acs.ohio-state.edu

Now that I have a copy of UUCP for linux, I decided that it was time to
see if my modem worked properly. So, I scanned all of the digest entries
that had anything to do with the modem and decided that, under version
0.95c+, COM2 was assigned to /dev/ttys2.

Following the FAQ, I put the following line in /etc/inittab

ttys2:console:/bin/getty -m 1200 ttys2

and rebooted. So far, so good.

Now, kermit, uucico, pcomm, and everything else will not recognize that
a modem exists. What have I done wrong?

If someone who has a working kermit, uucp, etc., would send detailed(!!)
instructions on how to get set up I would be very grateful.

Bill

===================================================================
[email protected] ..!osu-mps!cambria!bjones
[email protected]
===================================================================


[next article]
From: [email protected] (Drew Eckhardt)

Subject: Re: Linux SCSI support
Keywords: scsi, future domain, linux
Message-ID: <[email protected]>
Date: 10 Apr 92 23:25:38 GMT
References: <[email protected]> <1992Apr10.223301.24673@athena.
mit.edu>
Sender: [email protected] (The Daily Planet)
Organization: University of Colorado at Boulder
Lines: 46
Nntp-Posting-Host: romeo.cs.colorado.edu

In article <[email protected]> [email protected] (M. Sagg
af) writes:
>Is there any way Linux can work with a SCSI Future Domain 1670 host
>adapter? I'm a new user of Linux. Version .95a could not recognize my
>hard drive. I also have an RLL drive connected to an Adaptec 8-bit

People are working on a lowlevel driver for the 16 bit Future domain
boards. However, I'm not sure what the timeframe is on this, mail
out on the SCSI mailing list to find out who and the details.

>(XT) controller (in my 386 computer). Version .95a failed to recognize
>that one as well. Is there any way I can get Linux to work with any of
>these? Thanks.
>

The XT disk controllers are physically different from the AT disk
controllers. This doesn't matter when you're going through BIOS
and it looks all the same, but under a real operating system,
each piece of hardware must be dealt with individually at the
register level.

Either controller could be supported with an appropriate driver -
but since you can find an AT disk controller for ~$20, pulled
from equipment for upgrades, a driver for the XT controller
wouldn't be worth your time and effort.

As far as the SCSI host - the generic SCSI package sticks most of

the common code in one place, leaving relatively little for you
to implement in your driver at the lowest level. Basically, you're
concerned with getting data to/from the SCSI host- translation from
block device commands to SCSI, timeouts, error handling, etc are all handled
higher up the driver stack. You also have to provide an abort function,
SCSI bus reset function, probe, and info function.

Like I said, people are working on this, but if you have anything
to contribute, feel free to join the SCSI mailing list at

[email protected]

Adaptec, Seagate, and Ultrstor lowlevel drivers have been
incorporated into the generic package, which is available from

headrest.woz.colorado.edu

under

/pub/linux/scsi.95c+.tar.Z


[next article]
From: [email protected] (Joel M. Hoffman)

Subject: Re: Another bug?
Message-ID: <[email protected]>
Date: 11 Apr 92 02:53:30 GMT
References: <[email protected]> <1992Apr9.200244.28901@wam
.umd.edu> <[email protected]>
Sender: [email protected] (USENET News system)
Organization: University of Maryland at College Park
Lines: 31
Nntp-Posting-Host: rac2.wam.umd.edu

In article <[email protected]> [email protected].
FI (Linus Benedict Torvalds) writes:
>In article <[email protected]> [email protected] (Joel M. Hoffma
n) writes:
>>In article <[email protected]> [email protected].
FI (Linus Benedict Torvalds) writes:
>>>Well, only one known bug so far, but a couple of problems. I thought I'd
>>>mention them before anyone else does, and we'll call them "features" :^)
>>
>>More than once, I've gotten a message about mis-linked processes. I
>>usually get it when I'm compiling from one VC and using Kermit
>>interactively on another, but most of the time I'm compiling and using
>>Kermit....
>
>The mis-linked process problem seems to be linked with swapping: I've
>never seen it, and most (all?) that have reported it have only 2M ram -
>it seems to be a race-condition in the exit code, that breaks while
>swapping heavily. I think that running gcc and kermit on a 2M system
>should be enough to swap quite a bit...
>
>I have fixed some of the exit-code in 0.95c+ - I'd be very interested to
>hear if this problem is gone with the new kernel. The exit-code was
>"interesting" - I think tytso optimized the sibling linking a bit too
>much, and it broke when paging meant that the kernel sleeps in some new
>circumstances. I might be totally off base - apologies to tytso if I
>am.

I just purchased four more meg. This weekend, I'll see if I still
get the mis-linked siblings with 6M (little swapping -- I hope). If
I have time, I'll take out the 4M, then try 95c+ (*), etc. I'll
let you know.

-Joel
([email protected])


[next article]
From: [email protected] (david nugent)

Subject: smail 3.1.25
Message-ID:
Date: 11 Apr 92 07:12:28 GMT
Reply-To: [email protected]
Organization: Unique Computing Pty Ltd, Melbourne, Australia
Lines: 21

Has anyone got a working smail 3.1.25 for Linux yet? I managed to
get one compiled with a new machine description file and editing
the EDITME file to suit, and it seems to work - mostly. It
consistently traps _after_ delivering mail to local mailboxes,
otherwise it seems quite function.

I haven't been able to track it down further yet due to gdb not
functioning with the new gcc 2.1 beta + kernel + a.out.h, but I
thought I'd post this just in case the problem is known and
someone's already been though this before. If not, then I'll wait
for an updated gdb.

Regards,

david

..............................................................................
david nugent Public Access Usenet "Only Nixon can go to China"
[email protected] +61-3-792-3507 - ancient Vulcan proverb
3:632/400@fidonet, 58:4100/1@intlnet, 199:4242/5@rainbownet, 33:300/6@admin
PO Box 260, Endeavour Hills, Victoria, Australia, 3802.


[next article]
From: [email protected] (Linus Benedict Torvalds)

Subject: Re: kernel building problems
Message-ID: <[email protected]>
Date: 11 Apr 92 09:57:24 GMT
References: <[email protected]> <1992Apr10.162906
[email protected]>
Organization: University of Helsinki
Lines: 20

In article <[email protected]> [email protected] (David Engel) writes:
>
>Aside to Linus: Would you mind describing how you have your system set
>up? It might help aleviate problems like this if we knew which compiler
>and system libraries you are using.

I'm still using 2.0 (and an older version at that) as I've been too busy
to upgrade. I doubt the HJ-patch is really needed - I read through it,
and I'd guess 0.95c+ should compile pretty cleanly without it. Anybody
tried it? HJ's patch also assumes you have a 387, and you'll have to
edit the makefile if you don't.

I'll have to upgrade my compiler (probably this weekend), and the next
real version should have no problems with 2.1. I'm in fact still using
my original 1.40 for everything but the kernel - I've left the actual
beta-testing of the new compilers and libraries to everyone else, and
I'll step in now that they seem stable. Lazyness or prudency, you
decide 🙂

Linus


[next article]
From: [email protected] (Mika Pekka Liljeberg)

Subject: Re: kernel building problems
Message-ID:
Date: 11 Apr 92 11:14:00 GMT
References: <[email protected]>
<[email protected]>
<[email protected]>
Sender: [email protected] (Uutis Ankka)
Organization: Department of Computer Science, University of Helsinki, Finland
Lines: 32
In-Reply-To: [email protected]'s message of Sat, 11 Apr 1992 09: 57:24
GMT

In article <[email protected]> [email protected]
I (Linus Benedict Torvalds) wrote:
> In article <[email protected]> [email protected] (David Engel) writes
:
>>
>>Aside to Linus: Would you mind describing how you have your system set
>>up? It might help aleviate problems like this if we knew which compiler
>>and system libraries you are using.
>
> I'm still using 2.0 (and an older version at that) as I've been too busy
> to upgrade. I doubt the HJ-patch is really needed - I read through it,
> and I'd guess 0.95c+ should compile pretty cleanly without it. Anybody
> tried it? HJ's patch also assumes you have a 387, and you'll have to
> edit the makefile if you don't.

I compiled .95c first without and the with the HJ patch. There were no
problems either way. Then I moved the kernel includes under /usr/include/linux
with HJ's conv script. I hope this will become "official" practise, since
it makes everything ever so much simpler. How about it, Linus? 🙂

> I'll have to upgrade my compiler (probably this weekend), and the next
> real version should have no problems with 2.1. I'm in fact still using
> my original 1.40 for everything but the kernel - I've left the actual
> beta-testing of the new compilers and libraries to everyone else, and
> I'll step in now that they seem stable. Lazyness or prudency, you
> decide 🙂

In my case a bit of both. 😉

Mika
--
Mika Liljeberg Email: [email protected]
Helsinki University [email protected]
Dept. of Computer Science


[next article]
From: [email protected] (Janne Himanka)
Newsgroups: comp.os.linux,alt.os.linux

Subject: Which graphics card for a new Linux machine?
Message-ID:
Date: 11 Apr 92 13:53:18 GMT
Sender: [email protected]
Reply-To: [email protected]
Organization: Tilde's home bakery
Lines: 19

I'm about to buy a new machine, in which I will run mainly Linux. I
have an attractive offer of a 486 machine that has Microgram
motherboard with local bus (directly on the motherboard) SVGA card
with a S3 chip. That is supposed to be a very fast configuration, and
popular with (blech) MS-Windows. But I'm not sure if I should buy this
card, since it's new and I anticipate it won't be among the first cards
that the Linux (or BSD) X servers will have drivers to. So, should I
buy a tried and true ET4000 card instead?

I would appreciate a general comment on graphics cards from the
people who are working on Linux's graphics, since it might shed light
for others also.

-Janne

"What did you do with the Pharaoh + [email protected]
Great Outhouse 8?" + [email protected]
"We ate him. He was unspeakably + TKOP-HJH@FINOU
toothsome." + Oulu Univ., Dept. of Libr. & Inf. Sc.


[next article]
From: [email protected] (Peter Orbaek)

Subject: Re: Scheme, Lisp for Linux?
Keywords: scheme lisp linux-ports
Message-ID: <[email protected]>
Date: 11 Apr 92 14:50:52 GMT
References: <[email protected]>
Sender: [email protected] (Peter Orbaek)
Organization: DAIMI: Computer Science Department, Aarhus University, Denmark
Lines: 24

[email protected] (Dennis Petrocelli) writes:

> Is anyone working on porting any version of scheme or lisp to
>linux? I would be VERY appreciative of anyone's attempts to do so. I
>have successfully compiled a C implementation of Scheme, siod-2.7
>(Scheme in One Defun, from world.std.com), with gcc-1.4 on linux .95a.
>I am hoping that someone might have tried to port a more substantial
>version, such as scheme 7.1, from altdorf.ai.mit.edu.

I use SCM at home on my linux box. It is pretty painless to get it to
compile.

I would rather use that instead of a big hog as scheme-7.1. SCM also
adheres to the R4RS and the IEEE standard.

You can ftp SCM sources from "the Scheme repository" at nexus.yorku.ca.


- Peter.
--
Peter Orbaek ----------------- [email protected] | ///
Hasle Ringvej 122, DK-8200 Aarhus N, DENMARK | ///
| \\\///
"Strong typing is for people with weak memories" | \XX/


[next article]
From: [email protected] (Joseph Knapka)

Subject: Make dies under gcc2.1
Message-ID: <[email protected]>
Date: 11 Apr 92 15:22:31 GMT
Organization: University of Georgia, Athens
Lines: 12

I've got a terrible problem. I installed gcc-2.1 again today, and
suddenly "make" quit working entirely. It says "No system rules
(/usr/share/mk/sys.mk)". Does anyone know what's going on? It's true
there isn't even a directory /usr/share, but I don't know where such a
directory would come from in the first place. It wasn't in the gcc
distribution. And even my old non-shared-libs version of pmake doesn't
work (same error). Also, my previous four or five installations of
gcc-2.1 went fine, and everything worked.

Thanks for any advice,

Joseph


[next article]
From: [email protected] (Dale Ulan)

Subject: re: compiling kernel .95c+
Message-ID: <[email protected]>
Date: 11 Apr 92 15:34:42 GMT
Sender: [email protected]
Reply-To: [email protected] (Dale Ulan)
Organization: University of Alberta Electrical Engineering
Lines: 25
Nntp-Posting-Host: bode.ee.ualberta.ca


Compiling the kernel with GCC1.40 does not seem to work any more.

Yesterday, I wrote about problems with compilation of 0.95c+
under GCC1.40.

Last night I switched to GCC2.1, and the kernel compiled fine
(just very slowly... I've only got 2 meg of RAM).
The only problem I had was with the include files from /usr/include
for build.c
I ended up commenting out a whole bunch of lines from stdlib.h,
starting with mblen and ending with wcstombs.
(parse errors in that area). (at 3am it's easier than trying to
fix the include files 🙂 )

In any case, it now works, so perhaps the GCC1.40 provision could
be taken out of the makefile...

I tried reading the kernel code, but I guess I'm not much of a kernel
hacker, what I was wondering was about the 'mismatched parent link'
and the 'os' and 'ys' warnings.... and why they happen.... or am I
the only one that gets those?

----
Dale Ulan [email protected]


[next article]
From: [email protected] (Jean Liddle)

Subject: Local Bus support?
Message-ID: <[email protected]>
Date: 11 Apr 92 16:15:26 GMT
Organization: Illinois State University
Lines: 17

I am posting this for a friend:

Hello netters. I was just curious as to whether or not LINUX supports
32-bit local bus cards, such as 32-bit video cards and/or disk controllers.
This is _not_ the same as EISA. I understand that programs running over
the bios have no problem, but since LINUX manipulates the hardware directly
I wasn't sure.

Please reply to me or this newsgroup and I'll forward the responses
accordingly.

Jean.
--
Jean Liddle
Computer Science, Illinois State University
e-mail: [email protected]
[email protected]


[next article]
From: [email protected] (Michael Nolan)

Subject: 0.95c+ bootimage on tsx-11, will there be a rootimage?
Message-ID:
Date: 11 Apr 92 16:37:20 GMT
Sender: [email protected]
Reply-To: [email protected]
Organization: University of Nebraska - Lincoln
Lines: 6
Nntp-Posting-Host: helios.unl.edu

I see that a bootimage for 0.95c+ has arrived on tsx-11, will there be a
corresponding rootimage file? (Linus's message said he would post a new
kernel, but didn't the patches since 0.95a also involve stuff in rootimage?)
---
Michael Nolan
[email protected]


[next article]
From: [email protected] (Hongjiu Lu -- Graduate Student)

Subject: a bug in strtod.c
Message-ID: <[email protected]>
Date: 11 Apr 92 16:34:03 GMT
Sender: [email protected] (USENET News System)
Organization: Washington State University
Lines: 8
Originator: [email protected]

Hi,

There is a bug in strtod.c, which is taken from glibc.a 1.02. The
problem is the precision of strtod is very low. We need find a good one.
Any idea?


H.J.


[next article]
From: [email protected] (Will Clark)

Subject: Where is expr?
Message-ID: <[email protected]>
Date: 11 Apr 92 18:19:48 GMT
References:
Sender: [email protected] (News Service)
Organization: Stanford University
Lines: 14


I saw someone (I forget who) say that expr is part of the gnu shell
utilities v1.6. This is great, but I don't know where to find
it/them. I'm getting an error message while making the 0.95c+ kernel
(although I still get a bootable image) so I would guess that expr has
already been ported to Linux.

(Thanks to everyone who told me where to find dd two weeks ago.)

Will Clark
[email protected]
------------------------
SAIC does not share my opinions.
harlick@lindy is a friend's account, but you can send me email here, too.


[next article]
From: [email protected] (tim)

Subject: Re: kernel building problems
Message-ID: <[email protected]>
Date: 11 Apr 92 18:43:07 GMT
References: <[email protected]> <1992Apr10.162906
[email protected]> <[email protected]>
Organization: University of Minnesota
Lines: 13

In article <[email protected]> [email protected]
I (Linus Benedict Torvalds) writes:
>I'm still using 2.0 (and an older version at that) as I've been too busy
>to upgrade. I doubt the HJ-patch is really needed - I read through it,
>and I'd guess 0.95c+ should compile pretty cleanly without it. Anybody
>tried it? HJ's patch also assumes you have a 387, and you'll have to
>edit the makefile if you don't.
>
i compiled 0.95c+ without any changes whatsoever using gcc 2.1. i was,
however, running a 95c kernel which had also been compiled with gcc
2.1 if that makes any difference. i haven't had any problems with
either linux or the compiler that seem to be related to that.

tim


[next article]
From: [email protected] (Joseph Knapka)

Subject: User's Guide project
Message-ID: <[email protected]>
Date: 11 Apr 92 19:06:45 GMT
Organization: University of Georgia, Athens
Lines: 124

Hi, folks.

There's an embryonic project underway to collect/write a reasonably
coherent User's Guide to Linux. I include here an excerpt of recent
discussion from the man-page mailing list, and invite comments on the
plan outlined. We're especially looking for comments / criticism /
volunteerism on the following points:

1) Do you think there is a need for such a document?

2) Does the organization outlined seem reasonable? Should it be
expanded/changed? I know it's a bit vague at the moment, but it can't
be made less so without feedback...

3) Would you be willing to contribute a chapter on a particular
subject or software package? This is important --- the level of

technical expertise among those of us on the manpage list is, in many
cases, not sufficient to turn out the sort of documentation we're
lookin' to compile. We can, however, make sure that documentation is
edited into a readable-for-beginners form without distorting the
content. So we need help from the real experts.

Post, or send email to [email protected].

Thanks,

Joseph

----------------------- Begin included text ------------------------
From: [email protected] (michaelkjohnson)

>
>From: [email protected] (Joseph Knapka [MSAI])
>
> > What I envision is a single coherent document that consists of
> > chapters devoted to each major package (TeX, groff, gcc [whatever's
> > currently most stable], emacs... and maybe some not-so-major ones) and
>
>My list: (partial, my mind is not good enough to do a complete one 😉
>1) TeX -- We should encourage good typesetting
>2) Groff -- ditto.
> --notice that there is a lot of basic stuff available already on
> how to use these programs, like "A gentle intro to TeX." But
> we could certainly use good installation instructions that
> covered what all the files are for, which are optional, etc.
>3) emacs
>4) gcc
> --at least 1-4 need detailed instructions on what different files
> mean. I think that for typesetting packages, for instance, you
> would want to spend time on font files, how to generate and
> manipulate them, etc.
> --gcc itself would need not only information on how to get shared
> libs to work (for example), but also _basic_ information on how
> they are implemented. I think that a little enlightenment in
> these areas never hurts, as long as it is not done in a painfully
> obtuse way.
>5) ka9q
>6) kermit, pcomm, xcomm/xc, other comm packages
>7) uucp (now out at tsx-11! --thanks, Ed)
>8) gnu info
>9) X
>10)mail/smail/elm/sendmail/etc.
>
> > answers the following questions:
>
> > 1) How do I install this package? Best-case scenario. Including
> > Where do I get it?
>
> > 2) What are the major stumbling blocks encountered by others during
> > installation, and how do I fix them?
>
> > 3) Why do I need this package? Give me two or three good reasons why I
> > should devote valuable disk space to it.
>
> > And whatever else people think is appropriate/necessary. An advantage
> > of this format is that the individual chapters on each package can be
> > written by anyone, and the coordinator can edit them into a coherent
> > whole. Moreover, the individual chapters could be available separately
> > with the packages. Ultimately, the Beginner's Guide installation info
> > could be included, to create a genuine user's manual for Linux.
>
> > We can also include info on ports that have been done but which aren't
> > available at the Linux-aware ftp sites (i.e. GNU software that needs
> > only trivial Makefile changes to compile).
>
>Yes. Especially info on common changes that have to be made -- Things
>like take the g off gnu products, in general, because they are the
>only ones we are using generally. What things commonly have to be
>defined, (hopefully -DPOSIX 🙂 what differences people are likely to
>find in things like devices, (like no raw versions of devices) etc.
>
> > Thoughts?
>
> > Joseph
>
>As far as this goes, I think that it is good. I would also like to
>see "How to administer your linux system" and "Linux security
>features: what do you need?" and other such topics covered. I don't
>think that they belong in the same volume as one devoted to getting
>software to work, though perhaps in a companion volume...
>
>Perhaps some documentation on using Linux for a BBS would be worthwhile.
>
>We now have e-net support through ka9q. Once we add full network
>support, I imagine Brenden Kehoe (I don't think I spelled that right)
>wouldn't mind if we included "Zen and the Art of the Internet," a
>guide to the internet that he wrote which is available in texinfo
>format.
>
>Which brings up the point:
>A lot of the stuff we are looking at may have been written already.
>Keep your eyes open, use archie to find things, whatever -- We have
>not invented a lot of new needs -- Just created a way to express them.
>
>BTW: I am going to be writing a document on the man page project for
>the docs/man directory at tsx-11. If anyone has any ideas that they
>think should be in that document, please send them to me so that I can
>include them.
>
>michaelkjohnson
>[email protected]
>[email protected]

---------------------- End of included text ------------------------


[next article]
From: [email protected] (Eric_Munson)

Subject: ABC release questions
Message-ID: <[email protected]>
Date: 11 Apr 92 20:30:36 GMT
Organization: Western Washington University
Lines: 18


Recently I have tried to keep up with the current Linux releases but
trying to find the most up to date software, applying patches, and other
things has had me fall behind...I have a few questions about the ABC release
1. how will it be released? rawrite images, tar.Z files?
2. what will be in it?
3. will it be multiple files, I know i would like to put it all on
1.2 meg floppies for ease of re-installing, that is my major gripe
with the recent release of the 2.6 meg binaries file.

I'm really having fun playing with linux, keep up the good work..
Thanks..

--
----------------------------------------------------------------------------
Eric Michael Munson [email protected]
1148 Fairhaven 206-738-7181
Bellingham, WA 98225


[next article]
From: [email protected] (Ajay Shah)

Subject: Re: User's Guide project
Message-ID:
Date: 11 Apr 92 21:09:32 GMT
References: <[email protected]>
Sender: [email protected] (Ajay Shah)
Organization: University of Southern California, Los Angeles, CA
Lines: 20
NNTP-Posting-Host: alhena.usc.edu

[email protected] (Joseph Knapka) writes:

>There's an embryonic project underway to collect/write a reasonably
>coherent User's Guide to Linux. I include here an excerpt of recent
>discussion from the man-page mailing list, and invite comments on the
>plan outlined. We're especially looking for comments / criticism /
>volunteerism on the following points:
>1) Do you think there is a need for such a document?

Yes, I think such a document would be rather valuable.

Linux is great.. what more could a hacker ask for?
But we could make the results of the same work accessible to
a much larger class of users if we add "a little" effort
on documentation and easy-to-use installation/administration.

-ans.
--

Ajay Shah, (213)749-8133, [email protected]


[next article]
From: [email protected] (Werner Almesberger)

Subject: Re: compiling kernel .95c+
Message-ID: <[email protected]>
Date: 11 Apr 92 21:10:33 GMT
References: <[email protected]>
Sender: [email protected] (USENET News System)
Organization: Swiss Federal Institute of Technology (ETH), Zurich, CH
Lines: 21

In article <[email protected]> [email protected] (Da
le Ulan) writes:
> Compiling the kernel with GCC1.40 does not seem to work any more.

I've compiled it yesterday and haven't seen any unusual behaviour
yet. (GCC 1.40, 0.95c+ source, 0.95a and 0.95c+ running, 4 MB RAM,
no swap, 386/25.)

> In any case, it now works, so perhaps the GCC1.40 provision could
> be taken out of the makefile...

Please don't ! BTW, there are two new gotchas:
-O6 isn't recognized by GCC 1.40. I've changed it to -O
the Version [sp?] dependency forces unnecessary recompilation of
large parts of the source whenever make is run.

- Werner
--
_________________________________________________________________________
/ Werner Almesberger, ETH Zuerich, CH [email protected] /
/ IFW A44 Tel. +41 1 254 7213 [email protected] /
/_BITNET:_ALMESBER@CZHETH5A__HEPNET/CHADNET:_[20579::]57414::ALMESBERGER_/


[next article]
From: [email protected] (Steven Kraft)

Subject: Re: Local Bus support?
Message-ID: <[email protected]>
Date: 11 Apr 92 22:47:59 GMT
References: <[email protected]>
Reply-To: [email protected]
Lines: 29
Nntp-Posting-Host: dumaine.gac.edu

In article <[email protected]> writes:
! Hello netters. I was just curious as to whether or not LINUX supports
! 32-bit local bus cards, such as 32-bit video cards and/or disk controllers.
! This is _not_ the same as EISA. I understand that programs running over
! the bios have no problem, but since LINUX manipulates the hardware directly
! I wasn't sure.
! Jean Liddle
! Computer Science, Illinois State University
! e-mail: [email protected]
! [email protected]

I have an Orchid Superboard 486/33 with an Orchid Fahrenheit 1280/D 32 bit
local bus video card. Linux 0.12 works fine on it. I hope to get 0.95c
installed as soon as I can, and I am sure that will also work.

However, the fancy S3 accelerator chip functions of the Fahrenheit 1280/D are
not yet accessible to Linux since no driver has been written yet. I hope to
write that driver myself in the future, but I do not have enough info yet about

that card. I have ordered an S3 Programming Manual, and I will have it in a
few days. I cannot say how long it might take to write the driver though...

For now my video card looks just like any plain VGA card as far as Linux is
concerned.

However, a local bus disk controller may be a different matter. Many variables

are involved, and I have no experience here.

Steven Kraft [email protected] Micro/Minicomputer Consultant, Comp Sci major
Gustavus Adolphus College


[next article]
From: [email protected] (Niels Skov Olsen)
Newsgroups: comp.os.linux,alt.os.linux

Subject: Keyboard dies when I hit Caps-Lock (+SCSI)
Message-ID: <[email protected]>
Date: 11 Apr 92 23:27:03 GMT
Sender: [email protected]
Organization: Department of Computer Science, U of Copenhagen
Lines: 20

I just got my Adaptec 1542B the other day and installed the stuff
from sc.tamu.edu with Tommy Thorn's SCSI driver in it. It works
fine except that my keybord locks up completely when I hit
Caps-Lock. That's very annoying because Tab is right above it
and I use that one extensively for filename completion. I compiled
the kernel with the Danish keyboard - it didn't help. I 'heard' that
Drew Eckhardt was almost ready with incorporating low level Adaptec
support, so I thought I would poke around at headrest to see if
something had come up. There was a brand new vmunix which I installed
and booted. It detected my adapter and disk just fine, but tried to
load the root from device 0301. I changed that (vmunix[508]=02) and
rebooted. That didn't help so I cleared the bytes at 508 and 509
so that the rootimage would load from floppy. That worked (phew).
But still no talking to /dev/hda2. And the keyboard still locks
up...

Any clues? No problems in DOS. The keyboard is a cheap Taiwanese
one called DANREX KB-6252EA.

Niels


[next article]
From: [email protected] (Stephen Harris)
Newsgroups: comp.os.linux,alt.os.linux

Subject: Problems compiling 0.95c+
Message-ID: <[email protected]>
Date: 11 Apr 92 18:49:53 GMT
Organization: Spuddy's Public Usenet Domain
Lines: 25

Had a few problems compiling the c+ version, patching from 0.95c

These were:
(1) Lack of include/checkpoint.h, needed in char_dev.c
I just "touch"ed this file, to create it, and the compilation worked.
What should this include file do????

(2) Boot image finally made recognised my two printer ports then crashed
in a similar way to other people have described.
I compiled using 1.40 (newgcc.tar.Z)
Looking at the stderr from the makefile I found a lot of silly lines
"unknown option M%I%" or something like that.
(I already uncommented the GCC_OPT line)
Using the gccbin.tar.Z version of 1.40 the error became "unknwon
option -O6
Removing the "6" and recompiling produced a working bootable kernel.

mknod lp0 c 6 1 / mknod lp1 c 6 2, then my printers both worked.

Hope this helps people with a similar problem.
--
Stephen Harris
[email protected] ...!uknet!axion!spuddy!sweh

* Meeeeow ! Call Spud the Cat on > +44 203 638780 < for free Usenet access *


[next article]
From: [email protected] (Joel M. Hoffman)
Newsgroups: comp.os.linux,alt.os.linux

Subject: Remote dumb terminal loses char's with kermit
Message-ID: <[email protected]>
Date: 12 Apr 92 02:04:57 GMT
Sender: [email protected] (USENET News system)
Organization: University of Maryland at College Park
Lines: 13
Nntp-Posting-Host: rac2.wam.umd.edu

I just tried hooking up an old '286 I have to my Linux machine ('386 at
25MHz, 6M, 0.95c [b?]). I hooked up the machines with Laplink cable, ran

getty 19200 /dev/ttys2 &

on Linux, and tried kermit on the remote machine. But kermit kept
losing characters. Even at 9600 I had problems. And this is with a
null cable via two serial ports. My only guess is that Linux isn't
watching for Xon/Xoff. Is this possible? Anyone have any other
suggestions??

-Joel
([email protected])


[next article]
From: [email protected] (Jim Winstead Jr.)

Subject: Re: 0.95c+ bootimage on tsx-11, will there be a rootimage?
Message-ID: <[email protected]>
Date: 12 Apr 92 02:13:49 GMT
References:
Sender: [email protected] (The News System)
Organization: Harvey Mudd College, WIBSTR
Lines: 21

In article [email protected] writes:
>I see that a bootimage for 0.95c+ has arrived on tsx-11, will there be a
>corresponding rootimage file? (Linus's message said he would post a new
>kernel, but didn't the patches since 0.95a also involve stuff in rootimage?)

There won't be a corresponding root image release. None of the
patches since 0.95a affected the root disk except for a new fdisk,
which is available in source form on tsx-11 as system-0.95a.tar.Z.

So, you're okay using the 0.95a root image, unless you plan on mucking
around with extended partitons. (Even then, the 0.95c+ kernel does a
fdisk-like listing upon booting, so you could get by with that.)

The next release of Linux, 0.96, will have a new rootimage, and we'll
be changing the numbering scheme a bit. More on all this later....

--
Jim Winstead Jr. (CSci '95) | "Catch a fish!"
Harvey Mudd College | -Geddy Lee,
[email protected] | San Diego Sports Arena
Disclaimer: Mine, not theirs! | January 20, 1992


[next article]
From: [email protected] (Jim Winstead Jr.)
Newsgroups: comp.os.linux,alt.os.linux

Subject: Re: Problems compiling 0.95c+
Message-ID: <[email protected]>
Date: 12 Apr 92 02:21:13 GMT
References: <[email protected]>
Sender: [email protected] (The News System)
Organization: Harvey Mudd College, WIBSTR
Lines: 41

In article <[email protected]> [email protected] (Stephen Harris
) writes:
>Had a few problems compiling the c+ version, patching from 0.95c
>
>These were:
>(1) Lack of include/checkpoint.h, needed in char_dev.c
> I just "touch"ed this file, to create it, and the compilation worked.
> What should this include file do????

checkpoint.h is not needed - you're best off removing the references
to it completely. I think they're in kernel/chr_drv/lp.c and
fs/char_dev.c.

>(2) Boot image finally made recognised my two printer ports then crashed
> in a similar way to other people have described.
> I compiled using 1.40 (newgcc.tar.Z)
> Looking at the stderr from the makefile I found a lot of silly lines
> "unknown option M%I%" or something like that.

There is a typo in the kernel/Makefile, for the "dep" target - the
line:
(cd math; $(MAKE) (MFLAGS) dep)
should be:
(cd math; $(MAKE) $(MFLAGS) dep)

> Using the gccbin.tar.Z version of 1.40 the error became "unknwon
> option -O6
> Removing the "6" and recompiling produced a working bootable kernel.

This is new to gcc 2.0 - you took the right way to fix it. The best
way to fix it would be to upgrade compilers, of course. 🙂

>mknod lp0 c 6 1 / mknod lp1 c 6 2, then my printers both worked.

The minor numbers here are wrong. lp0 is (6,0), lp1 is (6,1), and lp2
is (6,2). The root image for 0.96 will have a MAKEDEV that makes
these correctly....
--
Jim Winstead Jr. (CSci '95) | "Catch a fish!"
Harvey Mudd College | -Geddy Lee,
[email protected] | San Diego Sports Arena
Disclaimer: Mine, not theirs! | January 20, 1992


[next article]
From: [email protected] (Clifford A Adams)

Subject: Needed: a guide to setting up swap space.
Summary: FAQ instructions on swap space are unclear and need additions.
Keywords: linux, swap, setup, installation, mkswap, swapon
Message-ID: <[email protected]>
Date: 12 Apr 92 05:39:14 GMT
Organization: University of New Mexico, Albuquerque
Lines: 61


I recently installed Linux on my HD (/dev/hda1 mostly
following I Reid's instructions), and things seem to be running OK. I
have 4 meg of ram (2.7meg reported free after buffers, etc), which is
enough for most activities. However, I'd like to add a few meg of
swap space (especially since I like to use the full GNU Emacs while
simultaneously compiling with GCC. 🙂 I have about 200k of various
FAQs and guides, but none of them explicitly say much about how to set
up swap space. I don't think I've been successful so far in figuring
it out...

Here are my problems (using 0.95c+ boot floppy and a mostly
0.95a root on /dev/hda1):

* "mkswap" was on my 0.95a rootdisk. There seems to be a
newer version (later file date) at tsx-11 under the binaries tree
(the later one is bigger). What are the differences, and which should
I use? For the rest of my experiments

* Looking at "mkswap" I noticed a "-c" option--what does it
do? Early on I thought that "-c" meant "create", and that otherwise
mkswap would set up and use the swapfile. At the moment it seems like
you use swapon to set up a swap file--at the first reading it seemed
like swapon was only for swap partitions...

* I made a file "/dev/swapfile" using emacs to create a file
of 1K spaces, then using cat to make 4K/16K/64K/256K/1meg files, and finally
the 6 meg /dev/swapfile. I didn't know if it would need a little more
space for tables or whatever so I added another 256K as extra.
* Is there a better way to create the swapfile?
* Do I need to have more space than the #blocks I tell mkswap?

* I noticed that the mkswap blocks are 1K--this should
probably be noted since its different than the filesystem blocks...

* Is there any way to tell how much swap space you are using or
have left?

...What I'd really like to see is a guide something like Reid's guide
to installation for setting up swap space--one that would cover both
swapfiles and swap partitions. (Perhaps it could be added as an
appendix to the guide?). Just a list of the commands and what they do
would be nice however...

Finally, my first impression of Linux is: "zippy". Its been a
long time since I've had the kind of quick response Linux gives, even
when I worked on "unloaded" machines (where I'm the only user, the load
avg is 0.00, etc...). Even when running GNU emacs the response is
excellent (on my 40mhz 386 w/64K cache). It may seem like a
small difference between the seeming immediate Linux response and the
more typical 2/3-1 second reaction time, but it feels nice... (Its also
nice not to wait for overloaded NFS fileservers, and being able to
mess with the kernal is a plus. 🙂

--Cliff

--
Clifford A. Adams --- "Understanding is inasmuch as becoming."
[email protected] | 457 Ash St. NE/Albuquerque, NM 87106 | (505) 242-4519
Programmer in the USENET Moderation Project: tools for advanced filtering
(many-moderator groups, keyword/summary addition, prioritizing newsreaders...)


[next article]
From: [email protected] (Drew Eckhardt)
Newsgroups: comp.os.linux,alt.os.linux

Subject: Re: Keyboard dies when I hit Caps-Lock (+SCSI)
Message-ID: <[email protected]>
Date: 12 Apr 92 07:41:48 GMT
References: <[email protected]>
Sender: [email protected] (The Daily Planet)
Organization: University of Colorado at Boulder
Lines: 42
Nntp-Posting-Host: caesar.cs.colorado.edu

In article <[email protected]> [email protected] (Niels Skov Ols
en) writes:
>I just got my Adaptec 1542B the other day and installed the stuff
>from sc.tamu.edu with Tommy Thorn's SCSI driver in it. It works
>fine except that my keybord locks up completely when I hit
>Caps-Lock. That's very annoying because Tab is right above it
>and I use that one extensively for filename completion. I compiled
>the kernel with the Danish keyboard - it didn't help. I 'heard' that
>Drew Eckhardt was almost ready with incorporating low level Adaptec
>support, so I thought I would poke around at headrest to see if
>something had come up. There was a brand new vmunix which I installed
>and booted. It detected my adapter and disk just fine, but tried to
>load the root from device 0301. I changed that (vmunix[508]=02) and
>rebooted. That didn't help so I cleared the bytes at 508 and 509
>so that the rootimage would load from floppy. That worked (phew).
>But still no talking to /dev/hda2. And the keyboard still locks
>up...
>
>Any clues? No problems in DOS. The keyboard is a cheap Taiwanese
>one called DANREX KB-6252EA.
>
>Niels

Recognizing the hard disk :

Yes : my SCSI drivers coexist with the stock WD1003 driver.
SCSI disks are MAJOR 8, MINOR 16 * [disk number], where disk
numbers are assigned dynamically, with the SCSI disk at the
lowest ID on the bus becoming 0, the next higher one 1, etc.

Where minor % 16 == 0 , the whole disk is used
minor & 0xf == 1, the first partition
== 2, the second partition
== 3, the third
== 4, the fourth.

5-15 are unused, and will become extended partitions after that
code is worked in.

The keyboard :

Does the keyboard lock up, period, do numlock, CAD still work,
or not?


[next article]
From: [email protected] (Terry Dawson)

Subject: A couple of simple previously answered (but missed) questions
Summary: Sorry, missed them thr first time
Keywords: stderr IDE linux questions
Message-ID:
Date: 12 Apr 92 05:15:35 GMT
Sender: [email protected]
Organization: Sydney University Computing Service, Sydney, NSW, Australia
Lines: 26
Nntp-Posting-Host: extro.ucc.su.oz.au


Sorry to waste bandwidth reasking questions, but I missed reading them the
first time (saw them, they didn't affect me at the time, so I didn't read
them ... theres a lesson to be learnt there somewhere I'm sure :-).

Firstly,
When compiling with gcc 1.4, any code tht references stderr causes gcc
to complain that '___stderr' is undefined, or words to that effect, I
am sure this is a result of me not have gcc properly installed, I'm also
sure its going to be a fairly simple fix ..
What have I got wrong where ?

Secondly,
I am planning on purchasing a larger hard drive to dedicate to Linux,
I'm looking at a 105Mb IDE drive, and I was wondering if any problems
have been reported with using IDE drives with Linux 0.95a ?

Emailed replies would be greatly appreciated.

Terry



--
Terry Dawson, [email protected], vk2ktj%[email protected]
+61 2 925 1556 (voice), +61 2 922 5973 (fax). __\*/__ 0^Ooooo _____


[next article]
From: [email protected] (Rik Faith)

Subject: Re: compile kernel with gcc2.1
Keywords: linux, kernel , gcc2.1
Message-ID: <[email protected]>
Date: 12 Apr 92 13:36:14 GMT
References: <[email protected].
fsu.edu>
Sender: [email protected]
Organization: University of North Carolina, Chapel Hill
Lines: 13

One simple way to get tools/build.c to compile under gcc 2.1 is just to
edit the Makefile entry for build.c and add the following options to the
command line:

-I/usr/include -I/usr/src/linux/include

This makes gcc search the standard area first and then the kernel
area--only those things which are NOT in the normal area are fetched from
the kernel area. This is the desired behaviour. Using the other order
[which is the same as just using -I/usr/src/linux/include] will get the
kernel area to be searched first.

Something similar works fine for ps, also.


[next article]
From: [email protected] (P Thomson)

Subject: swap ?
Keywords: mkswap,swapon,mkfs
Message-ID: <[email protected]>
Date: 12 Apr 92 10:51:41 GMT
Organization: Edinburgh University
Lines: 18

Hi can someone please explain the buisness of swap please, I have
read all the info in FAQ, beginner-guide followed the instruction
but still can't get it to work correctly

What I did was : created a swap device using pfdisk ie /dev/hda3
mkfs /dev/hda3
mkswap /dev/hda3 5400 (block) about 5Mbytes

and this seems to work ok ie no errors or complaints

next I'm trying to edit using ue a 1Mbyte file (for test purposes)

5 seconds later the the disk drive ( a 91 Mbytes space) starts
working over time (as expected) but then I get an out of memory
message and the system seems to hang


Thanks in advance Peter.


[next article]
From: [email protected] (Niels Skov Olsen)
Newsgroups: comp.os.linux,alt.os.linux

Subject: Re: Keyboard dies when I hit Caps-Lock (+SCSI)
Message-ID: <[email protected]>
Date: 12 Apr 92 17:16:58 GMT
References: <[email protected]>
Sender: [email protected]
Organization: Department of Computer Science, U of Copenhagen
Lines: 30

I wrote late one night:

>I just got my Adaptec 1542B the other day and installed the stuff
>from sc.tamu.edu with Tommy Thorn's SCSI driver in it. It works
>fine except that my keybord locks up completely when I hit
>Caps-Lock. That's very annoying because Tab is right above it
>and I use that one extensively for filename completion. I compiled
>the kernel with the Danish keyboard - it didn't help. I 'heard' that
>Drew Eckhardt was almost ready with incorporating low level Adaptec
>support, so I thought I would poke around at headrest to see if
>something had come up. There was a brand new vmunix which I installed
>and booted. It detected my adapter and disk just fine, but tried to
>load the root from device 0301. I changed that (vmunix[508]=02) and
>rebooted. That didn't help so I cleared the bytes at 508 and 509
>so that the rootimage would load from floppy. That worked (phew).
>But still no talking to /dev/hda2. And the keyboard still locks
>up...

>Any clues? No problems in DOS. The keyboard is a cheap Taiwanese
>one called DANREX KB-6252EA.

OK, now that I am awake I noticed that SCSI devices now have major 8!
That's fine, but my keyboard still locks up. Linux is still running,
it is just not receiving my keystrokes - I have to hit the reset
button (after waiting ~30 seconds for update to sync).

Am I the only one having such a problem???

Niels
[email protected]


[next article]
From: [email protected] (Greg Naber)
Newsgroups: comp.os.linux,alt.os.linux

Subject: Fidonet Distribution
Message-ID:
Date: 12 Apr 92 20:01:44 GMT
Reply-To: [email protected] (Greg Naber)
Organization: S'Qually Holler's, Renton WA
Lines: 141

Well I really put my foot in it this time;-)

I have been assigned/cajoled/volunteered to get this stuf distributed in
the Fidonet land. I do have the backing of SOME of the fidonet filemover
backbone structure to accomplish this.

What I am looking for is comments to the following text file before
setting up the distribution scheme. (Also more volunteers with dual
identities to help out;-)

This file is in fidonet jargon.
Please help me pick this apart, any comments appriciated.


Linux Distribution Network Policy Document For Fidonet - Rev 1 - 12 April 1992
Greg Naber
1:343/[email protected] or [email protected]


Area Name Area Description
========= ================

LNXSYS LNX - System Files Distribution.
contains: Linux system files, source,
system updates, drivers,
patches & modifications,
release notes, installation
instructions and boot/root images.

LNXCOMP LNX - Complier Files Distribution.
contains: complier related materials, i.e.
GCC-* source code, binaries, libraries,
and related installation instrutions.

LNXDOC LNX - Documentation Files Distribution.
contains: Manual (man pages) documentation,
for all areas. Update documentation
to all areas. Installation Instuctions.
Misc. documentation.

LNXUTIL LNX - Utilities File Distribution.
contains: All Linux related binary format
utilities and updates.

LNXGAME LNX - Games File Distribution.
contains: All Linux related binary format games.

LNXSOURC LNX - Source Code Distribution.
contains: All source code for utilities and games.

LNXMISC LNX - Miscellaneous Linux Related files.
contains: Any Linux related files not covered
by other file areas.

LNXUNIX LNX - Unix Related Files Distribution.
contains: Any files related to unix and MAY NOT be
related to Linux.

1. I would like to not see "CrippleWare" or "VaporWare" included. Only Public
Domain and ShareWare should be submitted. Demonstration software submitted will
be allowed only if the software is able to be used in some fashion that will
enable the end user to "Try before you buy". One thing to keep in mind is that
Linux itself and GCC is 'freeware'.

2. Due to the unix abilities to have numerous characters in the naming of
files and dos's limitation of 8 character with 3 character extension names, it
would be impracticable to allow any node to hatch files into the LNX areas.
Hatching will be restricted to LNX RC's. This will control the possible hatch-
ing of duplicate files under different names.

3a. The common compression utility used will be unix format to make
transfers between Dos and Unix easier for beginning systems to transfer files
between the two. The common naming of files will reflect this in the extension
of the file name. I.E. An extension of .Z will mean unix's compress program has
been used to compress the file. An extension of .TZ will mean the file is
'tarred' and compressed under unix.

3b. Occasionally files may be 'split' to reduced the overall size of a file
for transfer. In this case the file name will be .001, .002,
etc. It is to be assumed that these files will be compressed, they may also be
'tarred'. A corrisponding ascii text file should be included to discribe the
split file, .

3c. The text file .txt may also be used to describe a file and
it's related/asscociated, documentation/source/binary files located in other
areas.

3d. The corrisponding Dos utilities to manage these files will be made
available.

3e. Other compression programs common to Dos will be allowed for Ascii
files for the purpose of compression of related materials for transmittal,
and viewing on Dos based systems. Although all formats are supported by Unix
keep in mind that the utilities ARE NOT distributed on the original Linux
disks, and as such should not be used often.

4. There will be an Echo set up between all the the LNXNRC's called
LNXRCECHO. This Echo is strictly for the use of the PDNRC's in each Zone, and
for nodes that are a gateway to another Zone. Normal downstream nodes should
not have access to this echo.

5. There will be another Echo set up between all the the LNX nodes called
LNXMANAGE. This Echo is for the use of the LNX nodes in each Zone, and for
nodes that are a gateway to another Zone. Users should not be allowed access
to this area. This area is not mandatory, but suggested.

6. There will be another Echo set up for all the nodes picking up files
from the LNXRC's called LNXECHO. This Echo is for the users of individual
BBS's so those users of files that are sent through the LNX can have a
medium that is appropriate to discuss the installation of Linux and it's
related files. This area is not mandatory, but suggested.

7. ALL LNX nodes will have session-level passwords with each other. This
is simply to ensure the integrity of the LNX.

8. ANY file that comes to your system via the LNX should have the CRC
checked by TICK as it is processed. No archive advertisments should be placed
into the archive while the file is still being sent to another LNX node. This
idea is not meant to advertise a BBS, but to move files. This also corrupts
the CRC info, and it will cause TICK to bomb out on the CRC check. Nodes
should not re-archive the file until that node is sure ALL the downstream
nodes have it. Please pay attention to authors wishes when re-packing files.

9. The LNX is free, and should always be just that. However, sometimes
it costs money to move these files, so there is no problem with asking LOCAL
downstream nodes to help pay for the phone call when picking up the LNX. The
LNX should be allowed to travel as far and to as many nodes as possible.

10. Any changes/additions to this policy should be coordinated with the
International LNX Coordinator, Greg Naber at 1:260/[email protected] alternate
address is [email protected] for those with Internet/UUCP access.



Catch Ya Later!

Greg
---
S'Qually Holler's BBS Renton, Wa. 1-206-235-0270 USR HST D/S
1:343/[email protected]
squally.halcyon.com


[next article]
From: [email protected] (Jim Burns)

Subject: Re: Problems with serial port drivers??
Message-ID: <[email protected]>
Date: 12 Apr 92 20:35:10 GMT
References:
Organization: Georgia Institute of Technology
Lines: 19

in article , [email protected]
.FI (Mika Pekka Liljeberg) says:

Sorry - we don't get alt.os.unix here, so I couldn't crosspost.

> using the same IRQ line simultaneously. Isn't it possible to poll all the
> UARTS on the IRQ line in the interrupt handler? Ie. you'd need a slightly

You're not *supposed* to be able to do this. The ISA interrupts are level
triggered, which I believe means that after the ICU latches the first
interrupt, a 2nd one won't still be active after the first one is serviced.
You might try polling all devices on any interrupt, not just poll for the
*first* device that needs service. Don't know how successful that would be.
The MCA interrupts are level triggered, so this is not supposed to be a
problem there.
--
BURNS,JIM (returned student)
Georgia Institute of Technology, 30178 Georgia Tech Station,
Atlanta Georgia, 30332 | Internet: [email protected]
uucp: ...!{decvax,hplabs,ncar,purdue,rutgers}!gatech!prism!gt0178a


[next article]
From: [email protected] (P Thomson)

Subject: swapping continued ?
Message-ID: <[email protected]>
Date: 12 Apr 92 20:55:40 GMT
Organization: Edinburgh University
Lines: 12


Sorry, I should of added the line

swapon /dev/hda3


to my previous posting


Thanks to the couple of guys whom pointed this out

Peter


[next article]
From: [email protected] (Blake McBride)

Subject: passwd program?
Summary: Is there a passwd program for linux?
Keywords: passwd
Message-ID:
Date: 13 Apr 92 02:55:34 GMT
Organization: Netcom - Online Communication Services (408 241-9760 guest)
Lines: 7

Is there a passwd program for linux? If so where may I find it?

--
Blake McBride Home (615) 790-8521
3020 Liberty Hills Drive Work (615) 790-1088
Franklin, TN 37064 [email protected]
U.S.A.


[next article]
From: [email protected] (Rob Coleman)

Subject: Parse error
Message-ID: <[email protected]>
Date: 13 Apr 92 03:32:52 GMT
Organization: Georgia Institute of Technology
Lines: 23

I had the parse error problem for quite a while, but had I been able to
see through my own stupidity, I would have fixed it a lot sooner. The
problem began with the distribution version of tar (at least the one
that I got on the .95a root disk) being bad. It would randomly replace
colons (:) with null bytes ($0), and in an executable, this really
mangled the file. I've known some people to never have a problem with
this version of tar, but obviously a priviledged few run into it 🙂 I
was instructed to get either the tar in the utils.tar.Z file on tsx-11,
or just build a new tar (from tar-1.10.tar.Z at prep.ai.mit.edu) and use
them rather than the one that came on the root disk.

These solutions didn't seem to work because I kept replacing a the file
named "tar" in /usr/bin with the new version of tar, and the file
actually being called was in /bin. So, each time I tried a different
version of tar, it was calling the old, corrupt one. So, ever since
I corrected that little oversight, all has been well. This same problem
was encountered on a friend's system, so it may very well fix everyone's
problem. Hope that helps, it did for me!
--
Rob Coleman ][ Science, like nature, must also be tamed.
[email protected] ][ With a view towards its presevation..
Georgia Tech, P.O. Box 30449 ][
Atlanta, GA 30332 ][ -Rush/Natural Science


[next article]
From: [email protected] (Greg Martin )

Subject: Help needed with OS topic
Message-ID: <[email protected]>
Date: 13 Apr 92 02:40:31 GMT
Sender: [email protected]
Organization: University of Missouri - Rolla
Lines: 24

I seem to remember a discussion going on a while ago
on alt.os.linux pertaing to micro-kernel design versus
monolithic kernel design. I'm in an OS class this
semester and we have to write a paper, and I'd like
to write it on this subject. Unfortunately, I'm having
a lot of trouble finding information on the subject.
(after spending 4 hours in the library today, and
coming up with nothing, I'm a bit frustrated.)

Can anybody point me to articles/materials/etc with
information about the advantages/disadvantages in
these designs?

Thanks,
Greg


""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""
" O softly singing lute, / See with my tears thou time do keep. "
" Yet softly, gentle strings, / Agree with love that cannot sleep. "
" Sorrow hyst whenas it sings. / When tears do fall then sighs arise. "
" So grief oft shines in most sad eyes, / Yea love through heart it dies. "
" -Francis Pilkington, 1624 "
" Greg Martin [email protected] What disclaimer??? "


[next article]
From: [email protected] (Paul Nulsen)

Subject: HD timeout under 0.95c+ - what is the problem?
Summary: Is the HD timeout error an indication of faulty hardware
Keywords: HD timeout 0.95c+
Message-ID: <[email protected]>
Date: 13 Apr 92 03:08:41 GMT
Organization: University Of Wollongong
Lines: 26


My system is a 33 MHz 80486 using the AMI bios. It has 8 Mbyte memory and a
200 Mbyte Seagate IDE disk. The disk seems to be fine under DOS (before
making the changes listed below).

I have the problem that others have reported of HD timeout errors under
Linux - and using the 0.95c+ kernel does not seem to improve matters. The
problem is so bad on my system that Linux is unusable, so I can't even try
to investigate the cause.

I have tried: slowing down the AT bus (to CLKI/6, which I guess means about
5.5 MHz); adding a wait state to the AT bus cycle; slowing down processor
accesses to main memory (adding wait states) and slowing the processor
speed. Each of these things improve matters a little, but the problem
remains.

Does anybody know: 1) what exactly is the cause of the HD timeout problem;
2) whether it indicates that my drive or IDE controller are faulty?

I have asked this question before and got no response. I would be grateful
for any sort of answer because I need to sort this out before the warranty
on the machine expires.

Thanks
Paul Nulsen
[email protected]


[next article]
From: [email protected] (E Elu)

Subject: news programs
Message-ID: <[email protected]>
Date: 12 Apr 92 11:44:52 GMT
Sender: [email protected] (Usenet system)
Organization: Monash University
Lines: 12


I am trying to get a news feed up and running on my Linux system, and I
was wondering if anyone has managed to get a newsreader an delivery agent
such as rn and c-news up and running. I have had no joy as yet though I
managed to get the news reader Tass compiled but it failed the linking
with libcurses.a in gcc 2.1. I included the curses.h file and put a
-lcurses at the end of the compilation string, but still no go.
Any suggestions?

thanks in advance

Ernie Elu


[next article]
From: [email protected] (Guido Muesch)

Subject: Re: Scheme, Lisp for Linux?
Keywords: scheme lisp linux-ports
Message-ID:
Date: 13 Apr 92 10:50:47 GMT
References: <[email protected]>
Sender: [email protected] (Newsfiles Owner)
Organization: RBI - RWTH Aachen
Lines: 32
Nntp-Posting-Host: messua

[email protected] (Dennis Petrocelli) writes:


> Is anyone working on porting any version of scheme or lisp to
>linux? I would be VERY appreciative of anyone's attempts to do so. I
>have successfully compiled a C implementation of Scheme, siod-2.7
>(Scheme in One Defun, from world.std.com), with gcc-1.4 on linux .95a.
>I am hoping that someone might have tried to port a more substantial
>version, such as scheme 7.1, from altdorf.ai.mit.edu.

>Dennis Petrocelli
>[email protected]

> BTW, way to go LINUS, et al !!!!!!!!!!!!!!!!!

> OS/2 - half the OS at twice the price!

Well I am just a beginner to Linux but I tried to compile 'scm' which is
also a version of scheme. I had nearly no problems with it. The only
problem concerns floatingpoint, because there is no support for single
precision float in the gnulib (I think). I also compiled an old version
of Xlisp (1.7 or so) with -Dfloat=double because of the missing single
precision float. This one works fine whereas scm produces a math trap
error. I did this all using gcc 1.40 .....

Cheers Guido

--
/\ bitnet: odiug%[email protected]
/ \ email: [email protected]
__/ \ ________________________________________________________________
\/ snail: Guido Muesch, Ruetscher Str.165/1413, 5100 Aachen


[next article]
From: [email protected] (A. V. Le Blanc)

Subject: New interim release of Linux-0.95c
Summary: Install Linux 0.95c painlessly(?)
Message-ID: <[email protected]>
Date: 10 Apr 92 10:43:34 GMT
Followup-To: [email protected]
Organization: Computing Centre, University of Manchester
Lines: 232

I have now released the latest 'interim' version of Linux.
Let me remind anyone who might misunderstand, that this version
has no 'official' character. It is an experiment and, I hope, a
convenience ofor others, and it may provoke some interesting
discussion. To install it you will require the following files,
available by anonymous ftp from ftp.mcc.ac.uk (hpb.mcc.ac.uk),
130.88.200.7, in /pub/linux/mcc-interim/0.95c:

boot-xx.Z
util-xx.Z

where 'xx' is either 'UK' (for UK keyboards) or 'US' (for US keyboards).
If you use DOS to get these files onto floppies, you will also require

rawrite.exe
uncomp.exe

available from the same site in the directory
/pub/linux/mcc-interim/dos-utils. The 0.95c directory contains two
README files, one for the version of fdisk contained in this release,
and one describing the release in general, which follows:

This is the current version of my MCC (Manchester Computing Centre)
interim release of Linux. This is intended to double as an
installation/recovery system for a version of Linux 0.95c.
It installs a system using shared libraries based on gcc-2.1.

WARNING: this is an experimental release of a beta test version of
software compiled with a beta test compiler. There are some
problems with the kernel, and others with some of the utilities,
'make' in particular.

You need to get from the images directory the files *-UK.Z
(for UK keyboards) or *-US.Z (for US keyboards).
If you are creating the disks on a DOS system, you will also need
rawrite.exe and uncomp.exe or something similar; copies of these
and of their documentation are available in the directory
/pub/linux/mcc-interim/dos-utils.

Assuming you have transferred (in binary mode) the files rawrite.exe,
uncomp.exe, boot-xx.Z, and root-xx.Z to your DOS system, type

C:\>uncomp boot-xx
C:\>uncomp util-xx
C:\>rawrite
Enter source file name: util-xx
Enter destination drive: a:
Please insert a formatted diskette into drive A: and press -ENTER- :

At this point put a high density disk which has been formatted at
least once into your A drive. After rawrite has finished, then
remove the disk and continue:

C:\>rawrite
Enter source file name: boot-xx
Enter destination drive: a:
Please insert a formatted diskette into drive A: and press -ENTER- :

Once rawrite has finished, you may reboot your system, leaving the boot
disk in the A drive (assuming you want to install Linux on this system).
After the initial boot, you should see something like this:

Loading................................
Press to see SVGA-modes available or any other key to continue.

After pressing anything but , you should get information about
your system, ending with this:

Now you may take the boot disk out of your 0 (A) drive and
put the utilities disk in instead.
After you have done this, tell me if your 0 (A) disk is
a 3.5 inch disk (3) or a 5.25 inch disk (5): _

Assuming you have no problems, and that your system is Linux-compatible,
you should now see this after giving the correct answer:

/dev/ram /
/dev/fd0xxxxx /mnt
Further instructions can be found in the file /mnt/README.
Type 'more /mnt/README' to look at it.
#

If you follow instructions, you should see this:

# more /mnt/README
This is the root disk for installing the MCC (Manchester Computing
Centre) interim version of Linux. If you have followed the
instructions so far, and if nothing has gone wrong, you must now
set up your hard disk:

Use fdisk to create a primary partition for Linux. You may also
wish to set up other partitions for Linux, and/or a partition of
at least 5 megabytes for swap space. Give the command V (verify)
before writing your partition table.

Note: The fdisk supplied is my fdisk, version 0.92, which should
be able to read and create extended partitions correctly (I hope).
My fdisk deals with only one device at a time. The default is
/dev/hda. If you wish to repartition your second hard disk using
fdisk, you must type 'fdisk /dev/hdab'. Please note and believe
the following warning:

AFTER USING fdisk, REBOOT AND REMOUNT THIS DISK ON /mnt.

The warning means that if you change your partition table, the
operating system (Linux) does not know about the changes until
after you reboot. Naturally, you will type 'sync' a couple of times,
remove the utilities disk, and replace the boot disk before rebooting.
The message continues:

Then use mkfs and mkswap to initialise the partitions you created
above. If you do not remember their bloksize, use fdisk with the
command p to display it (and then q to exit). mkfs and mkswap
have this syntax:

mkxxx [-c] /dev/hdxxx nnnnn

where [-c] optionally checks the partition, hdxxx is the partition's
device as reported by fdisk, and nnnnn is the partition's size in
blocks as reported by fdisk.

Use the command 'swapon /dev/hdxxx' to activate swap space made by
mkswap.

Mount the new partition on /root and type install_root.

Note that to mount the partition, you will type 'mount /dev/hda1 /root'
or something similar. If you do type 'install_root', you will be asked
(among other messages)

Do you wish to install Linux now? Type 'no' to stop.
Yes (y) or no (n): _

At this point, since it is possible that you are installing Linux on
a partition which you have earlier used as a Linux file system,
or you may have aborted a previous installation attempt, the
program says:

If you wish, we can delete any files and directories already
on your new root disk. Do you wish to clean your new root disk?
Type 'no' to leave any files which may already be there.
yes (y) or no (n): _

If you continue, you will have a final chance to exit:

This is your last chance to stop before installing Linux.
Do you want to continue installing? Type 'no' to stop.
Yes (y) or no (n): _

If you type 'y' ( or 'yes'), you will have a long list of things
being installed (I hope). At the end of this list, you will see
the following message:

Linux has now been installed. You may now give the umount
command to unmount the utilities disk. Type 'mount' to see
the device name of the utilities disk. Then you can set up
a boot disk for your new hard disk root partition:

cd /root/etc
rdev Image /dev/hda3
replacing ^ with your root device
cp Image /dev/fd0H1440
replacing ^ with fd0h1200 if your 0 (A) drive is 5.25 inches.
sync

If you obey these instructions you can reboot from the floppy you have
just written. Note that this floppy must have been formatted at some
time.

The installed system takes up about 2.8 Mbytes of disk space.
It contains no source whatever (except for shell scripts).
In addition to Linux-0.95c, it contains nearly all binaries from:

ps-2.1 (from hlu) find-3.5 mtools-2.0.5
bash-1.12 gawk-2.13.2 patch-2.0.12u6
compress-4.01 grep-1.5 sed-1.08
diff-1.15 joe-0.1.4 shellutils-1.6
elvis-1.5 less-1.77 tar-1.10
fgrep-1.1 make-3.62 textutils-1.3
fileutils-3.2

as well as programs from BSD and various things posted on the
*.os.linux lists. The installed system (though not the boot/utilities
disks) contains (modified) versions of the poe-igl init as well as
the following commands in binary form:

awk basename cat chgrp chmod chown
clear cmp comm compress cp csplit
ctags cut date dd df diff
diff3 dir dirname doshell du echo
egrep elvis elvprsv elvrec encrypt env
ex expand expr false fdisk fgrep
find fmt fold free fsck gawk
getty grep groups head hexdump hostname
id install joe join kill last
less lesskey ll ln login logname
ls lsf make mattrib mcd mcopy
mdel mdir mformat mkdir mkfifo mkfs
mknod mkswap mlabel mmd more mount
mrd mread mren mtype mv mwrite
nice nl nohup od paste patch
pathchk pr printenv printf ps rdev
ref rm rmdir sed setterm sh
sleep sort split strings stty sum
swapon sync tac tail tar tee
test top touch tr true tty
umount uname uncompress unexpand uniq
update users vdir vi virec wc
whereis who whoami write xargs yes

It also contains the help files for joe, less, and more. The
installation disks actually contain the following commands, which
are intended to be useful for recovering from messes:

cat chgrp chmod chown compress cp
dir doshell du fdisk fsck install_root
ll ln ls lsf mkdir mkfs
mknod mkswap more mount mv rdev
rm rmdir setterm sh stty swapon
sync tar umount uncompress vdir

There is virtually no documentation yet other than the installation
instructions. I shall be putting the source to this in
/pub/linux/mcc-interim/0.95c and its subdirectories over the next
few days. Please let me know of any improvements which could be
made to this release.

-- Owen
A. V. Le Blanc
Computing Centre
University of Manchester
[email protected]


[next article]
From: [email protected] (A. V. Le Blanc)

Subject: Re: HELP! Problems with rootimage.
Message-ID: <[email protected]>
Date: 10 Apr 92 10:56:18 GMT
References: <[email protected]
.ac.uk>
Followup-To: [email protected]
Organization: Computing Centre, University of Manchester
Lines: 16

In article <[email protected]> [email protected] writes:
>In article , [email protected] (Timo J
aakko Sillanp{{) writes:
>> I get rootimage.Z from nic.funet.fi (yes, binary mode is on),
>> uncompress it and write it to a floppy disk (1.44Mb) using
>> rawrite.exe.
>
>the rootimage is (I think !) for a 1.2Mb floppy. Try to find a rootimage for a

>1.44 disk - this could be the problem ...

The root and boot images from nic.funet.fi should work with both
3.5 and 5.25 inch high density floppies without any changes.
Be sure that the floppies are high density, though; both boot and
root floppies must be, even though the boot floppy is small.

-- Owen
[email protected]


[next article]
From: [email protected] (A. V. Le Blanc)

Subject: Re: Job Ctrl && $$->programmers (How?)
Keywords: jobs, bash
Message-ID: <[email protected]>
Date: 10 Apr 92 11:07:39 GMT
References: <[email protected]> <[email protected]>
Followup-To: [email protected]
Organization: Computing Centre, University of Manchester
Lines: 24

In article <[email protected]> [email protected]
I (Linus Benedict Torvalds) writes:
>Yes, there are several versions of bash existing - some of them are
>/very/ old. I don't know if the original bash.Z is still somehwre to be
>found, but it was not too long ago, and that is the version I used under
>0.01, long before linux had job control. That version certainly doesn't
>support job control.
>
>Versions of bash that support job control (including the builtins fg,
>bg, kill, jobs etc) are available on older root-disks, notably the 0.95
>(not "a") rootdisk. I don't know where to find a "standalone" version
>of bash - I know somebody ported bash 1.12 (the one on the rootdisk is
>1.11 patch-level 1, although I think "help" returns 0.00.1).

The version of bash which I ported to 0.12 is still available, still
works, and has job control. You can get it as bash-1.12.tar.Z in the
directory /pub/linux/mcc-bin by anonymous ftp from ftp.mcc.ac.uk
(= hpb.mcc.ac.uk), [130.88.200.7]. The version of bash which I
ported for my new 'interim' release of Linux (0.95c) is available
from the same machine, but does not support job control. This is
because I have not yet been able to compile it with job control
using gcc-2.1 and its new libraries.

-- Owen
[email protected]


[next article]
From: [email protected] (Lars Lindner)

Subject: tar files on floppy.
Message-ID: <[email protected]>
Date: 13 Apr 92 10:49:34 GMT
Sender: [email protected]
Organization: Ericsson Telecom AB
Lines: 23
Nntp-Posting-Host: eos10c02.ericsson.se

OK here's one more silly question.

When I have both boot and root images installed on HD and i want to
install more SW downloaded as .tar.Z files, in what way vill i get
them on a floppy to be able to install them on my Linux machine.

* Shall i use "rawrite" from DOS to put the files on floppy?

* I rememer reading something like "Do not rawrite compressed files.".
Whay??

* How to do with files lager than the capasity of my floppy when not compressed?

* Is there another way to install .tar.Z files via floppy?

My Linux machine uses 5 1/4" floppies and the only way for me to write
floppies in that format is to use another DOS machine.

Thanks in advance from a nowise.


Lars Lindner
[email protected]


[next article]
From: [email protected] (Drew Eckhardt)

Subject: Re: tar files on floppy.
Message-ID: <[email protected]>
Date: 13 Apr 92 13:20:29 GMT
References: <[email protected]>
Sender: [email protected] (The Daily Planet)
Organization: University of Colorado at Boulder
Lines: 44
Nntp-Posting-Host: caesar.cs.colorado.edu

In article <[email protected]> [email protected] (Lars Lind
ner) writes:
>OK here's one more silly question.
>
>When I have both boot and root images installed on HD and i want to
>install more SW downloaded as .tar.Z files, in what way vill i get
>them on a floppy to be able to install them on my Linux machine.
>
>* Shall i use "rawrite" from DOS to put the files on floppy?
>

Either that, OR mtools.

>* I rememer reading something like "Do not rawrite compressed files.".
> Whay??

When you rawrite to a floppy, you write to the raw device. The file size,
etc are lost, and are not encoded in a compressed file as they
are with a .tar file. Solutions to this problem are to :

1. Tar the .tar.Z file again, and rawwrite the resulting tar.

2. Use dd, and use a combination of bs, count to truncate the .Z file
to the correct length when you read it from the
raw device.


>* How to do with files lager than the capasity of my floppy when not compressed
?
>

1. Split them into smaller parts, read them with the normal
methods, and join them under linux with
cat foo.1 foo.2 foo.3 big.tar

2. use a multi-volume tar. Linux doesn't create these too well, but
reading isn't a problem. There is also a GNUish Tar
for DOS that should? be able to write these.

3. Uncompress, untar, and retar in smaller pieces.



>* Is there another way to install .tar.Z files via floppy?

mtools, dd, tar methods as mentioned above.


[next article]
From: [email protected] (David Engel)

Subject: Help Removing Bad Directory
Message-ID: <[email protected]>
Date: 13 Apr 92 15:03:54 GMT
Organization: Optical Data Systems, Inc.
Lines: 17

I ran into my first serious, file-system problem this weekend. Luckily,
the only things lost were either backed-up or could easily be recreated.
Fsck -a was able to repair all problems but one. I have a directory that
I can't remove. Fsck reports an error about '.' and '..' not being first
and can't, or won't, repair the damage. Rm -rf reports some error about
the directory not being mounted on device xyz and fails saying that the
directory isn't empty.

Does anybody know of a way, short of mkfs, to remove this directory. If
not, I'll live with it and wait until we get FFS (or equivalent) before
running mkfs.

David
--
David Engel Optical Data Systems, Inc.
[email protected] 1101 E. Arapaho Road
(214) 234-6400 Richardson, TX 75081


[next article]
From: [email protected] (Juergen Henke)

Subject: Re: tar files on floppy.
Message-ID:
Date: 13 Apr 92 15:56:36 GMT
References: <[email protected]> <1992Apr13.132029.8256@colorado.
edu>
Sender: [email protected] (USENet News Administration BelWue)
Organization: /users/juh/.organization
Lines: 25
In-Reply-To: [email protected]'s message of Mon, 13 Apr 1992 13:20:29
GMT

In article <[email protected]> [email protected] (Dre
w Eckhardt) writes:

When you rawrite to a floppy, you write to the raw device. The file size,
etc are lost, and are not encoded in a compressed file as they
are with a .tar file. Solutions to this problem are to :

[...]

I've done this now several times, with GNU-tar (the normal linux tar) it's no
problem to read a rawritten (:-)) compressed tar from a floppy, just use
(for example)

tar xvZf /dev/PS0
^
|
compress/uncompress option

Juergen


--
_________________________________________________________________________
Juergen Henke, e-mail [email protected], PSI-mail PSI%4505016002::JUH_IPA
Fraunhofer-Institut f. Produktionstechnik u. Automatisierung
Nobelstrasse 12, D-7000 Stuttgart 80


[next article]
From: [email protected] (Jim Nelson)

Subject: Re: Problems with serial port drivers??
Message-ID: <[email protected]>
Date: 13 Apr 92 15:27:45 GMT
References: <[email protected]>
Organization: Chaos Central, Fargo ND
Lines: 33

In article <[email protected]> [email protected] (Jim Burns) writes:
>in article , [email protected]
i.FI (Mika Pekka Liljeberg) says:
>> using the same IRQ line simultaneously. Isn't it possible to poll all the
>> UARTS on the IRQ line in the interrupt handler? Ie. you'd need a slightly
>
>You're not *supposed* to be able to do this. The ISA interrupts are level
>triggered, which I believe means that after the ICU latches the first
>interrupt, a 2nd one won't still be active after the first one is serviced.
>You might try polling all devices on any interrupt, not just poll for the
>*first* device that needs service. Don't know how successful that would be.
>The MCA interrupts are level triggered, so this is not supposed to be a
>problem there.


Dunno about MCA. Never had the pleasure. I do know that *sometimes* you
can move a comx to another IRQ. If you have an internal modem, then you can
change it's irq if you like voiding any warranties you might have on your
hardware. One of the pins on the modems bus fingers is the apropriate IRQ
line. A little soldering can re-arrange these. A friend of mine is using
this techniqu with great success. He has his modem on com3 and a mouse on
com1, and they both work simultaneously.

I have the doc for it at home, I think. Email me and I'll try to find it
and mail it back to you.

Dependng on your hardware, the above may or may not work for you.

--
Jim, in the Land of the Lost. |Disclaimer: These are probably
ObQuote: Do Not Meddle in the Affairs of | opinions. I'm probably not
Wizards, For You are Crunchy, | supposed to have any.
and Good with Ketchup. |



[next article]
From: [email protected] (Chris P. Sullivan)

Subject: Questions from a beginner...
Keywords: 1.44 A: drive, compiling, shells
Message-ID: <[email protected]>
Date: 13 Apr 92 15:08:08 GMT
Organization: University of Massachusetts Dartmouth
Lines: 35

Hello everyone,

First, let me applaud everyone who has helped develop and/or document Linux,
my roommate and I installed it on my machine this weekend, and it went VERY
smoothly (as smooth as Unix-type os's can go, I suppose), this is of course
comparing it to the installation of Minix, which was pure hell.
Now that I have the basic 0.95a system installed, I have a couple of questions.

First, the only thing in the installation process that did not work was the
install, mkdev, mktree scripts, for some reason they would not execute,
it would just respond with a :not found error. So we ended up printing
the scripts and doing it by hand. My limited UNIX experience leads me to
believe it has to do with the wrong shell for these scripts. WHAT OTHER SHELLS
ARE AVAILIBLE FOR LINUX? HOW DO YOU INSTALL THEM? Do you need to re-compile
the kernal to make them work?

Next, following the list of MKNOD's in the mkdev script we set up PS0,
according to the documentation, this makes ps0 an A: drive 1.2MB. When
we were RAWRITEing to floppies (1.44) to move them to Linux, if the file
was larger than 1.2MB, we would get RESET-FLOPPY CALLED errors. When we
would mount the floppy, it would work to a point and we would get scattered
RESET-FLOPPY CALLED errors. HOW WOULD WE MAKE LINUX UNDERSTAND THAT WE HAVE
AND A: DRIVE THAT IS 1.44MB?

Finally, we have access to DECstation 3100s running ULtrix, would it be
for us to compile things on the workstations and then download the executable
to my PC/Linux machine since these machines have far more disk/memory capacity
than mine does? Or will I run into major incompatability problems?

Any responses will be appreciated.

Chris Sullivan
UMass Dartmouth

[email protected]


[next article]
From: [email protected] (Steven Kraft)
Newsgroups: comp.os.linux,alt.os.linux

Subject: New project: S3 Acc. video support for Linux.
Keywords: S3, 86C911, X, local bus
Message-ID: <[email protected]>
Date: 13 Apr 92 17:07:30 GMT
Reply-To: [email protected]
Lines: 113
Nntp-Posting-Host: perdita.gac.edu

To all who are interested in (or curious about) direct (local) bus video and S3

accelerated video cards being used with Linux:

I have to decided to start discussion and development of S3 accelerated

video support in Linux.
Consider the following scenario: Suppose you are about to buy a new
machine for Linux, and you find a motherboard with a proprietary direct bus
SVGA card with an S3 chipset. It is very fast, but you are not sure if you
should buy it since it's new and drivers may not be available yet. You could
also buy an S3 accelerated card for the ISA bus and stay with your current
motherboard (or at least save a few bucks on the cost of a direct bus system),
and get the benefits of S3 speed without the moderately higher price of a new
motherboard with a direct bus. So, should you buy a tried and true ET4000 card

on an ISA bus system instead of newer technology?
(BTW, as I will mention later, Orchid does have the ProDesigner IIS/D
which will let you have direct bus video and stay with an older video
standard.)
I have an Orchid Superboard 486/33 motherboard with an Orchid
Fahrenheit 1280/D 32 bit local bus video card. Linux 0.12 works fine on it. I

hope to get 0.95c installed as soon as I can, and that will also work.
However, the fancy S3 accelerator chip functions of the Fahrenheit
1280/D are not yet accessible to Linux (X Windows being one goal) since no
driver has been written yet. I hope to write that driver myself in the future,

but I am still collecting info on that card.
This morning I received S3 Incorporated's manual on the "86C911 GUI
Accelerator." I ordered it from Orchid Technology for $25. If you want one
for yourself, just call Orchid's Tech Support and they will sell it to you.
(Numbers are below.) The manual looks good, but I have only had a few minutes
to glance through it. It states that: "the S3 86C911 is a high-performance
GUI accelerator specifically designed to accelerate applications running under
Windows, Presentation Manager, and X." I will try to contact S3 Inc. to see
what they might have already on the X driver.

I cannot say how long it might take to write the driver though...


The beta version will be posted to an ftp site as soon as I have gotten

that far and, of course, X Windows must be working also.
For now my video card looks just like any plain VGA card as far as
Linux is concerned.
To answer the question of how much faster the direct bus systems are
than those on an ISA bus, I will discuss it briefly. The Fahrenheit 1280/D is
at least twice as fast as its 16 bit counterpart (the Fahrenheit 1280)
depending on what kind of operation it is performing, according to benchmark
results published by Orchid. Since the Fahrenheit 1280 (the 16 bit card) is
the second fastest S3 card reviewed recently in Byte magazine, it is no slouch
either. The 32 bit Fahrenheit 1280/D is a very fast video card, because it has

both a fast S3 chipset design and a very fast direct bus.
Orchid does have another video card besides the Fahrenheit 1280/D for
their Superboard motherboard's direct bus video slot. It is the Prodesigner
IIs/D, a very fast (non S3-accelerated) VGA card for those who want the
standardization of VGA and the speed of direct bus video.
Now some info on the Fahrenheit 1280's modes. The 1280 * [960|1024]
modes are interlaced. The refresh rate can be set to 43 or 48 Hz. The 640 *
480 and 800 * 600 modes can be set to 60 or 72 Hz, and 1024 * 768 can be set to

60 or 70 Hz. The primary way of setting refresh rates is through switches which

are easily accessible through a slot in the card's back panel. A small
screwdriver is all it takes, no need to open the computer. The other way is a
resident driver for DOS (size of about 4k in memory I think) called FSCAN. It
prompts you through setting the horizontal and vertical position adjustments
(to center the image at each resolution) and the refresh rates. Mine is set to

72 Hz at 640 * 480 and 800 * 600, 60 Hz at 1024 * 768, and 48 Hz at 1280 *
[960|1024]. These are the maximum refresh rates for all but the 1024 * 768
mode. My Samsung Syncmaster 4 is supposedly able to work at 70 Hz refresh at
1024 * 768 but it doesn't for some reason. Since the DOS utilities won't work
in Linux, refresh rates will need to be set using the switches.
I will be writing this driver in my spare time :-), and it might be a
while until I am done, but I will gladly work together with others even before
a beta release is ready. In the end it should be a lot faster than a generic
ET4000 card because of 32-bit communication and bitblt's, etc. And until this
driver is ready, you can use these cards as standard VGA.
The various S3-based cards are basically the same except for setting
the modes or whatever. (I could be wrong here.) I think that they are very
much the same, and that if any changes are needed, #define's will do the trick.

I think that local bus video cards look just like a card on an ISA bus as far
as software drivers are concerned. They have a memory address just like any
other video card. The difference is that the data is sent to the card at 32
bits. (Correct me if I am wrong.)
Accelerated video will become a standard among OS/2 and Win users, so
it will drop in price. The number of Linux users who will have these will
increase. And we will be able to use X Windows once all the software is
written.

Soon I will start a channel on the linux-activists list for this
discussion. If you have anything to contribute, please do, either through
email to me, the soon-to-be-created list channel, or comp.os.linux.

If you would like to talk to Orchid about their systems, you can reach
them at:
Orchid Technology
45365 Northport Loop West
Fremont, CA 94538

General line (800) 7-ORCHID
Tech Support (510) 683-0323
BBS (510) 683-0327
BBS (9600) (510) 683-0555

P.S.
I will also be working on a driver for Fractint, a DOS app that has
been ported to X Windows and named XFractint. We will, of course, be able to
use XFractint once we get X Windows going under Linux to satisfy those of us
who like fractal graphics, but some may want to have a driver available for DOS

Fractint also. Email me if you would like to be a beta tester for the DOS
Fractint video driver.
Orchid has stated their intention to release an OS/2 2.0 driver for
their Fahrenheit series in the near future. If I get OS/2 before then, it
won't be too much fun in 640*480. (But what could be as much fun as Linux,
anyway?)


Steven Kraft [email protected] Comp Sci major
Micro/Minicomputer Consultant
Gustavus Adolphus College, St. Peter, MN 56082


[next article]
From: [email protected] (Douglas E. Quale)

Subject: better graphics for Nethack
Message-ID: <[email protected]>
Date: 13 Apr 92 14:32:16 GMT
Sender: [email protected] (The News)
Organization: Undergraduate Projects Lab, UW Madison
Lines: 17

If there are any Nethackers out there, you can use the nicer DECgraphics
display if you make a small addition to your termcap. I had hesitated
posting this before because earlier releases didn't quite get all the
vt100 graphics characters right, but 0.95c+ has changed some of console.c
and everything seems ok now. Great work.

Use ``export NETHACKOPTIONS=DECgraphics'' and make the following
addition to your termcap entry to add the alternate character set start
and end strings. I also added the km (has meta key) flag so we can end
all the ``Why doesn't the Alt key work in emacs?'' questions. These
additions probably belong in the root disk distribution of /etc/termcap.

km:as=^N\E(0:ae=\E(B:

--
Doug Quale
[email protected]


[next article]
From: [email protected] (Bowen Goletz)

Subject: Re: Questions from a beginner...
Keywords: 1.44 A: drive, compiling, shells
Message-ID: <[email protected]>
Date: 13 Apr 92 17:08:57 GMT
References: <[email protected]>
Sender: [email protected]
Organization: Purdue University Computing Center
Lines: 33

In article <[email protected]> [email protected] (Chri
s P. Sullivan) writes:
>Hello everyone,
>
>First, the only thing in the installation process that did not work was the
>install, mkdev, mktree scripts, for some reason they would not execute,
>it would just respond with a :not found error. So we ended up printing
>the scripts and doing it by hand. My limited UNIX experience leads me to

My experience was similar. I executed the shells using sh . The
problem you had was that the current directory "." was not in the path.
typing ./filename or adding . to the path solves this "problem."

>believe it has to do with the wrong shell for these scripts. WHAT OTHER SHELLS
>ARE AVAILIBLE FOR LINUX? HOW DO YOU INSTALL THEM? Do you need to re-compile
>the kernal to make them work?

The shells that I know of are ash, bash (many tcsh features), and I've heard of
tcsh and zsh in the works. Kernel recompile is generally not necessary (I've
never had to).

>RESET-FLOPPY CALLED errors. HOW WOULD WE MAKE LINUX UNDERSTAND THAT WE HAVE
>AND A: DRIVE THAT IS 1.44MB?

Re-run the mkdev script as described above. PS0 is the desired device name
for a 1.44MB floppy (unit 0).

-bg

--
________________________________________________________________________________
Bowen Goletz
Purdue University Midi 'n Music
[email protected] [email protected]


[next article]
From: [email protected] (0134)

Subject: Re: Patch to preset the SVGA mode
Summary: LINUX kills CMOS table
Keywords: Linux, Problems, CMOS
Message-ID: <[email protected]>
Date: 13 Apr 92 16:59:41 GMT
Sender: [email protected]
Organization: Universitaet Bielefeld
Lines: 14

HELP !!


On our 40 MHZ 386 0.95a kills the CMOS-Setup (PRam)

It took some time to get the machine to boot again.

Linux has done this several times already, while trying to mkfs or pfdisk, the
last time when I tried to login. My first attempt to login failed, the second
caused the machine to reboot and then I had to rebuild the PRam. (after zapping
it that is)

folker
[email protected]


[next article]
From: [email protected] (david.a.cuka)

Subject: Re: Questions from a beginner...
Keywords: 1.44 A: drive, compiling, shells
Message-ID: <[email protected]>
Date: 13 Apr 92 18:07:51 GMT
References: <[email protected]>
Organization: AT&T
Lines: 30

In article <[email protected]> [email protected] (Chri
s P. Sullivan) writes:
>Hello everyone,
>
[installation glee deleted, but still present]
>
>First, the only thing in the installation process that did not work was the
>install, mkdev, mktree scripts, for some reason they would not execute,
>it would just respond with a :not found error. So we ended up printing
>the scripts and doing it by hand. My limited UNIX experience leads me to
>believe it has to do with the wrong shell for these scripts. WHAT OTHER SHELLS
>ARE AVAILIBLE FOR LINUX? HOW DO YOU INSTALL THEM? Do you need to re-compile
>the kernal to make them work?

try "sh < install"
or
"chmod +x install"
"./install"

[floppy problems deleted]

>
>Any responses will be appreciated.
>
>Chris Sullivan
>UMass Dartmouth
>
>[email protected]


DAC


[next article]
From: [email protected] (Kevin Cummings)

Subject: Re: Modems and com port access
Message-ID: <[email protected]>
Date: 13 Apr 92 18:38:42 GMT
References: <[email protected]>
Sender: cummings@hammer (Kevin Cummings)
Organization: Prime Computer R&D
Lines: 41

In article <[email protected]>, [email protected]
hio-state.edu (William A Jones) writes:
> Now that I have a copy of UUCP for linux, I decided that it was time to
> see if my modem worked properly. So, I scanned all of the digest entries
> that had anything to do with the modem and decided that, under version
> 0.95c+, COM2 was assigned to /dev/ttys2.
>
> Following the FAQ, I put the following line in /etc/inittab
>
> ttys2:console:/bin/getty -m 1200 ttys2
>
> and rebooted. So far, so good.
>
> Now, kermit, uucico, pcomm, and everything else will not recognize that
> a modem exists. What have I done wrong?

While getty is running, it "owns" the serial port device (/dev/ttys2).
Kermit, uucp, pcomm, etc can't open it to use it. In order to fix this
problem, you need to get something like the acu.pkg that was posted
recently. On my Sun system (at work), the getty is intelligent enough to
release the serial line when some other process wants to use it for outgoing
(instead of waiting for incoming). The current getty in LINUX is not so
smart. The acu.pkg indictased that it might be better at this (tho I haven't
tried it yet!)

> If someone who has a working kermit, uucp, etc., would send detailed(!!)
> instructions on how to get set up I would be very grateful.

Gladly, after I download uucp and get it working B^).

=================================================================
Kevin J. Cummings Prime Computer Inc.
20 Briarwood Road 500 Old Connecticut Path
Framingham, Mass. Framingham, Mass.

Work: [email protected]
Home: [email protected]

Std. Disclaimer: "Mr. McKittrick, after careful consideration,
I've come to the conclusion that your new
defense system SUCKS..." -- War Games
=================================================================


[next article]
From: [email protected] (Kevin Cummings)
Newsgroups: comp.os.linux,alt.os.linux

Subject: Re: New project: S3 Acc. video support for Linux.
Keywords: S3, 86C911, X, local bus
Message-ID: <[email protected]>
Date: 13 Apr 92 18:11:02 GMT
References: <[email protected]>
Sender: cummings@hammer (Kevin Cummings)
Organization: Prime Computer R&D
Lines: 88

In article <[email protected]>, [email protected] (Steven Kraft) w
rites:
> To all who are interested in (or curious about) direct (local) bus video and S
3
> accelerated video cards being used with Linux:
>
> I have to decided to start discussion and development of S3 accelerated

> video support in Linux.

Being the owner of a brand spanking new Diamond Stealth VRAM S3 based video
card with 1MB of memory and the Hi-Color DAC, I share your interest in the S3
based video support category.

> This morning I received S3 Incorporated's manual on the "86C911 GUI
> Accelerator." I ordered it from Orchid Technology for $25. If you want one
> for yourself, just call Orchid's Tech Support and they will sell it to you.
> (Numbers are below.) The manual looks good, but I have only had a few minutes

> to glance through it. It states that: "the S3 86C911 is a high-performance
> GUI accelerator specifically designed to accelerate applications running under

> Windows, Presentation Manager, and X." I will try to contact S3 Inc. to see
> what they might have already on the X driver.

Great news! Maybe I should order up a copy, since Diamond doesn't provide
access to ALL the S3 supported modes in it's video BIOS. On the other hand,
Diamond does provide a VESA BIOS, but it doesn't support all the S3 modes either
.

> Now some info on the Fahrenheit 1280's modes. The 1280 * [960|1024]
> modes are interlaced. The refresh rate can be set to 43 or 48 Hz. The 640 *

> 480 and 800 * 600 modes can be set to 60 or 72 Hz, and 1024 * 768 can be set t
o
> 60 or 70 Hz.

So far, all this is pretty S3 standard, the Diamond supports these refresh rates
as well.

> The primary way of setting refresh rates is through switches whi
ch
> are easily accessible through a slot in the card's back panel. A small
> screwdriver is all it takes, no need to open the computer.

The Stealth only has 4 swithches on the back of its card. Two of them help
select refresh rates. I guess I'm lucky. Two of these four states support the
NEC 3 and NEC 4/5 series monitors (I have a NEC 3D).

> The other way is a

> resident driver for DOS (size of about 4k in memory I think) called FSCAN. It

> prompts you through setting the horizontal and vertical position adjustments
> (to center the image at each resolution) and the refresh rates.

Diamond supplies an SMODE program to fine tune your video configuration, and sel
ect
other monitors or special frequencies/syncs/etc.

> Since the DOS utilities won't work
> in Linux, refresh rates will need to be set using the switches.

Yes, agreed, however, if we can figure out what the BIOSes are doing when they s
et
modes, maybe we can duplicate some of that in protected mode code so we don't ne
ed
the BIOSes (they're hard to access once LINUX gets running!).

> The various S3-based cards are basically the same except for setting
> the modes or whatever. (I could be wrong here.) I think that they are very
> much the same, and that if any changes are needed, #define's will do the trick
.
> I think that local bus video cards look just like a card on an ISA bus as far

> as software drivers are concerned. They have a memory address just like any
> other video card. The difference is that the data is sent to the card at 32
> bits. (Correct me if I am wrong.)

Uh, BIOS interfaces are probably card specific, though VESA support can help her
e.
As far as card interfaces, 8, 16, and 32 bits are all possible. My card is a 16
bit
card with jumpers on board to restrict access to 8 bits if necessary (not recomm
ended
in an AT+ class machine!).

> Soon I will start a channel on the linux-activists list for this
> discussion. If you have anything to contribute, please do, either through
> email to me, the soon-to-be-created list channel, or comp.os.linux.

Great! Please include me in your discussions!

Use either E-Mail address below if you start up a mailing list. Thanx!

=================================================================
Kevin J. Cummings Prime Computer Inc.
20 Briarwood Road 500 Old Connecticut Path
Framingham, Mass. Framingham, Mass.

Work: [email protected]
Home: [email protected]

Std. Disclaimer: "Mr. McKittrick, after careful consideration,
I've come to the conclusion that your new
defense system SUCKS..." -- War Games
=================================================================


[next article]
From: [email protected] (Drew Eckhardt)

Subject: Re: Problems with serial port drivers??
Message-ID: <[email protected]>
Date: 13 Apr 92 18:46:43 GMT
References: <[email protected]>
<[email protected]>
Sender: [email protected] (The Daily Planet)
Organization: University of Colorado at Boulder
Lines: 24
Nntp-Posting-Host: caesar.cs.colorado.edu

In article <[email protected]> [email protected] (Jim Nelson) writes
:
>In article <[email protected]> [email protected] (Jim Burns) writes
:
>>in article , [email protected]
ki.FI (Mika Pekka Liljeberg) says:
>>> using the same IRQ line simultaneously. Isn't it possible to poll all the
>>> UARTS on the IRQ line in the interrupt handler? Ie. you'd need a slightly
>>
>>You're not *supposed* to be able to do this. The ISA interrupts are level
>>triggered, which I believe means that after the ICU latches the first
>>interrupt, a 2nd one won't still be active after the first one is serviced.
>>You might try polling all devices on any interrupt, not just poll for the
>>*first* device that needs service. Don't know how successful that would be.
>>The MCA interrupts are level triggered, so this is not supposed to be a
>>problem there.
>

100% false with ISA. ISA interrupts are edge triggered. It has nothing
to do with the ICU not letting you have multiple interrupts. The ISA
IRQ lines are tristated, and if more than one is enabled at a time,
they short each other out. Not as bad as it sounds, but it doesn't
work.

You are correct about MCA though - it uses level active IRQ lines,
and they are open collector.



[next article]
From: [email protected] (Greenup)
Newsgroups: comp.os.linux,alt.os.linux

Subject: all kinds of questions
Summary: HELP! I WANT MY DOS GOODIES W/O THE PENALTIES!
Message-ID: <[email protected]>
Date: 13 Apr 92 19:15:37 GMT
Sender: [email protected] (USENET News System)
Organization: Montana State University
Lines: 28


Hello netters!
I just booted up for the first time two days ago, and now I'm a happy
Linux user. Well, almost. I do have a few questions...

1. is "agetty" different than "getty", if so,
where do I find it,
what's the difference?
2. why doesn't "tar +help | more" work?
3. does "sync" do a diskpark, or do I have to round up a routine for that?
4. I've heard that there's some binary incompatibility coming up.
does the new work with the old or does the old work with the new,
when (+/- a week) is this expected to take effect?
is this tied in with the new compiler, or a new kernal goodie?
5. has anyone done any work on something similar to messydos's "snipper"
(a cut & paste print screen) or "buffit" (a scrollback buffer)
6. will the ABC release make it out before I leave school for the year?
(twice, one extra to make sure it doesn't have big bugs 🙂 )
(leave school ~= May 14)

MANY THANKS--
john Greenup
[email protected]

--
john Greenup | snail mail:
Internet: [email protected] | 414 Langford Hall
Bitnet : not yet | Bozeman, MT 59771


[next article]
From: [email protected] (Kevin Cummings)

Subject: Re: Needed: a guide to setting up swap space.
Keywords: linux, swap, setup, installation, mkswap, swapon
Message-ID: <[email protected]>
Date: 13 Apr 92 18:52:15 GMT
References: <[email protected]>
Sender: cummings@hammer (Kevin Cummings)
Organization: Prime Computer R&D
Lines: 123

In article <[email protected]>, [email protected] (Clifford A Adams) writes
:
> I recently installed Linux on my HD (/dev/hda1 mostly
> following I Reid's instructions), and things seem to be running OK. I
> have 4 meg of ram (2.7meg reported free after buffers, etc), which is
> enough for most activities. However, I'd like to add a few meg of
> swap space (especially since I like to use the full GNU Emacs while
> simultaneously compiling with GCC. 🙂 I have about 200k of various
> FAQs and guides, but none of them explicitly say much about how to set
> up swap space. I don't think I've been successful so far in figuring
> it out...

It was pretty easy for me. First off there are two ways to set up
swap space. You can create a swap file in your file system and page to it, or
you can set aside a disk partition, and page to that. I choose the latter.

Step 1) Using pfdisk, I made the disk partition. It was a little difficult
finding the right fdisk to figure out the exact device name to use,
but a little common sense came through in the end. For me /dev/hda1
is my primary DOS partition, and /dev/hda2 is my extended DOS partition.
Using pfdisk, I set up partitions 3 and 4 and labeled them as
type Linux/Minix file-system, and Linux-swap partitions. The latter
wasn't really necessary, I could have labeled it as anything.
So, I have LINUX root on /dev/hda3, and LINUX swap on /dev/hda4.
(If you already have LINUX on your hard drive, you only have to setup
your swap partition. I just happened to do both at the same time.)

Step 2) I ran "mkswap -c /dev/hda4 xxxx" where xxxx is the amount of space
reported by fdisk as being available on the device. The -c
tells mkswap to check the media for bad spots (unrecoverable errors).
You don't want to page onto one of those!

Step 3) After the partition is ready for swapping, I turned it on via
"swapon /dev/hda4" to see if there would be any problems and ran that
way for a while.

Step 4) To make the swap area appear permanently in my system, I modifed /etc/rc
to do the swapon for me during system boot.

The prcedure for a file system swap file is similar to the above, except that th
e
swap device being used by the mkswap and swapon commands is a file instead of
a device. Create a file big enough to hold the amount of swap space you want,
and use that pathname in place of /dev/hdax in the above sequence.

> Here are my problems (using 0.95c+ boot floppy and a mostly
> 0.95a root on /dev/hda1):
>
> * "mkswap" was on my 0.95a rootdisk. There seems to be a
> newer version (later file date) at tsx-11 under the binaries tree
> (the later one is bigger). What are the differences, and which should
> I use? For the rest of my experiments

I don't know, I used the one on my root disk without problems.

> * Looking at "mkswap" I noticed a "-c" option--what does it
> do? Early on I thought that "-c" meant "create", and that otherwise
> mkswap would set up and use the swapfile. At the moment it seems like
> you use swapon to set up a swap file--at the first reading it seemed
> like swapon was only for swap partitions...

Like I said above, the -c causes a media check for bad spots.

mkswap configures the disk for swapping. (Like mkfs configures a disk
to be used as a file system.) swappon is the command which actually
tells the OS to use the new swap space.

> * I made a file "/dev/swapfile" using emacs to create a file
> of 1K spaces, then using cat to make 4K/16K/64K/256K/1meg files, and finally
> the 6 meg /dev/swapfile. I didn't know if it would need a little more
> space for tables or whatever so I added another 256K as extra.
> * Is there a better way to create the swapfile?

Under SUNOS, they have a mkfile command. I haven't seen one for LINUX yet.

> * Do I need to have more space than the #blocks I tell mkswap?

Nope, but if you have more space than you've configured with mkswap, it'll never
get used.

> * I noticed that the mkswap blocks are 1K--this should
> probably be noted since its different than the filesystem blocks...

So swapping swaps 1K bytes at a time. That is a memory page size.

> * Is there any way to tell how much swap space you are using or
> have left?

Under SUNOS I do a "pstst -T". I haven't seen a LINUX command yet to tell me
how much is actually in use. And seeing as gcc-1.4 didn't know when my disk
was full (other that getting unexpected EOF error messages during compiliation),
I wouldn't be surprised if LINUX doesn't detect "out of swap space" cleanly eith
er.
For now, I hope that 4MB of memory and 9MB of swap space is enough to not run ou
t of
until we get something that tells us swap usage.

> ...What I'd really like to see is a guide something like Reid's guide
> to installation for setting up swap space--one that would cover both
> swapfiles and swap partitions. (Perhaps it could be added as an
> appendix to the guide?). Just a list of the commands and what they do
> would be nice however...
>
> Finally, my first impression of Linux is: "zippy". Its been a
> long time since I've had the kind of quick response Linux gives, even
> when I worked on "unloaded" machines (where I'm the only user, the load
> avg is 0.00, etc...). Even when running GNU emacs the response is
> excellent (on my 40mhz 386 w/64K cache). It may seem like a
> small difference between the seeming immediate Linux response and the
> more typical 2/3-1 second reaction time, but it feels nice... (Its also
> nice not to wait for overloaded NFS fileservers, and being able to
> mess with the kernal is a plus. 🙂

Yeah, ain't it great?

=================================================================
Kevin J. Cummings Prime Computer Inc.
20 Briarwood Road 500 Old Connecticut Path
Framingham, Mass. Framingham, Mass.

Work: [email protected]
Home: [email protected]

Std. Disclaimer: "Mr. McKittrick, after careful consideration,
I've come to the conclusion that your new
defense system SUCKS..." -- War Games
=================================================================


[next article]
From: [email protected] (Bowen Goletz)
Newsgroups: comp.os.linux,alt.os.linux

Subject: Re: all kinds of questions
Message-ID: <[email protected]>
Date: 13 Apr 92 19:44:02 GMT
References: <[email protected]>
Sender: [email protected]
Followup-To: comp.os.linux
Organization: Purdue University Computing Center
Lines: 33

In article <[email protected]> [email protected] (Gree
nup) writes:
>
>Hello netters!
> I just booted up for the first time two days ago, and now I'm a happy
>Linux user. Well, almost. I do have a few questions...
>
>1. is "agetty" different than "getty", if so,

No significant difference, if any, just renamed. getty is the
more recent of the two.

>2. why doesn't "tar +help | more" work?

The help output is through stderr insted of stdio (why?) If you
want to pipe stderr, use the amperstand (&). For example: tar &> tar.help.
Then more the tar.help file. For some reason, stderr pipes aren't supported
with bash. You should be able to do a &| to pipe the output directly to
more in the first place.

>3. does "sync" do a diskpark, or do I have to round up a routine for that?

Sync flushes the disk cache. Very important! Most modern drives
auto-park on powerdown.

[abc release and binary incompatibility deleted]

-bg

--
________________________________________________________________________________
Bowen Goletz
Purdue University Midi 'n Music
[email protected] [email protected]


[next article]
From: [email protected] (Kevin Cummings)

Subject: Re: Questions from a beginner...
Keywords: 1.44 A: drive, compiling, shells
Message-ID: <[email protected]>
Date: 13 Apr 92 19:34:51 GMT
References: <[email protected]> <[email protected]>
Sender: cummings@hammer (Kevin Cummings)
Organization: Prime Computer R&D
Lines: 23

In article <[email protected]>, [email protected] (Bowen Golet
z) writes:
> In article <[email protected]> [email protected] (Ch
ris P. Sullivan) writes:
> >RESET-FLOPPY CALLED errors. HOW WOULD WE MAKE LINUX UNDERSTAND THAT WE HAVE
> >AND A: DRIVE THAT IS 1.44MB?
>
> Re-run the mkdev script as described above. PS0 is the desired device name
> for a 1.44MB floppy (unit 0).

Better yet, go find 0.95c or 0.95c+. This was supposed to fixed
the uncommon reset floppy problem that only a few people had.

=================================================================
Kevin J. Cummings Prime Computer Inc.
20 Briarwood Road 500 Old Connecticut Path
Framingham, Mass. Framingham, Mass.

Work: [email protected]
Home: [email protected]

Std. Disclaimer: "Mr. McKittrick, after careful consideration,
I've come to the conclusion that your new
defense system SUCKS..." -- War Games
=================================================================


[next article]
From: [email protected] (David Engel)

Subject: Re: all kinds of questions
Message-ID: <[email protected]>
Date: 13 Apr 92 20:51:38 GMT
References: <[email protected]>
Organization: Optical Data Systems, Inc.
Lines: 15

[email protected] (Bowen Goletz) writes:
: Then more the tar.help file. For some reason, stderr pipes aren't supported
: with bash. You should be able to do a &| to pipe the output directly to
: more in the first place.

Actually, standard error pipes are supported by bash. You just have to use
the traditional /bin/sh syntax like this:

command1 2>&1 | command2

David
--
David Engel Optical Data Systems, Inc.
[email protected] 1101 E. Arapaho Road
(214) 234-6400 Richardson, TX 75081


[next article]
From: [email protected] (Philippe Backouche)

Subject: System Architecture
Message-ID: <[email protected]>
Date: 13 Apr 92 20:13:46 GMT
References: <[email protected]> <[email protected]>
Sender: [email protected]
Reply-To: [email protected] (Philippe Backouche)
Organization: SCRI, Florida State University
Lines: 7


Could anyone explain me the difference of architecture between Linux, Mach an
d
Minix ? What makes Linux be so fast ?

Thanks for your answers....

Philippe.


[next article]
From: [email protected] (mr79)

Subject: When will "Linux 1.0" come to the world ?
Message-ID: <1992Apr13.215610.22632%[email protected]>
Date: 13 Apr 92 21:56:10 GMT
Sender: [email protected] (Net News)
Organization: National Tsing Hua University (HsinChu)
Lines: 5

Linux 0.95c is still in the period of development.
Everything is rather unstable.
when will Linux 1.0 come to the PC 386/486 world ?

J.H.Lin


[next article]
From: [email protected] (Drew Eckhardt)
Newsgroups: comp.os.linux,alt.os.linux

Subject: Re: all kinds of questions
Message-ID: <[email protected]>
Date: 13 Apr 92 22:34:28 GMT
References: <[email protected]>
Sender: [email protected] (The Daily Planet)
Organization: University of Colorado at Boulder
Lines: 39
Nntp-Posting-Host: ophelia.cs.colorado.edu

In article <[email protected]> [email protected] (Gree
nup) writes:
>
>Hello netters!
> I just booted up for the first time two days ago, and now I'm a happy
>Linux user. Well, almost. I do have a few questions...
>
>1. is "agetty" different than "getty", if so,

No real obvious differences.

>3. does "sync" do a diskpark, or do I have to round up a routine for that?

No. sync merely commits the buffercache to disk. Buffers will be
committed to disk when buffercache is full, or a sync() syscall
is performed. To prevent heinous catastrophees, update is
run from /etc/rc. update syncs the disks every 30 seconds.

Most hard disks these days autopark, so that shouldn't be a problem.

>4. I've heard that there's some binary incompatibility coming up.
> does the new work with the old or does the old work with the new,

Files linked with the new ld do not work with older kernels,
but older binaries run on the new kernel.

> when (+/- a week) is this expected to take effect?

The new a.out.h is already used in the .95c+ kernel,
and the new gcc2.1 ld uses it.

> is this tied in with the new compiler, or a new kernal goodie?

it's a header file upgrade to make us more 'normal'

>5. has anyone done any work on something similar to messydos's "snipper"
> (a cut & paste print screen) or "buffit" (a scrollback buffer)

There is a cut and paste via mouse kernel patch, and the kernel
supports screen dumps to a file.


[next article]
From: [email protected] (Rob Coleman)

Subject: Losing Characters
Message-ID: <[email protected]>
Date: 13 Apr 92 22:51:05 GMT
Organization: Georgia Institute of Technology
Lines: 25

Last night I was transferring a file using xcomm and the rz that is
distributed with it, and all was working well until the program went
to disk to write the data that it had just received. Using a USRobotics
Dual Standard communicating at 19.2kbps, a CRC error would be generated
each time Linux went to write something out to disk. This problem was
not at all due to processor load (I could run 4 infinite loops in other
tty's and not have it affect the transfer at all), but due seemingly to
interrupt latency. I have the modem running off of a 16550 UART, but
my tests show that at powerup, the 16 byte buffer is turned off. Should
I need to use the buffer for Linux to keep up with 19.2kbps transfers?
I've got some pascal code that turns the buffer on and off, and I'm
working on getting that translated into Linux useable code. When I get
that done, I'll stick it up (if there's interest).

I've got a 486/33 with two fairly quick Seagate hard drives, so I wouldn't
think that there would be much problem keeping up. Is there something I
may be doing wrong? I've also got a 5 meg swap partition set up..

Thanks in advance..

--
Rob Coleman ][ Science, like nature, must also be tamed.
[email protected] ][ With a view towards its presevation..
Georgia Tech, P.O. Box 30449 ][
Atlanta, GA 30332 ][ -Rush/Natural Science


[next article]
From: [email protected] (Drew Eckhardt)

Subject: Re: When will "Linux 1.0" come to the world ?
Message-ID: <[email protected]>
Date: 13 Apr 92 22:45:27 GMT
References: <1992Apr13.215610.22632%[email protected]>
Sender: [email protected] (The Daily Planet)
Organization: University of Colorado at Boulder
Lines: 18
Nntp-Posting-Host: ophelia.cs.colorado.edu


In article <1992Apr13.215610.22632%[email protected]> [email protected]
.tw (mr79) writes:
>Linux 0.95c is still in the period of development.
>Everything is rather unstable.

The only things unstable in Linux right now is caused by the lack of
limits and a reasonable scheme to insure that there is enough virtual
memory for every process running, and tar not handling EOF and consequently
multi-volume tars correctly.

I think that before Linux reaches 1.0, it definately needs to have the limit
system calls implemented, and a "real" filesystem on top of VFS,
such as Berkely FFS, or LFS.

To be on a par with what BSD / HURD / etc will be once they are out
of development, an X port should be finished,
and support for TCP / IP over "standard" ethernet boards, as well as
SLIP / CSLIP in the kernel should be added. QIC-02 tape support
would be nice too.


[next article]
From: [email protected] (Steven Kraft)
Newsgroups: comp.os.linux,alt.os.linux

Subject: Re: New project: S3 Acc. video support for Linux.
Message-ID: <[email protected]>
Date: 14 Apr 92 00:29:54 GMT
References: <[email protected]>
Reply-To: [email protected]
Lines: 38
Nntp-Posting-Host: nim.gac.edu

In article <[email protected]> writes:
! In article <[email protected]>, [email protected] (Steven
Kraft) writes:
! > To all who are interested in (or curious about) direct (local) bus video
! Being the owner of a brand spanking new Diamond Stealth VRAM S3 based video
! card with 1MB of memory and the Hi-Color DAC, I share your interest in the S3
! based video support category.

Great!

! > The primary way of setting refresh rates is through switches which
! > are easily accessible through a slot in the card's back panel. A small
! > screwdriver is all it takes, no need to open the computer.
!
! The Stealth only has 4 swithches on the back of its card. Two of them help
! select refresh rates. I guess I'm lucky. Two of these four states support

The Fahrenheit is similar. Two switches select refresh, one controls a speed
up option which can "hang" some systems (this switch is not used on my Fahr
1280/D since it can always run at full speed in its own direct bus slot), and
the 4th switch has no use on either the 16 or 32 bit versions!

! Diamond supplies an SMODE program to fine tune your video configuration, and
! select other monitors or special frequencies/syncs/etc.

Orchid has other utilities as well, such as FMODE.

! =================================================================
! Kevin J. Cummings Prime Computer Inc.
! 20 Briarwood Road 500 Old Connecticut Path
! Framingham, Mass. Framingham, Mass.
!
! Work: [email protected]
! Home: [email protected]

Steven Kraft [email protected] Comp Sci major
Micro/Minicomputer Consultant
Gustavus Adolphus College, St. Peter, MN 56082


[next article]
From: chojnows@osiris (Brian Chojnowski)

Subject: Where to get a hold of Linux
Keywords: linux
Message-ID: <[email protected]>
Date: 13 Apr 92 23:57:41 GMT
Sender: [email protected] (News)
Organization: University of Illinois at Urbana
Lines: 3


Where can I get ahold of this operating system? And what terms are
there to its use?


[next article]
From: [email protected]

Subject: help with back-space key in 0.95c+(boot) and 0.95a(root)!
Keywords: back-space key, 0.95c+ (boot), 0.95a (root)
Message-ID: <[email protected]>
Date: 14 Apr 92 00:22:01 GMT
References:
Sender: [email protected]
Followup-To: [email protected]
Distribution: na
Organization: Michigan State University
Lines: 8

i was able to make linux boot on my 386sx (4Mb) and use the root
on floppy. after the login as root it came out that back-space
does not work -- pressing of backspace doesn't move the cursor
and pressing of return sends the buffer as is! also, ^H is echoed
on the screen as is dosen't affect the info in the keyboard buffer.
any pointers will be greatly appreciated.

-->ishwar ([email protected])


[next article]
From: [email protected] (Jim Winstead Jr.)
Newsgroups: comp.os.linux,alt.os.linux

Subject: Re: all kinds of questions
Message-ID: <[email protected]>
Date: 14 Apr 92 01:35:38 GMT
References: <[email protected]> <1992Apr13.223428.29499@colo
rado.edu>
Sender: [email protected] (The News System)
Organization: Harvey Mudd College, Claremont, CA 91711
Lines: 19

In article <[email protected]> [email protected] (D
rew Eckhardt) writes:
>In article <[email protected]> [email protected] (Gre
enup) writes:
>>
>>Hello netters!
>> I just booted up for the first time two days ago, and now I'm a happy
>>Linux user. Well, almost. I do have a few questions...
>>
>>1. is "agetty" different than "getty", if so,
>
>No real obvious differences.

To clear up any questions: The 'getty' on the 0.95a root disk is the
*exact* same program as 'agetty' on the 0.95 root disk. I renamed it
for the 0.95a release to be consistent with other Unix implementations.
--
Jim Winstead Jr. (CSci '95) | "Catch a fish!"
Harvey Mudd College | -Geddy Lee,
[email protected] | San Diego Sports Arena
Disclaimer: Mine, not theirs! | January 20, 1992


[next article]
From: [email protected] (Jim Winstead Jr.)

Subject: Re: When will "Linux 1.0" come to the world ?
Message-ID: <[email protected]>
Date: 14 Apr 92 01:41:39 GMT
References: <1992Apr13.215610.22632%[email protected]>
Sender: [email protected] (The News System)
Organization: Harvey Mudd College, WIBSTR
Lines: 29

In article <1992Apr13.215610.22632%[email protected]> [email protected]
.tw (mr79) writes:
>Linux 0.95c is still in the period of development.

I follow you here....

>Everything is rather unstable.

"Everything"? "rather unstable"? What are you basing these
judgements on? My experience with Linux indicates it is more stable
than DOS+DESQview on my machine, and allows me to get even more done,
with the small exception of word processing. (Which I hope to
overcome, once I figure out how the !@$*!@ to set up TeX. I
appreciate the absolute lack of instructions included with the TeX
distributed for Linux. Grr.)

>when will Linux 1.0 come to the PC 386/486 world ?

How about when Linux is stable? Really, Linus has said before that he
had intended Linux 1.0 to be released when it was able to run GCC. It
is obviously far beyond that point now, with the addition of SCSI
support, printer support, virtual console support, serial port
support, etc.

Linux 1.0 will be sooner if you help....
--
Jim Winstead Jr. (CSci '95) | "Catch a fish!"
Harvey Mudd College | -Geddy Lee,
[email protected] | San Diego Sports Arena
Disclaimer: Mine, not theirs! | January 20, 1992


[next article]
From: [email protected] (Dennis Flaherty)

Subject: Re: Patch to preset the SVGA mode
Message-ID: <[email protected]>
Date: 13 Apr 92 19:13:37 GMT
References: <[email protected]> <[email protected]>
Organization: Miles, Inc. Diagnostics Divis.
Lines: 121

In article <[email protected]> [email protected] (Andy I. Liaw) writes
:
> In article <[email protected]> [email protected] (Dennis Flahert
y) writes:
> >
> >If you are tired of hitting RETURN and selecting the
> >same SVGA mode every time you boot Linux, apply this
> >patch and rebuild the kernel. It allows you to preset
> >the mode in the Makefile, so that the setup program
> >proceeds as if you had hit RETURN and your favorite number.
>
> Could you please also post the same thing for presetting
> the mode to VGA (or EGA)? Maybe it's trivial, but I'm
> not a C programmer and have no idea how this is done.
>
> Thanks!

Sure. This took some doing, and I'd appreciate some feedback.
I moved the # sign from the Makefile to boot/setup.S, so I now
use "2" instead of "\#2" in the Makefile. And now if you set
SVGA_MODE to "VGA" it will compile a kernel that boots EGA/VGA
mode, no questions asked (well, it doesn't wait for you to
answer :-,).

I've assembled all three versions; SVGA_MODE undefined prompts
you for a RETURN key and then a number, SVGA_MODE=VGA acts like
you hit another key and sets EGA/VGA mode, and SVGA_MODE=2 sets
my favorite SVGA mode. They all seem to work right for me.

Dennis

Here's the revised patch:

*** Makefile Mon Apr 6 15:41:45 1992
--- Makefile Mon Apr 13 13:48:02 1992
***************
*** 44,49 ****
--- 44,56 ----
AS86 =as86 -0 -a
LD86 =ld86 -0

+ #
+ # If you want to preset the SVGA mode, uncomment the next line
+ # and set SVGA_MODE to whatever number you want. I use 2.
+ # Set it to -DSVGA_MODE=VGA if you just want the EGA/VGA mode.
+ #
+ #SVGA_MODE= -DSVGA_MODE=2
+
AS =as
LD =ld
#LDFLAGS =-s -x -M
***************
*** 124,130 ****
$(LD86) -s -o boot/setup boot/setup.o

boot/setup.s: boot/setup.S include/linux/config.h
! $(CPP) -traditional boot/setup.S -o boot/setup.s

boot/bootsect.s: boot/bootsect.S include/linux/config.h
$(CPP) -traditional boot/bootsect.S -o boot/bootsect.s
--- 131,137 ----
$(LD86) -s -o boot/setup boot/setup.o

boot/setup.s: boot/setup.S include/linux/config.h
! $(CPP) -traditional $(SVGA_MODE) boot/setup.S -o boot/setup.s

boot/bootsect.s: boot/bootsect.S include/linux/config.h
$(CPP) -traditional boot/bootsect.S -o boot/bootsect.s
*** boot/setup.S Mon Apr 6 13:32:48 1992
--- boot/setup.S Mon Apr 13 12:56:03 1992
***************
*** 245,250 ****
--- 245,251 ----
mov es,ax
lea si,msg1
call prtstr
+ #ifndef SVGA_MODE
flush: in al,#0x60 ! Flush the keyboard buffer
cmp al,#0x82
jb nokey
***************
*** 256,264 ****
--- 257,268 ----
ja nokey
cmp al,#0x9c
je svga
+ #endif
+ #if !defined(SVGA_MODE) || SVGA_MODE == VGA
mov ax,#0x5019
pop ds
ret
+ #endif
svga: cld
lea si,idati ! Check ATI 'clues'
mov di,#0x31
***************
*** 497,502 ****
--- 501,509 ----
call prtstr
pop si
add cl,#0x80
+ #if defined(SVGA_MODE) && SVGA_MODE != VGA
+ nonum: mov al,#SVGA_MODE ! Preset SVGA mode
+ #else
nonum: call getkey
cmp al,#0x82
jb nonum
***************
*** 508,513 ****
--- 515,521 ----
zero: sub al,#0x0a
nozero: sub al,#0x80
dec al
+ #endif
xor ah,ah
add di,ax
inc di

--
Dennis T. Flaherty [email protected]
Miles, Inc. [email protected]
Diagnostics Division
My doctor says Mylanta, but my boss says Alka-Seltzer!


[next article]
From: [email protected] (Alex Liu)

Subject: Re: Needed: a guide to setting up swap space.
Keywords: linux, swap, setup, installation, mkswap, swapon
Message-ID:
Date: 14 Apr 92 05:02:46 GMT
References: <[email protected]> <[email protected]>
Organization: None to Speak of
Lines: 24
NNTP-Posting-Host: aludra.usc.edu

[email protected] (Kevin Cummings) writes:

>> * Is there any way to tell how much swap space you are using or
>> have left?

>Under SUNOS I do a "pstst -T". I haven't seen a LINUX command yet to tell me
>how much is actually in use. And seeing as gcc-1.4 didn't know when my disk
>was full (other that getting unexpected EOF error messages during compiliation)
>I wouldn't be surprised if LINUX doesn't detect "out of swap space" cleanly eit
>For now, I hope that 4MB of memory and 9MB of swap space is enough to not run o
>until we get something that tells us swap usage.

Actually, there is a program to do that... Along with the PS program, there
is a program called "FREE". This reports the ammount of memory and the ammount
of swap pages available. (Among other things). You need to have a recompiled
kernel to be able to use PS, (and FREE) tho.

BTW, I had run out of swap space a couple of times, and Linux did exit cleanly.

--
_____________________________________________________________________________
Alejandro Liu |EMail: [email protected] |All mispellings are intentional
3131 Mc Clintock #373F |Voice: 213-745-2431 |Anything mentioned here is not
Los Angeles, CA 90007 | |necessarily true.


[next article]
From: [email protected] (Alex Liu)

Subject: Re: Patch to preset the SVGA mode
Message-ID:
Date: 14 Apr 92 05:06:45 GMT
References: <[email protected]> <[email protected]> <1992Apr13
[email protected]>
Organization: None to Speak of
Lines: 23
NNTP-Posting-Host: aludra.usc.edu

[email protected] (Dennis Flaherty) writes:

>> >
>> >If you are tired of hitting RETURN and selecting the
>> >same SVGA mode every time you boot Linux, apply this
>> >patch and rebuild the kernel. It allows you to preset
>> >the mode in the Makefile, so that the setup program
>> >proceeds as if you had hit RETURN and your favorite number.

How about asking the user to press RETURN, but if he/she doesnt press
anything in 5 secs (or 10 or whatever) just choose a compiled default?

This might be trivial, but I am not an assembly programmer and I wouldnt
know where to start.

BTW, it would be even nicer to be able to switch modes any time, is this
possible?

--
_____________________________________________________________________________
Alejandro Liu |EMail: [email protected] |All mispellings are intentional
3131 Mc Clintock #373F |Voice: 213-745-2431 |Anything mentioned here is not
Los Angeles, CA 90007 | |necessarily true.


[next article]
From: [email protected] (Mika Matti Jalava)

Subject: gcc2.1 and gnulib?
Message-ID: <[email protected]>
Date: 13 Apr 92 16:26:12 GMT
Sender: [email protected] (Usenet pseudouser id)
Organization: Helsinki University of Technology, Finland
Lines: 7
Nntp-Posting-Host: kaira.hut.fi

What might I be doing wrong now? I've been trying to compile the
kernel and the build utility that comes with it. ld (the one from
binutils.tar.Z and also an old one) complains about not finding
gnulib. That isn't distributed with 2.1misc.tar.Z or 2.1lib.tar.Z.
Where do I find it or is there something else wrong?

Mika


[next article]
From: [email protected] (P Thomson)

Subject: Bad blocks (Help Please !)
Message-ID: <[email protected]>
Date: 13 Apr 92 22:13:08 GMT
Organization: Edinburgh University
Lines: 14

Hi, Can some one tell me if it is possible to hack/mask out bad blocks
after 15+ Mbytes of Linux is installed, without having to take the system
down and start again


Thanks in advance Peter.


P.S also a note on tar and /dev/at0. This might not be worth much to
anyone but it bugged me for a couple of hours today. If you use rawrite
to write a large file to a floppy. It is best to reformat the floppy
afterwards for the next time you are going to use it. I kept getting
tar skipping.. messages in my tar files when I knew there should'nt be any
Formatting the disk cured the problem.


[next article]
From: [email protected] (Jiansheng Zhao)

Subject: compile 95c kernel: fs.h not found?
Message-ID:
Date: 14 Apr 92 05:12:52 GMT
Sender: [email protected] (Usenet News Maintenance)
Organization: University of British Columbia, Vancouver, B.C., Canada
Lines: 4
Nntp-Posting-Host: chilko.ucs.ubc.ca

When compiling kernel 95c+, I got the following error message:
tools/build.c:28 linux/fs.h:No such file or directory
I remember someone has posted solution for this problem, but can't remember
exactly how. BTW, there was a file fs.h.


[next article]
From: [email protected] (P Thomson)

Subject: GCC2.1 (Wonderful !)
Message-ID: <[email protected]>
Date: 14 Apr 92 02:20:15 GMT
Organization: Edinburgh University
Lines: 53


Hi Again, Thought I'd mail and pitch linux again (not that most of
the subscribers here are not already sold =8^) ).


There seem to be some band width about problems with gcc2.1, I would
like to say that after installing it and getting the swap space
sorted out I've had no problems. In fact running it on 2M bytes memory
4M swap space I managed to compile the best part of 1M byte of source
code this evening (in about 30 minutes) the amount of disk traffic
was not nearly as bad as I thought it might be. The source code was
the empire type game dominion ( I've stuck a bit on the bottom about
were you can find it) and it compiled with the minimum of effort
(1 tweek the makeworld.c file )


So thanks to all concerned, Linux is just so more sexier than dos.

Peter.

P.S. thanks to the guys whom mailed me with the answers to my
problems during installation.

----------------------------------------------------------------------
WHAT IS THIS?
-------------

This is release of Dominion (version 1.06), a world simulation and
conquest game. Dominion runs on every version of UNIX on which we
have tested it.

The dominion source code is available by anonymous ftp on
max.physics.sunysb.edu [129.49.21.100], in the directory
pub/dominion-version. For example, to get version 1.06 you would
type:

ftp -i max.physics.sunysb.edu (or ftp -i 129.49.21.100)

cd pub/dominion-1.06
ls
binary
mget DOM*

The files you will retrieve are "shell archives", so after
uncompressing, you un-bundle them with the Bourne shell.

Send bug reports to:
[email protected]

Send flames to /dev/tty :-).

Don't forget that this is copylefted free software. Read the file COPYING.



[next article]
From: [email protected] (Brett J. Vickers)

Subject: Re: Needed: a guide to setting up swap space.
Keywords: linux, swap, setup, installation, mkswap, swapon
Message-ID: <[email protected]>
Date: 14 Apr 92 06:33:15 GMT
References: <[email protected]> <[email protected]>
Reply-To: [email protected] (Brett J. Vickers)
Organization: Univ. of Calif., Irvine, Info. & Computer Sci. Dept.
Lines: 22
Nntp-Posting-Host: bonnie.ics.uci.edu

[email protected] (Kevin Cummings) writes:
>Step 3) After the partition is ready for swapping, I turned it on via
> "swapon /dev/hda4" to see if there would be any problems and ran that
> way for a while.

Hmmm. I created a swap file of 1024 blocks as a test, ran mkswap
on it, and then tried to run swapon on it. What I got was the
following:

Unable to find swap-space signature
Error adding 'swapfile' as swap
swapon: EINVAL

Any ideas on why I can't use it as a swapfile? mkswap seemed
to work just fine on it. It's using swapon that's giving me
a problem.

Brett

--
[email protected] | "Wherever the real power in a Government lies, there
[email protected] | is the danger of oppression." - James Madison


[next article]
From: [email protected] (Bernard Joseph Gardner)

Subject: Re: all kinds of questions
Message-ID: <[email protected]>
Date: 15 Apr 92 06:30:02 GMT
References: <[email protected]> <[email protected]>
Sender: [email protected]
Reply-To: [email protected] (Bernard Joseph Gardner)
Organization: Basser Dept of Computer Science, University of Sydney, Australia
Lines: 65


In article <[email protected]> [email protected] (Bowen Goletz

) writes:
In article <[email protected]> [email protected] (Gre
enup) writes:
>
>Hello netters!
> I just booted up for the first time two days ago, and now I'm a happy
>Linux user. Well, almost. I do have a few questions...

>2. why doesn't "tar +help | more" work?

The help output is through stderr insted of stdio (why?) If you
want to pipe stderr, use the amperstand (&). For example: tar &> tar.help.
Then more the tar.help file. For some reason, stderr pipes aren't supported
with bash. You should be able to do a &| to pipe the output directly to
more in the first place.

Another solution is 'tar +help 2>&1 |more' (This combines stderr with stdout
which is then piped through more)

Bernard.

Newsgroups: alt.os.linux
Subject:
Summary:
Expires:
Sender:
Reply-To: [email protected] (Bernard Joseph Gardner)
Followup-To:
Distribution:
Organization: Basser Dept of Computer Science, University of Sydney, Australia
Keywords:


Newsgroups: comp.os.linux
Subject: Re: all kinds of questions
Summary:
Expires:
References: <[email protected]> <[email protected]>
Sender:
Reply-To: [email protected] (Bernard Joseph Gardner)
Followup-To:
Distribution:
Organization: Basser Dept of Computer Science, University of Sydney, Australia
Keywords:


In article <[email protected]> [email protected] (Bowen Goletz
) writes:
In article <[email protected]> [email protected] (Gre
enup) writes:
>
>Hello netters!
> I just booted up for the first time two days ago, and now I'm a happy
>Linux user. Well, almost. I do have a few questions...

>2. why doesn't "tar +help | more" work?

The help output is through stderr insted of stdio (why?) If you
want to pipe stderr, use the amperstand (&). For example: tar &> tar.help.
Then more the tar.help file. For some reason, stderr pipes aren't supported
with bash. You should be able to do a &| to pipe the output directly to
more in the first place.

Another solution is 'tar +help 2>&1 |more' (This combines stderr with stdout
which is then piped through more)

Bernard.


[next article]
From: [email protected] (Bernard Joseph Gardner)

Subject: Apology and question
Message-ID: <[email protected]>
Date: 15 Apr 92 06:54:19 GMT
Sender: [email protected]
Reply-To: [email protected] (Bernard Gardner)
Followup-To: comp.os.linux
Organization: Basser Dept of Computer Science, University of Sydney, Australia
Lines: 9

Well, first, an apology for the multiply included garbage in that previous post,
I really hate the news software here 🙂

Anyway, my real question is with regard to cut and paste using the mouse. Has
anyone implemented such a beastie?

If not, I might make it one of my pet projects.

Bernard.


[next article]
From: [email protected] (A. V. Le Blanc)

Subject: problems with tcsetattr
Message-ID: <[email protected]>
Date: 13 Apr 92 13:38:16 GMT
Followup-To: [email protected]
Organization: Computing Centre, University of Manchester
Lines: 7

Has anyone else been having problems with tcsetattr?
This may be a kernel bug or not, but several programs on my system,
bash in particular, seem to have problems occasionally with
tcsetattr failing with an error of 'invalid argument'.

-- Owen
[email protected]


[next article]
From: [email protected] (A. V. Le Blanc)

Subject: Re: Another bug?
Message-ID: <[email protected]>
Date: 13 Apr 92 13:43:36 GMT
References: <[email protected]> <1992Apr9.200244.28901@wam
.umd.edu> <[email protected]>
Followup-To: [email protected]
Organization: Computing Centre, University of Manchester
Lines: 13

In article <[email protected]> [email protected].
FI (Linus Benedict Torvalds) writes:
>
>The mis-linked process problem seems to be linked with swapping: I've
>never seen it, and most (all?) that have reported it have only 2M ram -
>it seems to be a race-condition in the exit code, that breaks while
>swapping heavily. I think that running gcc and kermit on a 2M system
>should be enough to swap quite a bit...

I have a 4Mb system and have seen this problem frequently, but
I do use it very heavily.

-- Owen
[email protected]


[next article]
From: [email protected] (Mika Pekka Liljeberg)
Newsgroups: comp.os.linux,alt.os.linux

Subject: Re: Tcsh for linux, fixed release
Message-ID:
Date: 14 Apr 92 08:35:44 GMT
References:
Sender: [email protected] (Uutis Ankka)
Organization: Department of Computer Science, University of Helsinki, Finland
Lines: 14
In-Reply-To: [email protected]'s message of Tue, 14 Apr 1992 01: 02:38
GMT


Tcsh has now been uploaded to tsx-11.mit.edu, too.
Tsx-11 was unreachable last night, so I couldn't
upload it then.

Nic.funet.fi already seems to have tcsh on display.
Fast work! 🙂

Mika

--
Mika Liljeberg Email: [email protected]
Helsinki University [email protected]
Dept. of Computer Science


[next article]
From: [email protected] (Hongjiu Lu -- Graduate Student)

Subject: Re: GCC2.1 (Wonderful !)
Keywords: gcc, 2.1, 1.40
Message-ID: <[email protected]>
Date: 14 Apr 92 06:50:32 GMT
References: <[email protected]>
Sender: [email protected] (USENET News System)
Organization: Washington State University
Lines: 22

In article <[email protected]> [email protected] (P Thomson) writes:
>
>
>There seem to be some band width about problems with gcc2.1, I would
>like to say that after installing it and getting the swap space
>sorted out I've had no problems. In fact running it on 2M bytes memory
>4M swap space I managed to compile the best part of 1M byte of source
>code this evening (in about 30 minutes) the amount of disk traffic
>was not nearly as bad as I thought it might be. The source code was
>
>

It seems that gcc 2.1 can run with 2MB RAM. I think it's the time we
retire gcc 1.40. Lots of problems about kernel and gcc 2.1 are related
to the old gcc 1.40. With gcc 2.1, we can use the shared libs as the
standard. So we can squeeze more stuffs into rootimage. I really miss
bash. I had to copy it from 0.12 rootimage.

What do you guys think?


H.J.


[next article]
From: [email protected] (Niels Skov Olsen)
Newsgroups: comp.os.linux,alt.os.linux

Subject: Re: Keyboard dies when I hit Caps-Lock (+SCSI)
Message-ID: <[email protected]>
Date: 13 Apr 92 22:49:13 GMT
References: <[email protected]> <1992Apr12.074148.6395@colorad
o.edu>
Sender: [email protected]
Organization: Department of Computer Science, U of Copenhagen
Lines: 36

[email protected] (Drew Eckhardt) writes:

[My keyboard locks up when Caps-Lock is hit both in scsi-version
at sc.tamu.edu (DK and US keyb.) and in the scsi0.95c+ version at
headrest.woz.coloradu.edu.]

[scsi devices now use major 8]

>The keyboard :

>Does the keyboard lock up, period, do numlock, CAD still work,
>or not?

Ok I checked it more thoroughly.

Following keys gives TOTAL lock up but with Linux stil running:
Pause
Scroll Lock (gives a task-info)
Caps Lock
Num Lock


When these keys are hit all following keys will make a beep instead
of normal function except Return/Enter which will restore the normal
state:
PgUp
PgDn
Home
End
Ins
Del

The scsi0.95c+ version from headrest is only tested with the supplied
keyboard driver since I cannot yet compile it.

Niels


[next article]
From: [email protected] (Joerg Pommnitz)

Subject: compress doesn't work
Summary: problems using compress
Keywords: compress
Message-ID:
Date: 14 Apr 92 11:11:42 GMT
Sender: [email protected] (Owner of all binaries)
Organization: tu-chemnitz
Lines: 27

I have installed LINUX 0.95a on my harddisk without any problems. In order to
recompile the kernel I got gcc2.1.
Now I have got some trouble. compress "refuses" to uncompress the 2.1misc file.
The file is ok, I checked it on a SUN.
What happens is:
compress starts to work (tar ztvf * shows the contents of the file until it
dies).
Then compress hangs. The process table indicates, that PC doesn't change.
So I think it's somewhere in the kernel.
If I switch to another VC, I can't do any input.
So I can wait a loooong time, but nothing happens. If I interrupt the process
using ^C some messages appear on the screen:

Out of Memory !
Out of Swap Memory!
Memory fault!

But I have 4M RAM and a 6M swapfile !!!!
The same problem occured when I tried to unpack elvis-1.5 (a correct file, too).
My hardware: i386-33, 4M RAM, ST1144A, AMI-BIOS,OAK-VGA.
The LINUX-FS is about 25 M.
There is no special swap device, only a swap file.

I know about at least one other machine with the same problem (using the same
distribution) and another machine where everything is perfect.

It's very frustrating to be unable to do anything, please assist !


[next article]
From: [email protected] (Niels Skov Olsen)
Newsgroups: comp.os.linux,alt.os.linux

Subject: Re: Keyboard dies when I hit Caps-Lock (+SCSI)
Message-ID: <[email protected]>
Date: 14 Apr 92 11:19:58 GMT
References: <[email protected]> <1992Apr12.074148.6395@colorad
o.edu> <[email protected]>
Sender: [email protected]
Organization: Department of Computer Science, U of Copenhagen
Lines: 13

I wrote:

>Following keys gives TOTAL lock up but with Linux stil running:
> Pause
> Scroll Lock (gives a task-info)
> Caps Lock
> Num Lock

Clarification: Everything keep running but it seems as if the
wiring of the keyboard interrupt to the service routine is
being broken after one of these keys are hit.

Niels


[next article]
From: [email protected] (Joel M. Hoffman)
Newsgroups: comp.os.linux,alt.os.linux

Subject: Re: Keyboard dies when I hit Caps-Lock (+SCSI)
Message-ID: <[email protected]>
Date: 14 Apr 92 13:09:32 GMT
References: <[email protected]> <1992Apr12.074148.6395@colorad
o.edu> <[email protected]>
Sender: [email protected] (USENET News system)
Organization: University of Maryland at College Park
Lines: 31
Nntp-Posting-Host: rac2.wam.umd.edu

In article <[email protected]> [email protected] (Niels Skov Ols
en) writes:
>[email protected] (Drew Eckhardt) writes:
>
>[My keyboard locks up when Caps-Lock is hit both in scsi-version
>at sc.tamu.edu (DK and US keyb.) and in the scsi0.95c+ version at
>headrest.woz.coloradu.edu.]
>
>[scsi devices now use major 8]
>
>>The keyboard :
>
>>Does the keyboard lock up, period, do numlock, CAD still work,
>>or not?
>
>Ok I checked it more thoroughly.
>
>Following keys gives TOTAL lock up but with Linux stil running:
> Pause
> Scroll Lock (gives a task-info)
> Caps Lock
> Num Lock

Are you sure the keyboard is locked up? On my system, many of those
keys provide a ``hidden feature.'' They convert everything into
control sequences until Linux is rebooted. So, `a' becomes C-a (which
looks invisible), `c' becomes C-c, etc. BTW, you can still logout, as
`d' become C-d, and if you have a ``sync'' in your history (bash), you
can use `p' (which becomes C-p) to retrieve it.

-Joel
([email protected])


[next article]
From: [email protected] (david.a.cuka)

Subject: Linux Printer
Summary: Does it work?
Keywords: linux, printer
Message-ID: <[email protected]>
Date: 14 Apr 92 14:31:16 GMT
Followup-To: comp.os.linux
Organization: AT&T
Lines: 17


I tried last night to get my parallel printer working with linux. Based
on some previous postings, I tried

"mknod /dev/lp0 c 6 0"

followed by a test

"cat /etc/rc > /dev/lp0"

I got a "ENODEV", so I tried using minor device number 1, which did not
work either. Is there some addtional setup that I am missing? I
am running on v0.95a. The printer is a Panasonic KXP-1024 parallel
printer. Is there a "Guide to installing a printer" yet?

DAC
Obligatory Praise: "Linux is Great!"


[next article]
From: [email protected] (Jon Tombs)
Newsgroups: comp.os.linux,alt.os.linux

Subject: Phantom optimisation
Message-ID: <[email protected]>
Date: 14 Apr 92 14:04:01 GMT
Sender: [email protected]
Organization: Robotics Research Group, Engineering Science Dept, Oxford, UK.
Lines: 19
Originator: [email protected]


Everyone seems to be pushing the optimisation level of gcc higher and higher
the new kernel make file uses -O6. I've not found any gnu doc that mentions
over -O2, and a quick look at the source shows no obvious use of >2.

To test this I built the entire kernel -O6 and -O2.
cmp on the resultant Image found _no_ differences in the binaries!!

Are we victims of GCC hype!!?
--
Jon
"I got in today just as the sun was crashing"









[next article]
From: [email protected] (Adrian Wallaschek)

Subject: Questions and suggestions
Keywords: linux
Message-ID: <[email protected]>
Date: 14 Apr 92 14:18:55 GMT
Sender: [email protected]
Organization: University of Oldenburg, Germany
Lines: 23

Hi everybody!

I've got some questions:

- is there anybody working on a serial driver for multiport cards as
ast-compatible cards, etc. (There is a freeware-driver for commercial
*nix-versions called FAS which does a quite good job!) ?

- under DOS there is a package called FDREAD which makes it possible to
write 1.44 MB on 5.25"-floppies without problems. I think it should be
possible to modify the floppy driver that way, it could handle those
disks. (Perhaps you'd like some extra 240k on your distribution disk)
Is there a nearer description on the floppy-devices, esp. the minor
device-numbers ?

- how soon can we expect a pre^n-release of the new filesystem? (We
are working with three users on a 386/33 and even though we have
fast SCSI (Adaptec) the system actually pauses every user if we
e.g. untar a file of 300k!)

prefect
[email protected]
(I didn't make the address, sorry!)


[next article]
From: [email protected] (Joseph Knapka)

Subject: Hitchhiker's Guide to Linux
Summary: User's Guide volunteers wanted
Message-ID: <[email protected]>
Date: 14 Apr 92 14:27:44 GMT
Organization: University of Georgia, Athens
Lines: 52

Hi, folks. Here's a list of Linux User's Guide chapters, and people
who are writing/compiling them. Volunteerism is HIGHLY encouraged.

0: Linux overview
I: Installation
IIa:gcc-1.40
IIb: gcc-2.x (This will be the standard by the time UG is available.
[email protected] is working on a gcc reference manual.)
III: Communication packages (kermit, etc.)
IV: mtools
V: emacs ...............................Me
VI: groff (& man?)
VII: TeX, LaTeX
------- These next three will involve close cooperation -------
VIII: uucp (Yay!) (Mr. Carp, could you do this one?)
IX: mail
X: ka9q
----------------------------------------------------------
XI: X11
XII: system administration/security ...Chuck Boyer & me
XIII: configuring and building your Linux kernel
XIV: random gnu software (general porting info)

The first chapter (0, that is) should cover basically what the
INFO-SHEET does.

Chapter I should not only cover the physical installation of Linux,
but should also discuss the standard *nix utilites, what they're for,
and where to get them --- i.e., it should be a guide to putting
together a functional system. [email protected] says:
> With the ABC release, this chapter should only be a minor
> revision of what Iain has already done, in my opinion.
Yes. But should definitely include a pointer to a good book
on Unix for regular-guy users, IMHO. Also, I think the filesystem
standard document should be here.

The chapters devoted to software packages should cover 1) Where to get
the package; 2) Best-case installation procedure; 3) Typical
installation problems and fixes; 4) When available, a pointer to a
genuine user's manual (i.e., "Lamport, Leslie. LaTeX: a document
preparation system. Addison-Wesley, Reading, MA, 1986").

The chapter on GNU software should cover things like appropriate
compiler flags for Linux and so forth. And why they're appropriate.

It's been suggested that "texinfo" be the standard vehicle for this
project. Is this a good idea? I will have to find out about texinfo.

Comments or (preferably) volunteers?

Bye,
Joseph


[next article]
From: [email protected] (Lawrence C Foard)

Subject: I'm back (I think)
Message-ID: <[email protected]>
Date: 14 Apr 92 01:08:19 GMT
Sender: [email protected] (News)
Organization: Worcester Polytechnic Institute
Lines: 11
Nntp-Posting-Host: goofy.wpi.edu

Between various disasters I've been unable to keep up with linux stuff for a
while.

Some jerk mail bombed wintermute using a forged sendsys, this meant that I
didn't get any of the mail that was supposed to get there since it had to
be taken off the net.

Then I lost the use of my account there because of it. But fortunitly WPI
is getting comp.os.linux now, so until WPI's news feed dies again I'm back 🙂

If any ones mail to me boinged send it again to [email protected]


[next article]
From: [email protected] (Kevin Cummings)
Newsgroups: comp.os.linux,alt.os.linux

Subject: Re: all kinds of questions
Message-ID: <[email protected]>
Date: 14 Apr 92 15:21:17 GMT
References: <[email protected]> <1992Apr13.223428.29499@colo
rado.edu>
Sender: cummings@hammer (Kevin Cummings)
Organization: Prime Computer R&D
Lines: 29

In article <[email protected]>, [email protected] (
Drew Eckhardt) writes:
> In article <[email protected]> [email protected] (Gr
eenup) writes:
> >3. does "sync" do a diskpark, or do I have to round up a routine for that?
>
> No. sync merely commits the buffercache to disk. Buffers will be
> committed to disk when buffercache is full, or a sync() syscall
> is performed. To prevent heinous catastrophees, update is
> run from /etc/rc. update syncs the disks every 30 seconds.
>
> Most hard disks these days autopark, so that shouldn't be a problem.

Unless you are not running with "new" disks. There are an AWFUL lot of MFM
disk out there that don't auto-park; like most of them! So unless you are
shutting down your system to reboot DOS (or something else), your disk
heads will not be parked by LINUX. Maybe this is the cause of some of the
trashed file systems some people are experiencing?

=================================================================
Kevin J. Cummings Prime Computer Inc.
20 Briarwood Road 500 Old Connecticut Path
Framingham, Mass. Framingham, Mass.

Work: [email protected]
Home: [email protected]

Std. Disclaimer: "Mr. McKittrick, after careful consideration,
I've come to the conclusion that your new
defense system SUCKS..." -- War Games
=================================================================


[next article]
From: [email protected] (Kevin Cummings)

Subject: Re: Linux Printer
Keywords: linux, printer
Message-ID: <[email protected]>
Date: 14 Apr 92 15:47:44 GMT
References: <[email protected]>
Sender: cummings@hammer (Kevin Cummings)
Organization: Prime Computer R&D
Lines: 41

In article <[email protected]>, [email protected]
(david.a.cuka) writes:
>
> I tried last night to get my parallel printer working with linux. Based
> on some previous postings, I tried
>
> "mknod /dev/lp0 c 6 0"
>
> followed by a test
>
> "cat /etc/rc > /dev/lp0"
>
> I got a "ENODEV", so I tried using minor device number 1, which did not
> work either. Is there some addtional setup that I am missing? I
> am running on v0.95a.

I am running 0.95c+ from tsx-11. When I boot, it tells me that lp1 and lp2
both exist in my system (note: no lp0). This is amusing since under
dos, my printer (when it was working) was connected to LPT1. I suspect
that the line printers here suffer from the same problem as the serial
ports. Under DOS (actually, under POST), the BIOS checks a list of IO
addresses sequentially and assigns the first addresses that is successful
to xxx1, the second to xxx2, etc trying upt to 2 or 4 COM ports, and up to
3 LPT ports. Is it true that unde LINUX there is a one to one correspondence
between an address and a LINUX device? If so, then the parallel port
that is found on the old CGA/Monochrome cards is lp0, and the parallel ports
more commonly found on serial/parallel cards can be either lp1 or lp2.
I don't know what addresses are used by printer ports off the top of
my head, and I don't know which one is used on printer ports on motherboards.

=================================================================
Kevin J. Cummings Prime Computer Inc.
20 Briarwood Road 500 Old Connecticut Path
Framingham, Mass. Framingham, Mass.

Work: [email protected]
Home: [email protected]

Std. Disclaimer: "Mr. McKittrick, after careful consideration,
I've come to the conclusion that your new
defense system SUCKS..." -- War Games
=================================================================


[next article]
From: [email protected] (Scott Dunn)
Newsgroups: comp.os.linux,alt.os.linux

Subject: Miscellaneous questions.
Message-ID: <[email protected]>
Date: 14 Apr 92 12:00:35 GMT
Sender: [email protected] (News Administrator)
Organization: Dept of Computer Science, Heriot-Watt University, Scotland
Lines: 49

Finally managed to get linux installed and working properly. 0.95c seems
to have solved the HD timeout problem I was having. Thanks to everyone
who tried to help. I think I replied to everyone, but if I didn't, sorry,
I deleted all the saved messages before replying to all.(silly)

Some miscellaneous questions (probably daft) :


1. I am using tcsh, how do I set up paths to include and lib directories?
This doesn't seem to be a problem, recompiled the kernel with no paths
set up, other than path for binaries etc.
However, gcc -g (debug info) gives the following link error :
can't find libg or something like that.
I can't find what it's looking for.

2. What is a working version of tar and un/compress?
A lot of problems seem to be solved by getting new versions of the
aforementioned. However, nobody has indicated how to identify which
are the dodgy copies.
Sometimes when uncompressing it seems to get stuck in an infinite
loop which takes all system resources. ie I can't login to another
terminal and kill the process. Is this what is known as a 'race
condition' ? Is this indicate a dodgy copy of un/compress ?

3. What are shared libraries?
What are the implications of using shared libraries?
I got a load of object files, I think it was 2.1-shared-A.tar.Z which
apparently use them. Certainly, once they were 'made' they were
much smaller than the 'normal' equivelants.
Are there limitations when using shared libraries?
Are they like runtime libaries?

4. Why are ccp and cc1plus(?) in my /lib directory with links to
/usr/lib/..../ccp, cc1plus?
When I first tried to use gcc, it failed saying it couldn't
find them. I just stuck /lib in my path, but this seems wrong
to me. I didn't move the aforementioned into /usr/bin on the
assumption that they must have been put /lib for some reason.

5. A number of utilities crash when I give valid syntax in an
inapropriate situation. e.g.
ls -l *.doc
when there are no docs in the current directory. Is this a bug
in the kernel, the utilities (do they have to be recompiled), or
some peculiarity of my system?


Thanks in advance,

Scott.


[next article]
From: [email protected] (Thomas Wuensche)

Subject: German FTP-Server ready/Mirroring tsx-11
Message-ID:
Date: 14 Apr 92 09:49:51 GMT
Sender: [email protected] (Mr. News)
Reply-To: [email protected]
Organization: Leibniz-Rechenzentrum, Muenchen (Germany)
Lines: 10

I have set up ftp-server ftp.fgb.mw.tu-muenchen.de [129.187.200.1].
It mirrors /pub/linux on tsx-11.mit.edu (Thanks to Ted for his help).

There will soon be installed the ftpd from wuarchive.wustl.edu,
that allows downloading subdirectories by creating a tar-file on
the fly.

Enjoy it,

Thomas


[next article]
From: [email protected] (Steve Ansell)

Subject: Re: Linux Printer
Keywords: linux, printer
Message-ID: <[email protected]>
Date: 14 Apr 92 17:20:28 GMT
References: <[email protected]>
Sender: [email protected]
Organization: Computing Services, UC Davis
Lines: 27

In article <[email protected]> [email protected]
(david.a.cuka) writes:
>
>I tried last night to get my parallel printer working with linux. Based
>on some previous postings, I tried
>
> "mknod /dev/lp0 c 6 0"
>
>followed by a test
>
> "cat /etc/rc > /dev/lp0"
>
>I got a "ENODEV", so I tried using minor device number 1, which did not
>work either. Is there some addtional setup that I am missing? I
>am running on v0.95a. The printer is a Panasonic KXP-1024 parallel
^^^^^^

Unless you applied the kernel patches yourself (you don't mention it above)
you don't have the lp driver support in the kernel. You need to get a copy
of v0.95c+ which has the lp driver support in the kernel already. Once you
have that there shouldn't be much to it other than what you show above.


--
-Steven T. Ansell
Unix Consultant
Computing Services U.C.D.



[next article]
From: [email protected] (Al Clark)

Subject: Re: When will "Linux 1.0" come to the world ?
Keywords: stable
Message-ID:
Date: 14 Apr 92 18:07:32 GMT
References: <1992Apr13.215610.22632%[email protected]>
Organization: Netcom - Online Communication Services (408 241-9760 guest)
Lines: 14

In article <1992Apr13.215610.22632%[email protected]> [email protected]
.tw (mr79) writes:
>Linux 0.95c is still in the period of development.
>Everything is rather unstable.
>when will Linux 1.0 come to the PC 386/486 world ?
>
> J.H.Lin

The only thing unstable is the configuration management [ 🙂 ]. Linux
is actually more stable than the early (1.06) Interactive Systems Unix
386/ix that I have. The only thing I miss is the ability to run DOS
programs. But I find I would rather have Linux and do some rebooting.
--
Al - [email protected] - My opinions are my own.
*** Practice random kindness and senseless acts of beauty! ***


[next article]
From: [email protected] (Greg Lee)

Subject: Re: Questions and suggestions
Keywords: linux
Message-ID: <[email protected]>
Date: 14 Apr 92 18:31:28 GMT
References: <[email protected]>
Sender: [email protected] (News Service)
Organization: University of Hawaii
Lines: 82
Nntp-Posting-Host: uhunix.uhcc.hawaii.edu

In article <[email protected]> Adrian.Walla
[email protected] (Adrian Wallaschek) writes:
}...
} - under DOS there is a package called FDREAD which makes it possible to
} write 1.44 MB on 5.25"-floppies without problems. I think it should be
} possible to modify the floppy driver that way, it could handle those
} disks. (Perhaps you'd like some extra 240k on your distribution disk)
} Is there a nearer description on the floppy-devices, esp. the minor
} device-numbers ?

I've been using 3.5" disks formatted by fdform18 (whose msdos tsr is fdread.exe)
for a month or so with no apparent problem. It gets me 1722 1k blocks per diske
tte.
(In msdos: fdformat a: f172, or something similar.) My drive A: is 3.5", but you
might
be able to adapt my changes for extra room on 5.25" disks.

(1) I've patched floppy.c to replace one of the floppy types I don't use with a
type
for the 1722 block 3.5" diskettes (context diff below is against 95c version
).
I also disabled track reads for diskettes with the extra sectors/track (21).
(2) There's a corresponding special file in /dev:
brw-rw-rw 1 root root 2, 12 Mar 13 17:54 /dev/f172
(3) For mtools (the version that reads disk types from /etc/mtools), there's a l
ine in
/etc/mtools:
H /dev/f172 12 82 2 21
to use, e.g., "mcd h:/" with the overstuffed diskettes.

*** linux/kernel/blk_drv/floppy.c Fri Mar 13 17:03:31 1992
--- /sys/linux/kernel/blk_drv/floppy.c Sun Apr 5 02:30:04 1992
***************
*** 92,98 ****
{ 0, 0,0, 0,0,0x00,0x00,0x00 }, /* no testing */
{ 720, 9,2,40,0,0x2A,0x02,0xDF }, /* 360kB PC diskettes */
{ 2400,15,2,80,0,0x1B,0x00,0xDF }, /* 1.2 MB AT-diskettes */
! { 720, 9,2,40,1,0x2A,0x02,0xDF }, /* 360kB in 720kB drive */
{ 1440, 9,2,80,0,0x2A,0x02,0xDF }, /* 3.5" 720kB diskette */
{ 720, 9,2,40,1,0x23,0x01,0xDF }, /* 360kB in 1.2MB drive */
{ 1440, 9,2,80,0,0x23,0x01,0xDF }, /* 720kB in 1.2MB drive */
--- 92,99 ----
{ 0, 0,0, 0,0,0x00,0x00,0x00 }, /* no testing */
{ 720, 9,2,40,0,0x2A,0x02,0xDF }, /* 360kB PC diskettes */
{ 2400,15,2,80,0,0x1B,0x00,0xDF }, /* 1.2 MB AT-diskettes */
! /*--gl { 720, 9,2,40,1,0x2A,0x02,0xDF },*/ /* 360kB in 720kB drive
*/
! { 3444,21,2,82,0,0x0C,0x00,0xCF }, /* 1.722MB diskette */
{ 1440, 9,2,80,0,0x2A,0x02,0xDF }, /* 3.5" 720kB diskette */
{ 720, 9,2,40,1,0x23,0x01,0xDF }, /* 360kB in 1.2MB drive */
{ 1440, 9,2,80,0,0x23,0x01,0xDF }, /* 720kB in 1.2MB drive */
***************
*** 374,380 ****
--- 375,387 ----
*/
static void transfer(void)
{
+ /**--gl
read_track = (command == FD_READ) && (CURRENT->errors < 4);
+ **/
+ if (floppy->sect <= 18) read_track = (command == FD_READ) && (CURRENT->e
rrors < 4);
+ else read_track = 0;
+ /****/
+
if (cur_spec1 != floppy->spec1) {
cur_spec1 = floppy->spec1;
output_byte(FD_SPECIFY);
***************
*** 543,549 ****
0, 0, 0, 0,
360, 360 ,360, 360,
1200,1200,1200,1200,
! 360, 360, 360, 360,
720, 720, 720, 720,
360, 360, 360, 360,
720, 720, 720, 720,
--- 550,557 ----
0, 0, 0, 0,
360, 360 ,360, 360,
1200,1200,1200,1200,
! /*--gl 360, 360, 360, 360, */
! 1722,1722,1722,1722,
720, 720, 720, 720,
360, 360, 360, 360,
720, 720, 720, 720,


--
Greg Lee


[next article]
From: [email protected] (David Darknell)

Subject: X-Windows status?
Message-ID:
Date: 14 Apr 92 19:14:41 GMT
Sender: [email protected]
Lines: 5


Whats the current status of X? Is there something available
to play with?

David


[next article]
From: [email protected] (Hongjiu Lu)
Newsgroups: comp.os.linux,alt.os.linux

Subject: Re: Miscellaneous questions.
Message-ID: <[email protected]>
Date: 14 Apr 92 16:51:02 GMT
References: <[email protected]>
Sender: [email protected] (USENET News System)
Organization: Washington State University
Lines: 72

In article <[email protected]>, [email protected] (Scott Dunn)
writes:
|> Finally managed to get linux installed and working properly. 0.95c seems
|> to have solved the HD timeout problem I was having. Thanks to everyone
|> who tried to help. I think I replied to everyone, but if I didn't, sorry,
|> I deleted all the saved messages before replying to all.(silly)
|>
|> Some miscellaneous questions (probably daft) :
|>
|> 1. I am using tcsh, how do I set up paths to include and lib directories?
|> This doesn't seem to be a problem, recompiled the kernel with no paths
|> set up, other than path for binaries etc.
|> However, gcc -g (debug info) gives the following link error :
|> can't find libg or something like that.
|> I can't find what it's looking for.

ranlib /usr/lib/libg.a

|>
|> 2. What is a working version of tar and un/compress?
|> A lot of problems seem to be solved by getting new versions of the
|> aforementioned. However, nobody has indicated how to identify which
|> are the dodgy copies.
|> Sometimes when uncompressing it seems to get stuck in an infinite
|> loop which takes all system resources. ie I can't login to another
|> terminal and kill the process. Is this what is known as a 'race
|> condition' ? Is this indicate a dodgy copy of un/compress ?
|>

I think those on 0.95a rootimage are no good.

|> 3. What are shared libraries?
|> What are the implications of using shared libraries?
|> I got a load of object files, I think it was 2.1-shared-A.tar.Z which
|> apparently use them. Certainly, once they were 'made' they were
|> much smaller than the 'normal' equivelants.
|> Are there limitations when using shared libraries?
|> Are they like runtime libaries?

Do not distribute the shared binaries. Distribute .a or static only.

|>
|> 4. Why are ccp and cc1plus(?) in my /lib directory with links to
|> /usr/lib/..../ccp, cc1plus?

How you got cc1plus? Cpp is for some programs which think it should be /lib or
/usr/lib.

|> When I first tried to use gcc, it failed saying it couldn't

Installation error. Read FAQ for gcc 2.1.


|> find them. I just stuck /lib in my path, but this seems wrong
|> to me. I didn't move the aforementioned into /usr/bin on the
|> assumption that they must have been put /lib for some reason.
|>
|> 5. A number of utilities crash when I give valid syntax in an
|> inapropriate situation. e.g.
|> ls -l *.doc
|> when there are no docs in the current directory. Is this a bug
|> in the kernel, the utilities (do they have to be recompiled), or
|> some peculiarity of my system?
|>

I don't know. Maybe VFS.

|>
|> Thanks in advance,
|>
|> Scott.

H.J.


[next article]
From: [email protected] (Karl R. Buck)

Subject: Re: GCC2.1 (Wonderful !)
Keywords: gcc, 2.1, 1.40
Message-ID:
Date: 14 Apr 92 19:59:57 GMT
References: <[email protected]> <[email protected]>
Organization: Kansas State University
Lines: 21
NNTP-Posting-Host: matt.ksu.ksu.edu

[email protected] (Hongjiu Lu -- Graduate Student) writes:

>It seems that gcc 2.1 can run with 2MB RAM. I think it's the time we
>retire gcc 1.40. Lots of problems about kernel and gcc 2.1 are related
>to the old gcc 1.40. With gcc 2.1, we can use the shared libs as the
>standard. So we can squeeze more stuffs into rootimage. I really miss
>bash. I had to copy it from 0.12 rootimage.

>What do you guys think?

Well, since you ask:-)
Some of you may remember my whining and complaining a while back about
not being able to get 2.1 to work. It ended up being a bogus copy of
tar from the 95a distribution. After fixing this I've no trouble at
all compiling .95c+ and other applications. If anyone is looking for
votes I think going with 2.1 is a great idea. I'm really looking
forward to converting my stuff to shared libs.

To the people (person) working on 2.1: Thanks, and good work!
--
Karl Buck, [email protected] 🙂


[next article]
From: [email protected] (william.j.bitner)
Newsgroups: comp.os.linux,alt.os.linux

Subject: Can't get Linux to boot
Keywords: help
Message-ID: <[email protected]>
Date: 14 Apr 92 19:50:55 GMT
Followup-To: poster
Organization: AT&T
Lines: 40

I ftp'd the boot/root/rawrite.c/FAQ/REL images down (remembering to
use binary mode with ftp), uncompressed them and used kermit
to bring them down to my DOS machine. FTP site was tsx-11.mit.edu
and version of Linux was 0.95a. Uncompressed, the bootimage
was 190980 bytes. I ran sum on the compressed images and the
first number agreed but the block count was different (different
size blocks on my machine). I tried both kermit and umodem to
get the uncompressed files down to my machine with no differences.
I formatted a 5.25 floppy and verified it really had 1.2M of
space available. I performed the rawrite using my bootable
drive "A" (the 5.25) and attempted to boot. The normal configuration
table comes up (from BIOS) and then Floppy drive spins. I get
two "graphics" chars. a newline and then Lo. Nothing further and
the keyboard is locked (IE type 16 keystrokes and it beeps). A
soft reset (Ctrl-Alt-DEL) will get me back to DOS if I remove the
floppy. I tried altering the machines configuration by disabling
cache, video BIOS copy , and BIOS copy (into ram). This made
no difference at all. I ftp'd the bootimage only (if the boot
works, there'll be time to get the rest) from banjo.concert.net
version 0.95c+ (new could be better, right??) and got the same
results except it printed Loa instead of "Lo ". What to do???

Machine. 386/25 with 64K cache OPTI chipset with AMI BIOS
4 MB. Ram
387 FPU
Floppy A 5.25 1.2M
Floppy B 3.5 1.44M
Hard Drive C: IDE Western Digital 202 MB (Master)
Hard Drive D: IDE Seagate ST-157a-0 44 MB (Slave)
Trident SVGA (1 MB)
4 Serial Ports
2 Parallel Ports
SoundBlaster Board


Help is needed!
Thanks,
Bill Bitner
[email protected]
708-879-0401


[next article]
From: [email protected] (LCDR Michael E. Dobson)

Subject: Re: tar files on floppy.
Message-ID: <[email protected]>
Date: 14 Apr 92 20:00:59 GMT
References: <[email protected]> <1992Apr13.132029.8256@colorado.
edu>
Organization: Naval Medical Research & Development Command
Lines: 34

In article [email protected] (Juergen Hen
ke) writes:
>In article <[email protected]> [email protected] (Dr
ew Eckhardt) writes:
>
> When you rawrite to a floppy, you write to the raw device. The file size,
> etc are lost, and are not encoded in a compressed file as they
> are with a .tar file. Solutions to this problem are to :
>
>[...]
>
>I've done this now several times, with GNU-tar (the normal linux tar) it's no
>problem to read a rawritten (:-)) compressed tar from a floppy, just use
>(for example)
>
> tar xvZf /dev/PS0
> ^
> |
> compress/uncompress option
>
I also just successfully installed a complete 0.95c+/gcc2.1 system via this
method but I used the command line:

tar xvpz
I am just about finished with linking everything in 2.1shared-A.tar.Z with
shared libs and symlinking into /usr/bin. A word of caution, DON'T use
shared libs or symlinks on anything in /bin or /etc if / and /usr are on
different partitions. They won't work until /usr is mounted. Keep a good
bootimage floppy and root floppy handy just in case.

--
LCDR Mike Dobson, Sys Admin for | Internet: [email protected]
nmrdc1.nmrdc.nnmc.navy.mil | UUCP: ...uunet!mimsy!nmrdc1!rdc30
AT&T 3B2/600G Sys V R 3.2.3 | BITNET: dobson@usuhsb or nrd0mxd@vmnmdsc
WIN/TCP for 3B2 R3.2 | MCI-Mail: 377-2719 or [email protected]


[next article]
From: [email protected] (Drew Eckhardt)

Subject: Re: Questions and suggestions
Keywords: linux
Message-ID: <[email protected]>
Date: 14 Apr 92 21:13:14 GMT
References: <[email protected]>
Sender: [email protected] (The Daily Planet)
Organization: University of Colorado at Boulder
Lines: 88
Nntp-Posting-Host: bigwig.cs.colorado.edu

In article <[email protected]> Adrian.Walla
[email protected] (Adrian Wallaschek) writes:
>Hi everybody!
>
>I've got some questions:
>
> - under DOS there is a package called FDREAD which makes it possible to
> write 1.44 MB on 5.25"-floppies without problems. I think it should be
> possible to modify the floppy driver that way, it could handle those
> disks. (Perhaps you'd like some extra 240k on your distribution disk)
> Is there a nearer description on the floppy-devices, esp. the minor
> device-numbers ?

It definately is possible, it's just a question of finding someone to do it.

> - how soon can we expect a pre^n-release of the new filesystem? (We
> are working with three users on a 386/33 and even though we have
> fast SCSI (Adaptec) the system actually pauses every user if we
> e.g. untar a file of 300k!)
>

That's not a file system problem, it's a SCSI driver thing

Only the kernel is scheduled when Linux is in the strategy routine
of a device driver.

With the SCSI drivers, there are two possible interfaces between
the midlevel driver and the lowlevel host specific driver. One
provides a generic interupt driven interface, and expects that a
completion function will be called when the SCSI host is done doing
its thing, and that the command function will return immediately,
allowing the strategy routine in sd.c to exit immediately,
and other tasks to be scheduled. With this interface,
you won't have that problem.

The other interface is mostly there to ease development,
and to support brain dead hardware that
can't cope with interrupts well. It
calls a command function that doesn't exit until
command completion.

Since the command() is called
from scsi_do_cmnd(), which is called from the strategy routine
of the sd driver, when this interface to the SCSI drivers is used,
the strategy routine doesn't exit until the command completes,
and user tasks aren't scheduled until exit.

On the Adaptec drivers, code is there supporting the queue_command()
interface, but there are bugs in the {compiler, kernel, adaptec low level driver
, midlevel SCSI driver, gcc} that keep it from working correctly.

The Seagate adapters have severe hardware restrictions that
led to me to use the command() interface.

I think
the current Ultrastor driver used this approach for ease
of development, and in anticipation of the new interface
designed to work with multiple outstanding commands,
out there because of multiple devices and disconnect / reconnect,
the support for multiple outstanding commands for a single LUN
under SCSI2, and inteligent host adapters that queue commands
internally and send them out in an optimal order.

So - no one has done much work with the queue_command() interface,
and since there are no other implementations other than
the one in the Adaptec driver, I can't tell if the bug
is there, in my code, the reset of the kernel, etc.

This situation will be rectified permanantly when
1. Some one else takes care of it.

2. Some one donates / loans me a better SCSI host more friendly
to this interface, like an Adaptec, with
full technical documentation.

3. I get time to implement a hack to deal with the seagate
hardware in this mode, which means either
disconnecting and reconnecting on an interrupt
(which can only be triggered by a device on
the BUS asserting SEL),
and personally starting to use queue_command(),
or doing something more bletcherous.


A temporary fix may be to call schedule() from
within the wait loop in the low level driver, but
I haven't looked at the side effects this will have.

Linus?


[next article]
From: [email protected] (David Engel)

Subject: Re: tar files on floppy.
Message-ID: <[email protected]>
Date: 14 Apr 92 22:13:09 GMT
References: <[email protected]>
Organization: Optical Data Systems, Inc.
Lines: 17

[email protected] (LCDR Michael E. Dobson) writes:
: I am just about finished with linking everything in 2.1shared-A.tar.Z with
: shared libs and symlinking into /usr/bin. A word of caution, DON'T use
: shared libs or symlinks on anything in /bin or /etc if / and /usr are on
: different partitions. They won't work until /usr is mounted. Keep a good
: bootimage floppy and root floppy handy just in case.

If you copy the shared lib to /lib instead of symlinking it (done by the
install script) you can use shared binaries in /bin and /etc. You can also
safely remove /usr/shared the next time you upgrade libraries. It's still
a good idea to have "emergency" boot and root floppies handy however.

David
--
David Engel Optical Data Systems, Inc.
[email protected] 1101 E. Arapaho Road
(214) 234-6400 Richardson, TX 75081


[next article]
From: [email protected] (Jim Winstead Jr.)

Subject: Re: tar files on floppy.
Message-ID: <[email protected]>
Date: 14 Apr 92 23:04:18 GMT
References: <[email protected]> FhG.de> <[email protected]>
Sender: [email protected] (The News System)
Organization: Harvey Mudd College, Claremont, CA 91711
Lines: 20

In article <[email protected]> [email protected]
c.nnmc.navy.mil (LCDR Michael E. Dobson) writes:
>I am just about finished with linking everything in 2.1shared-A.tar.Z with
>shared libs and symlinking into /usr/bin. A word of caution, DON'T use
>shared libs or symlinks on anything in /bin or /etc if / and /usr are on
>different partitions. They won't work until /usr is mounted. Keep a good
>bootimage floppy and root floppy handy just in case.

As long as /lib is on your root partition, and /lib/lib92.04.* is
actually present in that directory, and *not* a sym-link, you can use
shared binaries in /bin just fine. /lib/lib92.04.* really shouldn't
be symlinked....

The idea of keeping a good boot disk and root disk is valid, though.
I've had to go back to mine on a few occasions. (Usually the boot
disk, when I screw up a kernel compilation.)
--
Jim Winstead Jr. (CSci '95) | "Catch a fish!"
Harvey Mudd College | -Geddy Lee,
[email protected] | San Diego Sports Arena
Disclaimer: Mine, not theirs! | January 20, 1992


[next article]
From: [email protected] (Jiansheng Zhao)

Subject: can't compile selection/mouse driver
Message-ID:
Date: 14 Apr 92 22:42:40 GMT
Sender: [email protected] (Usenet News Maintenance)
Organization: University of British Columbia, Vancouver, B.C., Canada
Lines: 9
Nntp-Posting-Host: chilko.ucs.ubc.ca

I almost got everything work with linux0.95c+.
However, I still have trouble to compile selection to use a mouse.
gcc2.1 compiled new kernel all right. When compiling selection-1.0
I got the following error message:
stuff.c In function 'begsel'
stuff.c:73: 'TIOCLINUX' undeclared
stuff.c:86 ... ... ...
Maybe there is a simple fix for trained eyes, but I have been pulling
my hair ,:).


[next article]
From: [email protected] (Steven Kraft)

Subject: Samsung Sync 4 modes - reply for
Message-ID: <[email protected]>
Date: 15 Apr 92 00:20:15 GMT
Reply-To: [email protected]
Lines: 25
Nntp-Posting-Host: hotspur.gac.edu

Sorry that I have to post this here in comp.os.linux, since it is not really
relevant. But I need to answer mail from a fellow Linux enthusiast who doesn't

know how to get a proper address on his mail. 🙂

----- Transcript of session follows -----
421 Host maxwell not found for mailer ether.
550 David Darknell ... Host unknown

!Are you able to use your SyncMaster 4 at 1280x1024 without problems?
!
!David
First let me say that the Samsung Syncmaster 4 is only advertised as a
1024*768 monitor.
I don't use 1280*1024 for anything right now, but the test program
supplied with the Orchid Fahrenheit 1280/D uses that mode successfully on my
Syncmaster 4. It has flicker since it is 48Hz interlaced. I have to fiddle
with the size and position knobs some to center the image. It is a bit coarse
since the Syncmaster 4 is .31mm dot pitch.
If you are looking for advice on finding a GOOD monitor for something
serious like CAD at 1280*1024 I would suggest that you look elsewhere or you
may be disappointed.

Steven Kraft [email protected]
Micro/Minicomputer Consultant, Comp Sci major
Gustavus Adolphus College


[next article]
From: [email protected] (Jiansheng Zhao)

Subject: bootlin.com (boot linux from HD) works well!
Keywords: linux, harddisk, boot
Message-ID:
Date: 14 Apr 92 23:48:19 GMT
Sender: [email protected] (Usenet News Maintenance)
Organization: University of British Columbia, Vancouver, B.C., Canada
Lines: 7
Nntp-Posting-Host: chilko.ucs.ubc.ca

Thanks to [email protected], now I can boot linux from my
hard drive (dev/hdb2) directly. I used mutiple config.sys files and left
autoexec.bat empty when booting linux. There is only one line needed in
config.sys file:
SHELL=c:\bootlin.com c:\etc\boot95c (boot image file).

I used to have boot-image floppy handy when starting linux, not any more.


[next article]
From: [email protected] (A. V. Le Blanc)

Subject: Re: tar files on floppy.
Message-ID: <[email protected]>
Date: 14 Apr 92 08:03:55 GMT
References: <[email protected]>
Followup-To: [email protected]
Organization: Computing Centre, University of Manchester
Lines: 37

In article <[email protected]> [email protected] (Lars Lind
ner) writes:
>
>When I have both boot and root images installed on HD and i want to
>install more SW downloaded as .tar.Z files, in what way vill i get
>them on a floppy to be able to install them on my Linux machine.

If you have both Linux and MS-DOS on the same machine, the simplest
course of action is to use mtools (i.e., mread, mcopy, or mtype) to
access them.

If you do not have both Linux and MS-DOS on the same machine, the
simplest course of action is to copy the files to a dos disk, then
use mtools to read the floppy.

>* Shall i use "rawrite" from DOS to put the files on floppy?

The 'rawrite' program is intended for use with disk images. It CAN
write a tar file to a disk, and in fact this is how I originally
installed Linux on my machine, but it is probably simpler to use
copy and mtools.

>* I rememer reading something like "Do not rawrite compressed files.".
> Why??

The problem is, I think, that compress is written to read from a
file or a stream, and that its I/O does not handle a raw read from a
disk correctly. Anyway it didn't when I first installed Linux.

For larger files, it would be best to split them up, copy them over
on separate floppies, and concatenate them, that is, assuming that
you have no DOS partition on your Linux machine. Unfortunately I
cannot point you to any standard utility running on DOS which would
do this painlessly. I did this once with a GNUish DOS dd, which I
found in the anonymous ftp at, I believe, grape.ecs.clarkson.edu.

-- Owen
[email protected]


[next article]
From: [email protected] (A. V. Le Blanc)

Subject: GNU Fortran
Message-ID: <[email protected]>
Date: 14 Apr 92 11:17:24 GMT
Followup-To: [email protected]
Organization: Computing Centre, University of Manchester
Lines: 8

James Craig Burley, who is writing the GNU Fortran compiler,
announced in a recent post that he is getting a machine to
run Linux on. I expect that those who like FORTRAN (yeugh!)
will find the Linux version should work fairly well if JCB
is doing the development on a Linux system.

-- Owen
[email protected]


[next article]
From: [email protected] (A. V. Le Blanc)

Subject: Dr DOS and fdisk 0.91
Message-ID: <[email protected]>
Date: 14 Apr 92 12:27:58 GMT
Followup-To: [email protected]
Organization: Computing Centre, University of Manchester
Lines: 17

Someone reported recently that the Linux/Minix partitions
created by my fdisk 0.91 are recognised as DOS partitions
by Dr. DOS 6.0. I suspect what is happening is that
Dr. DOS is ignoring the high bit of the type field. If
this is so, it should certainly be reported to the Dr. DOS
suppliers, since it will cause problems for MINIX users
as well. The 8-bit field contains 00000001 for a DOS partition
with a 12-bit FAT, and 10000001 for MINIX/Linux. You can
always change to another number, e.g., 0x41, which I expect
even Dr. DOS will let alone.

A bug in the version of fdisk 0.91 makes it delete partition
5 incorrectly. This bug is fixed in fdisk 0.92, which I shall
be distributing soon.

-- Owen
[email protected]


[next article]
From: [email protected] (A. V. Le Blanc)

Subject: Re: Needed: a guide to setting up swap space.
Message-ID: <[email protected]>
Date: 14 Apr 92 13:16:40 GMT
References: <[email protected]> <[email protected]>
Followup-To: [email protected]
Organization: Computing Centre, University of Manchester
Lines: 21

In article <[email protected]> [email protected] (Kevin
Cummings) writes:
>In article <[email protected]>, [email protected] (Clifford A Adams) write
s:
>> * Is there any way to tell how much swap space you are using or
>> have left?
>
>Under SUNOS I do a "pstst -T". I haven't seen a LINUX command yet to tell me
>how much is actually in use.

The 'free' command from the 2.1ps distribution does this. If swapping
is not enabled, it displays only memory; otherwise it displays two lines,
one for memory and one for swap, each giving total, used, and free, and
memory also mentioning how much is shared.

Of course, you could always notice what swapon reports:

Adding Swap: 1061 pages (4345856 bytes) swap-space

or something of the sort.

-- Owen
[email protected]


[next article]
From: [email protected] (A. V. Le Blanc)

Subject: include/linux and include/asm (was Re: kernel building problems)
Message-ID: <[email protected]>
Date: 14 Apr 92 14:58:22 GMT
References: <[email protected]> <[email protected]
nki.FI>
Organization: Computing Centre, University of Manchester
Lines: 25

In article [email protected].
FI (Mika Pekka Liljeberg) writes:
>
>I compiled .95c first without and the with the HJ patch. There were no
>problems either way. Then I moved the kernel includes under /usr/include/linux
>with HJ's conv script. I hope this will become "official" practise, since
>it makes everything ever so much simpler. How about it, Linus? 🙂

The include files distributed with the kernel should not be part of
normal C programs, and are closely linked with the kernel. For the
purpose of compiling fundamental utilities, they are often useful, and
for this reason I have usually put at least those in linux/include/linux
and those in linux/include/asm into /usr/include/linux and /usr/include/asm
respectively.

I hope this will NOT become the official practice, in that I do NOT want
to see these files 'distributed' with the compiler; this would cause too
many problems coordinating versions. The include files used to compile
Linux MUST be able to be changed as Linux changes, and so are distinct
from the include files used by the C compiler. In any case, the only
compilation about which I have heard that these include files caused a
difficulty was the compilation of tools/build, which is, of course, not
part of the kernel, but a tool used to put the kernel together.

-- Owen
[email protected]


[next article]
From: [email protected] (David Engel)

Subject: Re: tar files on floppy.
Message-ID: <[email protected]>
Date: 15 Apr 92 01:12:18 GMT
References: <[email protected]>
Organization: Optical Data Systems, Inc.
Lines: 17

[email protected] (A. V. Le Blanc) writes:
: For larger files, it would be best to split them up, copy them over
: on separate floppies, and concatenate them, that is, assuming that
: you have no DOS partition on your Linux machine. Unfortunately I
: cannot point you to any standard utility running on DOS which would
: do this painlessly. I did this once with a GNUish DOS dd, which I
: found in the anonymous ftp at, I believe, grape.ecs.clarkson.edu.

I have a version of tar (really pax) which can read and write standard,
multi-volume, Unix tar-files on floppy. I think it's available on Simtel
(and mirrors) ... Yep, it is. Look for pax2exe.zip in the filutl area.

David
--
David Engel Optical Data Systems, Inc.
[email protected] 1101 E. Arapaho Road
(214) 234-6400 Richardson, TX 75081


[next article]
From: chojnows@osiris (Brian Chojnowski)

Subject: charset oriented GUI
Message-ID: <[email protected]>
Date: 15 Apr 92 01:06:36 GMT
Sender: [email protected] (News)
Organization: University of Illinois at Urbana
Lines: 13


I have been writting some user interface routines, and have just
started to add a crude windowing sytem to the bunch. There are a few
multi-user games that I am writting that Linux would be ideal for, but I
need a little graphics power. Since this is for Pc's, I figgured if I could
devise a standard of somekind, and someone was able to provide me with
source for a telnet program, and some idea as to where in the kernel or
device drivers to look, I could implement a windowing environment that wrote
directly to the video card. Because of my limited experience with Unix,
given telnet source, I would have no problem modifying that, but I have no
idea how to start on the kernel/device end.

-Brian


[next article]
From: [email protected]

Subject: mtools binaries for linux request!
Keywords: mootls binary, linux, archive
Message-ID: <[email protected]>
Date: 15 Apr 92 01:07:58 GMT
References:
Sender: [email protected]
Followup-To: [email protected]
Distribution: na
Organization: Michigan State University
Lines: 3

Where can I find 'mtools' binary for linux?
With Thanks,
- ishwar([email protected])


[next article]
From: [email protected] (Peter Williams 8169821)

Subject: man.1 files available for tex, mf and related programs
Message-ID: <[email protected]>
Date: 15 Apr 92 01:24:56 GMT
Sender: [email protected]
Organization: Design Computing Unit, The University of Sydney
Lines: 13
Nntp-Posting-Host: dean.arch.su.oz.au

In pub/linux on archsci.arch.su.oz.au (129.78.66.1) there are now available
the man.1 files for tex, latex, mf, dvips, etc.

They are in the file texmfman.tar.Z and should be detarred in the "man"
directory under which you wish them installed. They should automatically go into
the "man1" directory within the directory they are untarred in.

Don't forget to update your "whatis" file.

--
Peter Williams |e-mail: [email protected]
Key Centre for Design Quality |phone: +61-2-692 2053 or +61-2-660 6156
University of Sydney |+++++++++++++++++++++++++++++++++++++++


[next article]
From: [email protected] (Peter Williams 8169821)

Subject: Re: GCC2.1 (Wonderful !)
Keywords: gcc, 2.1, 1.40
Message-ID: <[email protected]>
Date: 15 Apr 92 01:52:54 GMT
References: <[email protected]> <[email protected]>
Sender: [email protected]
Organization: Design Computing Unit, The University of Sydney
Lines: 25
Nntp-Posting-Host: dean.arch.su.oz.au

In article <[email protected]>, [email protected] (H
ongjiu Lu -- Graduate Student) writes:
|>
|> It seems that gcc 2.1 can run with 2MB RAM. I think it's the time we
|> retire gcc 1.40. Lots of problems about kernel and gcc 2.1 are related
|> to the old gcc 1.40. With gcc 2.1, we can use the shared libs as the
|> standard. So we can squeeze more stuffs into rootimage. I really miss
|> bash. I had to copy it from 0.12 rootimage.
|>
|> What do you guys think?
|>
|>
|> H.J.

I agree and will undertake to make shared *.a versions of the various tex, mf,
dvips and associated programs available.

I am already running shared versions of these programs and the savings are
about 15-30% per program file. This is less than the savings made on many other
programs due to their large initial size. I will look into making further gains
when I have a better understanding of how to create shared "libraries".

--
Peter Williams |e-mail: [email protected]
Key Centre for Design Quality |phone: +61-2-692 2053 or +61-2-660 6156
University of Sydney |+++++++++++++++++++++++++++++++++++++++


[next article]
From: [email protected] (Rik Faith)

Subject: Re: Linux SCSI support
Keywords: scsi, future domain, linux
Message-ID: <[email protected]>
Date: 14 Apr 92 20:54:29 GMT
References: <[email protected]> <1992Apr10.223301.24673@athena.
mit.edu>
Sender: [email protected]
Organization: UNC-Chapel Hill Computer Science
Lines: 12

In article <[email protected]>, [email protected] (M. Sag
gaf) writes:
> Is there any way Linux can work with a SCSI Future Domain 1670 host
> adapter?

I'm one of the people working on the FD 1680 (which is a 1670 with floppy
support) driver. I expect to have it finished early in May (right now Ken
and I are somewhat swamped with classes). Perhaps you can e-mail me in May
and check out the progress.

--
Rik Faith: [email protected]
Paradox is the question of Chaos.


[next article]
From: [email protected] (Michael Pereckas)

Subject: Ghostscript
Summary: it works
Keywords: ghostscript port
Message-ID:
Date: 15 Apr 92 04:22:47 GMT
Sender: [email protected] (News)
Organization: University of Illinois at Urbana
Lines: 21

I have successfully compiled Ghostscript 2.4 with gcc 2.1.
It was a fairly easy compile. A few things needet changing, but I
managed and I'm a lousy programmer and I don't know C. My system was
to comment things out and hope for the best in most cases. Anyway, it
seems to work, and I've printed actual documents on my Alps Allegro 24,
which is Epson compatible. I have in the past had no success getting
the MS-DOS port to print anything but the included test files without
crashing with some scarry error messages, so I conclude the Linux is
useful, for that alone, plus with Linux I can do other work while the
printer grinds away at a typical 10-15 min per page, which is the usual
speed for this printer in graphics mode, and not the fault of GS.

I'm not prepared to provide diffs for the GS sources now, and my binary
only supports the epson driver and uses shared libs, so I'm not going
to upload it. I'm also not building another one, since it takes several
hours on my 2MByte system. I will try to answer questions by e-mail,
however.

Thanks to all who made Linux possible!

Michael Pereckas


[next article]
From: [email protected] (Michael Pereckas)

Subject: Yet another X question
Summary: how big will X be?
Keywords: X Windows
Message-ID:
Date: 15 Apr 92 04:33:30 GMT
Sender: [email protected] (News)
Organization: University of Illinois at Urbana
Lines: 18

I know it is sort of early to be asking this, but does anyone have a
guess as to how much disk space X Windows will
need? I'm trying to get a handle on how much money I'll have to
come up with to upgrade my system to run it when it becomes
available. I know better than to ask when that will be.

How much thought has gone into supporting graphics in a simpler manner?
I know X is useful, and many nice programs use it, but it is notoriously
big. A simple, small graphics interface for Linux might be useful for
people with small systems. For example, I need a DVI previewer, and
probably many other people do, too. If the only solution (I'm looking
to the future here) is X, then many will be shut out of really using
Linux until we can upgrade. I know that there has been some discussion
of this sort of thing recently, and I'd like to say that I am in favor
of porting X, and I like the idea of having X, but maybe it should not
be the only answer. Any thoughts?

Michael Pereckas


[next article]
From: [email protected] (Hongjiu Lu -- Graduate Student)

Subject: Re: Yet another X question
Keywords: X Windows
Message-ID: <[email protected]>
Date: 15 Apr 92 05:32:27 GMT
References:
Sender: [email protected] (USENET News System)
Organization: Washington State University
Lines: 15

In article [email protected] (Michael Pereckas)
writes:
>I know it is sort of early to be asking this, but does anyone have a
>guess as to how much disk space X Windows will
>need? I'm trying to get a handle on how much money I'll have to
>come up with to upgrade my system to run it when it becomes
>available. I know better than to ask when that will be.

From my own experience, a real X11R4 can take as little as 10MB to run.
on a 386. I only keep the minimum stufffs I need. Beside MIT stuffs,
I added some other programs, like xgif and xdvi. I can put off X source
from the net and compile it without much problem. I believe, for X11R5,
20MB should be a safe bet.


H.J.


[next article]
From: [email protected] (Einar Indridason)

Subject: Re: User's Guide project
Message-ID: <[email protected]>
Date: 15 Apr 92 09:29:04 GMT
References: <[email protected]> edu>
Sender: [email protected]
Lines: 24

In [email protected] (Ajay Shah) writes:

>[email protected] (Joseph Knapka) writes:

>>1) Do you think there is a need for such a document?

>Yes, I think such a document would be rather valuable.

>Linux is great.. what more could a hacker ask for?
>But we could make the results of the same work accessible to
>a much larger class of users if we add "a little" effort
>on documentation and easy-to-use installation/administration.


And perhaps some co-ordination regarding patches. Perhaps a few persons
would take each patch, and compare it to some "standard" source tree.

I have used Minix a bit, but haven't installed Linux (yet), but I'd
hate to see the "patch-mess", that minix has gotten into, hit Linux as
well.


--
[email protected]


[next article]
From: [email protected] (Michael Haardt)
Newsgroups: comp.os.linux,comp.sys.transputer

Subject: First small success in using Linux as transputer host
Message-ID: <[email protected]>
Date: 11 Apr 92 23:09:15 GMT
Sender: [email protected] (Newsfiles Owner)
Reply-To: [email protected] (Michael Haardt)
Organization: Gandalf - a 386-20 machine
Lines: 49
Nntp-Posting-Host: rama

I ported dmake 3.8 to get a decent make on Linux (it compiles without
much effort, as usual on Linux) and built my small transputer tools a
few minutes ago.

Well, I started the transputer memory test with the start server and:
wow! It worked, and *so* much faster than under MINIX. This simple
server uses /dev/port, and it was painful slow under MINIX. Once again
I am really impressed by the speed of Linux.

I am still working on my transputer assembler and linker and I am still
not sure if I will release it to public domain, but if there is
interest, I will make binaries for Linux available per ftp. I can't
make available the server, because it is a based on the server in "Das
Transputerbuch", but I will write my own server next time.

I will be away for vacation with my gf during next month. When I come
back, I will develop a concept for a device driver to use the link
interface at highest possible speed. I guess that the PC bus will be
the bottleneck, because the link is able to work at 20 MBit/second, but
it is used via an 8 bit port. Here are two ideas of mine:

- /dev/tlink, a character device with blocking read and write to the
link. Perhaps non blocking read/write may be supported, I don't know
if it is useful. ioctl's will control reset, error and analyse.

- /dev/tlink for the link, /dev/tlinkctl for controlling error, reset and
analyse with kind of message passing via fs. This makes ioctl obsolete
and I like it a lot since I read the Plan 9 papers, but on the other
side it is totally different from anything else in Linux 🙁

Any other suggestions?

I haven't looked at device drivers in Linux yet, how difficult it will
be and what has to be done to support blocking/non blocking operations.

The transputer should be fast enough to use polling without loosing
performance. I think the overhead of interrupts would even slow down
the data transfer, but I have no practical experience in that.

I also thought about a lock to allow only one process to deal with
/dev/tlink in the way like Parsytec uses semaphores for OpenLink, but I
decided to let the server program deal with that. People who open
devices in their programs should know what they do and no restrictions
should be put on them.

If you post follow-ups, please send me a copy per email during next
month. I will post a summary.

Michael


[next article]
From: [email protected] (Joel M. Hoffman)

Subject: Re: Yet another X question
Keywords: X Windows
Message-ID: <[email protected]>
Date: 15 Apr 92 12:59:30 GMT
References: <[email protected]>
Sender: [email protected] (USENET News system)
Organization: University of Maryland, College Park
Lines: 34
Nntp-Posting-Host: next.wam.umd.edu

In article <[email protected]> [email protected] (Ho
ngjiu Lu -- Graduate Student) writes:
>In article [email protected] (Michael Pereckas)
writes:
>> [Is there a smaller, simpler, graphics support update planned?]
>>
>
>From my own experience, a real X11R4 can take as little as 10MB to run.
>on a 386. I only keep the minimum stufffs I need. Beside MIT stuffs,
>I added some other programs, like xgif and xdvi. I can put off X source
>from the net and compile it without much problem. I believe, for X11R5,
>20MB should be a safe bet.

I also would like to see simpler graphics support. I really don't
want to clobber by system with X, and all I really need can be
provided much more simply. First, I'd like to have downloadable
character sets, so I can use the Hebrew routines I've been using under
DOS. Second, I'd like a .dvi previewer. So far, I've seen nothing
that's even nearly as good as the .dvi drivers that come with emTeX.
Any chance of porting those?? (E. Mattes? Please!) Finally,
sometime down the road I'd like to see a full DOS emulator for Linux,
and that will require graphics support.

It seems to me a ggod approach would be to emulate DOS's graphics
interface, and trap the reads and writes to the graphics card. That
would make all of the above tasks much easier, and wouldn't demand so
much of the system.

Comments?

-Joel
([email protected])

p.s. Please don't write back telling me how terrible DOS is. I have
many useful programs for DOS, and everything I sell is written for
DOS. I can't give it up.


[next article]
From: [email protected] (david.a.cuka)

Subject: Re: Linux Printer
Summary: Printer now works
Keywords: linux, printer
Message-ID: <[email protected]>
Date: 15 Apr 92 13:46:50 GMT
References: <[email protected]> <[email protected]
du>
Followup-To: comp.os.linux
Organization: AT&T
Lines: 24

In article <[email protected]> [email protected] (Steve Anse
ll) writes:
>In article <[email protected]> [email protected]
(david.a.cuka) writes:
>>
>>I got a "ENODEV", so I tried using minor device number 1, which did not
>>work either. Is there some addtional setup that I am missing? I
>>am running on v0.95a. The printer is a Panasonic KXP-1024 parallel
> ^^^^^^
>
>Unless you applied the kernel patches yourself (you don't mention it above)
>you don't have the lp driver support in the kernel. You need to get a copy
>of v0.95c+ which has the lp driver support in the kernel already. Once you
>have that there shouldn't be much to it other than what you show above.
>
>
>--
> -Steven T. Ansell
> Unix Consultant
> Computing Services U.C.D.


Thanks to all who replied, I downloaded 0.95c+ and the printer now works.
Also, my extended partition is now recognized by linux and by mtools.

DAC


[next article]
From: [email protected] (LCDR Michael E. Dobson)

Subject: Re: tar files on floppy.
Message-ID: <[email protected]>
Date: 15 Apr 92 13:16:48 GMT
References: <[email protected]> <1992Apr14.22130
[email protected]>
Organization: Naval Medical Research & Development Command
Lines: 28

In article <[email protected]> [email protected] (David Engel) writes:
>[email protected] (LCDR Michael E. Dobson) writes:
>: I am just about finished with linking everything in 2.1shared-A.tar.Z with
>: shared libs and symlinking into /usr/bin. A word of caution, DON'T use
>: shared libs or symlinks on anything in /bin or /etc if / and /usr are on
>: different partitions. They won't work until /usr is mounted. Keep a good
>: bootimage floppy and root floppy handy just in case.
>
>If you copy the shared lib to /lib instead of symlinking it (done by the
>install script) you can use shared binaries in /bin and /etc. You can also
>safely remove /usr/shared the next time you upgrade libraries. It's still
>a good idea to have "emergency" boot and root floppies handy however.
>
Good news. I had thought about this and will very likely do it. With only
a 4 Meg root partition, I ran out of space attempting to install the shadow
passwd utilities. Converting all binaries in the root partition to shared
libs will save lots of space. Now I just need to go back and grab the src
for the shadow stuff to compile and link it for shared libs. I assume I
should also cp the shared binaries into /usr/bin rather than using symlinks
to /usr/shared/bin.

Mike

--
LCDR Mike Dobson, Sys Admin for | Internet: [email protected]
nmrdc1.nmrdc.nnmc.navy.mil | UUCP: ...uunet!mimsy!nmrdc1!rdc30
AT&T 3B2/600G Sys V R 3.2.3 | BITNET: dobson@usuhsb or nrd0mxd@vmnmdsc
WIN/TCP for 3B2 R3.2 | MCI-Mail: 377-2719 or [email protected]


[next article]
From: [email protected] (Rogier Wolff)

Subject: Loadable Device drivers.
Keywords: devices, drivers, linux, project
Message-ID: <[email protected]>
Date: 15 Apr 92 12:55:11 GMT
Sender: [email protected] (UseNet News System)
Organization: Delft University of Technology, Dept. of Electrical Engineering
Lines: 22
Nntp-Posting-Host: dutecaj.et.tudelft.nl


Hi everyone,

I will be doing a project involving the implementation of
"loadable device drivers" in linux. This will facilitate developing
new device drivers: you should be able to:

gcc device.driver.c
unloaddevice
loaddevice a.out b

/etc/mount /dev/ /mnt

If anybody has suggestions, they are welcome. One thing though:
please don't tell me that you've done this already... I'd have to
go looking for another project.

Roger
--
If the opposite of "pro" is "con", what is the opposite of "progress"?
(stolen from [email protected])
EMail: [email protected] ** Tel +31-15-783644 or +31-15-142371


[next article]
From: [email protected] (Adam Thompson)

Subject: Re: Yet another X question
Keywords: X Windows
Message-ID: <[email protected]>
Date: 15 Apr 92 14:58:06 GMT
References: <[email protected]>
Organization: University of Manitoba, Winnipeg, Canada
Lines: 27

In <[email protected]> [email protected] (Hongjiu Lu
-- Graduate Student) writes:

>From my own experience, a real X11R4 can take as little as 10MB to run.
>on a 386. I only keep the minimum stufffs I need. Beside MIT stuffs,
>I added some other programs, like xgif and xdvi. I can put off X source
>from the net and compile it without much problem. I believe, for X11R5,
>20MB should be a safe bet.
>H.J.

Perhaps you're assuming he doesn't want to compile it 🙂

According to the docs in the R5 distribution, the source tree untars to
roughly 150-200 Megs. To compile, expect to need at least another 100Meg.

As far as compiled binaries go, using shared libraries, on a Sun i386,
the full MIT core plus about 50% again in added-on goodies takes somewhere
in the range of 60-70 Megs. Of course, to simply run R5 'functional', then,
yes, expect to use ~20 megs. Perhaps even a bit less if you only have one
WM on call...

-Adam Thompson
[email protected]
--
= Adam Thompson ---- Computer Engineering ---- University of Manitoba =
= [email protected] = "When you have eliminated the improbable, =
= ...!uunet!decwrl!alberta!\ = whatever is left, however impossible, =
= ccu.UManitoba.CA!umthom61 = must be the answer." =


[next article]
From: [email protected] (Kent Williams)

Subject: ALLWAYS IN2000 SCSI Controller
Message-ID: <[email protected]>
Date: 15 Apr 92 15:11:11 GMT
Sender: [email protected]
Lines: 12

To make it short, I've got one of these. Does anyone who's trying to run
Linux have one of these? I would try to bring linux up if I could
get my scsi controller to work.

Failing that, how big a disk would I need to get a development system
to boot??


--
Kent Williams | "Do you see your cerebellum as a lightbulb or a cog? I
[email protected] | saw mine as gristle so I fed it to the dog. But it
Quote: Bevis Frond | taste so bad, that she left it in the bowl ..."


[next article]
From: [email protected] (Johan W}hlin)

Subject: Re: tar files on floppy.
Message-ID: <[email protected]>
Date: 15 Apr 92 15:31:26 GMT
Sender: [email protected] (Usenet)
Organization: Royal Institute of Technology, Stockholm, Sweden
Lines: 13
Nntp-Posting-Host: noir.nada.kth.se

In article <[email protected]> [email protected]
c.nnmc.navy.mil (LCDR Michael E. Dobson) writes:

unrelated stuff deleted ..

>I am just about finished with linking everything in 2.1shared-A.tar.Z with
>shared libs and symlinking into /usr/bin. A word of caution, DON'T use
>shared libs or symlinks on anything in /bin or /etc if / and /usr are on
>different partitions. They won't work until /usr is mounted. Keep a good
>bootimage floppy and root floppy handy just in case.

Just put the shared lib in /lib and the symlink in /usr/.. Ohh, and do keep
good images around 8).
/Johan


[next article]
From: [email protected] (Andreas Mengel)

Subject: Need help with swapping
Keywords: linux, swap
Message-ID: <[email protected]>
Date: 15 Apr 92 16:18:47 GMT
Sender: [email protected]
Organization: University of Oldenburg, Germany
Lines: 41

Hi there!

First of all a BIG THANX to Linus and all the rest who have made this nice l
little *nix work. I just tossed aside my old ISC and installed it.
Looks great, even my C++-sources compile with the new gcc2.1 :-)))

But I also had some little problems:

- how can I set up the swap-space correctly ?
I tried mkswap /dev/hd03 10240 (10MB, 3rd partition) with and
without the "-c"-flag to mkswap. And I tried setting up a swap-file
with the same result: swapon complains about not being able to find
the swap-space-signature (I think its something like "SWAP-SPACE",
looked through the swap-file). Anything I am doing totally wrong ???
(B.t.w.: I am using an adaptec-1542b-scsi-controller with a 150MB
disk, if that matters)

- when I compiled my sources (C++) to object-files and put them together
into a library using "ar rc libfoo.a foo1.o foo2.o ..." I was not
able to link these libraries to my other sources. ld complained
about undeclared symbols in the text-segment but nm showed that all
the symbols were in the library. I helped me out creating a shared
object (with ld -x -r) and linking that.
Any suggestions.

Thanx, falcon

Keep working on Linux, this is the best thing I have seen since I first
saw X-Windows.
--
Bye,
Falcon
-------------------------------------------------------------------------------
Real Name: Andreas Mengel
Bitnet: 169371@DOLUNI1 | UUCP: [email protected]
Internet: [email protected]
falcon@aragorn | falcon@gimli | falcon@theoretica
+-------------------------------------------------+
|Some say I'm lazy, but others say that's just me.|
|Some say I'm crazy, I guess I'll always be! (GnR)|
+-------------------------------------------------+


[next article]
From: [email protected] (Nan-shan Chen)

Subject: help, incompatible video ..
Summary: My video card is incompatible, a hardware scroll does not work
Keywords: Softscroll
Message-ID:
Date: 15 Apr 92 17:07:48 GMT
Sender: [email protected] (News Uni-Paderborn)
Organization: Uni-GH Paderborn
Lines: 21
Nntp-Posting-Host: pbhrzx


Hi everybody,

I have just downloaded the Boot & Root image of LINUX. There is, however,
a serious problem in the first place. I could not even view the screens.
That's the same problem as in MINIX. But I could set it to software
scroll under MINIX just by pressing a key or doing:

setterm -softscroll

There is no "softscroll" option of setterm in LINUX.

What should I do?



--

////\
{ 0-0 }
\ o /


[next article]
From: [email protected] (Fureching Jeng)

Subject: expr
Keywords: Softscroll
Message-ID: <[email protected]>
Date: 15 Apr 92 17:53:01 GMT
References:
Sender: [email protected]
Reply-To: [email protected] (Fureching Jeng)
Organization: Bellcore MRE
Lines: 8
Nntp-Posting-Host: ipanema.bellcore.com

I have installed gcc 2.1 on my machine. Unfortunately, I can't run g++ without

"expr". Does anyone know where I can get "expr"?


Thanks,

FC


[next article]
From: [email protected] (Theodore Y. Ts'o)
Newsgroups: comp.os.linux,alt.config

Subject: Re: Too many duplicated messages
Message-ID: <[email protected]>
Date: 15 Apr 92 17:51:09 GMT
Sender: [email protected] (News system)
Organization: Massachusetts Institute of Technology
Lines: 34
Nntp-Posting-Host: tsx-11.mit.edu

>From: claris!wet!dhl%[email protected] (Dahai Li)
>Reply-To: claris!wet!dhl%[email protected] (Dahai Li)

>For the past week, there are many duplicated messages appearing in the
>linux-digests. And I did notice some one complained about it. However, I
>don's see any reponse, and they keep on coming.

I did send a response; as I mentioned, this is because of a bug in C
News about how it handles aliased newsgroups. (I had aliased
alt.os.linux to comp.os.linux, so that people who send articles to only
alt.os.linux would still appear in the digest.) However, it seems that
doing this is counterproductive at this point, since I suspect we are
annoying more people with the duplicated messages than we are helping by
catching the stray message which is sent to only alt.os.linux. So I
will be discontinuing this forwarding of alt.os.linux as of today. If
someone sends a message to only alt.os.linux, you will not see their
message in the Digest.

I asked about rmgroup'ing alt.os.linux a week or so ago, and some
people responded that comp.os.linux still hadn't been created yet at
their site. So last week I sent out newgroup messages for
comp.os.linux again. If those sites haven't created comp.os.linux by
now, they probably have lame sysadmins and it is unlikely they will
get around to creating those groups any time soon. So, I don't think
we really have an obligation to accomodate them much longer; having
two newsgroups is somewhat confusing, and we should really transition
everyone over to using comp.os.linux.

So, let me repeat my call. Shall we rmgroup alt.os.linux? I think its
time has come.
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
Theodore Ts'o bloom-beacon!mit-athena!tytso
308 High St., Medford, MA 02155 [email protected]
Everybody's playing the game, but nobody's rules are the same!


[next article]
From: [email protected] (Kaleb Keithley)

Subject: Re: Yet another X question
Keywords: X Windows
Message-ID: <[email protected]>
Date: 15 Apr 92 17:43:46 GMT
References:
Organization: Jet Propulsion Laboratory, Pasadena, CA
Lines: 17

In article [email protected] (Michael Pereckas)
writes:
>I know it is sort of early to be asking this, but does anyone have a
>guess as to how much disk space X Windows will
>need? I'm trying to get a handle on how much money I'll have to
>come up with to upgrade my system to run it when it becomes
>available. I know better than to ask when that will be.
>

When I built R5 under ESIX V4, (and didn't do contrib, docs, or the 100dpi
fonts, but did do shared AND static libs) it used >100 meg.


--

Kaleb Keithley [email protected]

Not authorized, in any way, shape, or form, to speak for anyone.


[next article]
From: [email protected] (Teresa Selling)

Subject: trouble with dialin login and vi
Message-ID: <[email protected]>
Date: 15 Apr 92 18:35:30 GMT
Sender: [email protected] (USENET News System)
Organization: Indiana University at South Bend
Lines: 9
Nntp-Posting-Host: peabody.iusb.indiana.edu

I can use vi fine on my system with Linux 0.95c+ installed, but when I
call into school I have trouble with vi messing up the lines and such.

I already do stty erase "^?" to make the backspace work. Is there something
I need to do to fix this other problem. I had no trouble with vi on the
school computer before (HP unix) with linux 0.12....any suggestions
appreciated.
--
Teresa Selling, IUSB, Indiana.


[next article]
From: [email protected] (Jay Maynard)

Subject: SVGA: 0.95c+ doesn't switch modes
Message-ID: <[email protected]>
Date: 15 Apr 92 18:37:20 GMT
Sender: [email protected]
Organization: UT Health Science Center Houston
Lines: 11
Nntp-Posting-Host: oac.hsc.uth.tmc.edu


I just put a Trident SVGA9000 in my machine. When I boot 0.95c+ (and 0.95a,
too), it shows the list of available SVGA modes. When I select one, though,
the card is never put into the other mode; instead, I still have a 25x80
screen. If I select a mode with more lines, the last (n-25) aren't shown on my
screen at all. What's going on here?
--
Jay Maynard, EMT-P, K5ZC, PP-ASEL | Never ascribe to malice that which can
[email protected] | adequately be explained by a .sig virus.
"While we liked developing Windows applications, we never inhaled." --
Philippe Kahn, in Infoworld, 13 Apr 1992


[next article]
From: [email protected] (Kevin Cummings)

Subject: Re: GCC2.1 (Wonderful !)
Keywords: gcc, 2.1, 1.40
Message-ID: <[email protected]>
Date: 15 Apr 92 19:15:09 GMT
References: <[email protected]> <[email protected]>
<[email protected]>
Sender: cummings@hammer (Kevin Cummings)
Organization: Prime Computer R&D
Lines: 29

OK, coming from one of the companies that pioneered shared libraries
for the mini-computer industry, and therefore having a VERY one-sided
view of them, I need to get a few questions answered about LINUX
shared libraries. So here goes:

Are the shared libraries dynamicly linked to at runtime? or are
the .a files bound at install time to whatever shared library
is in /lib/whatever? If the latter, then this would explain the
"extra" step necessary to "install" the programs that use them
and why .a files are necessary. Does this mean that these
programs that use shared libraries must be re-installed every time
the shared library is updated? (If so, then: Barf, puke, etc.)

Please excuse if these are newbie style questions, but working for
a proprietary computer manufacturer for 15 years is an easy way to not
become a UNIX expert.

=================================================================
Kevin J. Cummings Prime Computer Inc.
20 Briarwood Road 500 Old Connecticut Path
Framingham, Mass. Framingham, Mass.

Work: [email protected]
Home: [email protected]

Std. Disclaimer: "Mr. McKittrick, after careful consideration,
I've come to the conclusion that your new
defense system SUCKS..." -- War Games
=================================================================


[next article]
From: [email protected] (Brian Olsen)

Subject: Re: Linux Printer
Keywords: linux, printer
Message-ID: <[email protected]>
Date: 15 Apr 92 19:31:26 GMT
References: <[email protected]> <[email protected]
du> <[email protected]>
Sender: [email protected]
Organization: University of Missouri - Rolla
Lines: 15


Hello!
I'm running linux 0.95c+ and have been working on connecting up my printer
(Panasonic KXP-1123).
When I tried
/bin/mknod /dev/lp1 c 6 0
I got a 'device not found' error (to that effect)
using
/bin/mknod /dev/lp1 c 6 1
I was able to print to stuff, but I lost a lot of character in the printout
(about one in every three or four).

Suggestions anyone?
Brian
[email protected]


[next article]
From: [email protected] (Kevin Cummings)

Subject: Re: Yet another X question
Keywords: X Windows
Message-ID: <[email protected]>
Date: 15 Apr 92 19:29:03 GMT
References: <[email protected]>
<[email protected]>
Sender: cummings@hammer (Kevin Cummings)
Organization: Prime Computer R&D
Lines: 44

In article <[email protected]>, [email protected]
(Adam Thompson) writes:
> In <[email protected]> [email protected] (Hongjiu
Lu -- Graduate Student) writes:
>
> >From my own experience, a real X11R4 can take as little as 10MB to run.
> >on a 386. I only keep the minimum stufffs I need. Beside MIT stuffs,
> >I added some other programs, like xgif and xdvi. I can put off X source
> >from the net and compile it without much problem. I believe, for X11R5,
> >20MB should be a safe bet.
> >H.J.
>
> Perhaps you're assuming he doesn't want to compile it 🙂
>
> According to the docs in the R5 distribution, the source tree untars to
> roughly 150-200 Megs. To compile, expect to need at least another 100Meg.
>
> As far as compiled binaries go, using shared libraries, on a Sun i386,
> the full MIT core plus about 50% again in added-on goodies takes somewhere
> in the range of 60-70 Megs. Of course, to simply run R5 'functional', then,
> yes, expect to use ~20 megs. Perhaps even a bit less if you only have one
> WM on call...

I plan on trying to run X over a slip connection to a "mainframe" host.
What I need locally is the local server, and maybe a few local
applications (a window manager, xterm, xmh, etc). Most of the
other applications don't need to be local, but instead can run
from remote hosts. I'm hoping that this is the 20MB scenario. In fact
I'm willing to go to 40 or 50MB locally in this scenario. I'm not
interested in compiling ALL of the X distribution, but I imagine that
there will be a fair number of people out there who will want to, mostly
because they won't have a live network connection, and will need to have
everything locally.

=================================================================
Kevin J. Cummings Prime Computer Inc.
20 Briarwood Road 500 Old Connecticut Path
Framingham, Mass. Framingham, Mass.

Work: [email protected]
Home: [email protected]

Std. Disclaimer: "Mr. McKittrick, after careful consideration,
I've come to the conclusion that your new
defense system SUCKS..." -- War Games
=================================================================


[next article]
From: [email protected] (Hongjiu Lu)

Subject: Re: Need help with swapping
Keywords: linux, swap
Message-ID: <[email protected]>
Date: 15 Apr 92 18:12:55 GMT
References: <[email protected]>
Sender: [email protected] (USENET News System)
Organization: Washington State University
Lines: 16

In article <[email protected]>, Andreas.Men
[email protected] (Andreas Mengel)
|> - when I compiled my sources (C++) to object-files and put them together
|> into a library using "ar rc libfoo.a foo1.o foo2.o ..." I was not
|> able to link these libraries to my other sources. ld complained
|> about undeclared symbols in the text-segment but nm showed that all
|> the symbols were in the library. I helped me out creating a shared
|> object (with ld -x -r) and linking that.
|> Any suggestions.
|>

Use ranlib before linking.

Should that be in FAQ? I thought everybody knows it.


H.J.


[next article]
From: [email protected] (Teresa Selling)

Subject: rebuilding kernel & running a comm session at same time problem
Message-ID: <[email protected]>
Date: 15 Apr 92 20:38:47 GMT
Sender: [email protected] (USENET News System)
Organization: Indiana University at South Bend
Lines: 8
Nntp-Posting-Host: peabody.iusb.indiana.edu

I have found that when I am rebuilding the kernel and downloading from
an xcomm session I get constant HD timeout errors from the download session
using rz.

Is this some sort of buffer problem or what? And is it fixable?

--
Teresa Selling, IUSB, Indiana.


[next article]
From: [email protected] (Kevin Cummings)

Subject: Re: expr
Keywords: Softscroll
Message-ID: <[email protected]>
Date: 15 Apr 92 20:42:12 GMT
References: <1992Apr15.175301.27390@w
alter.bellcore.com>
Sender: cummings@hammer (Kevin Cummings)
Organization: Prime Computer R&D
Lines: 20

In article <[email protected]>, [email protected].
com (Fureching Jeng) writes:
> I have installed gcc 2.1 on my machine. Unfortunately, I can't run g++ without
> "expr". Does anyone know where I can get "expr"?

These is an expr.a in the /usr/shared/obj directory of the 2.1shared-A.tar
file that is kinda part of gcc 2.1. Finish loading it on your
machine and install it along with your 2.1 shared library.

=================================================================
Kevin J. Cummings Prime Computer Inc.
20 Briarwood Road 500 Old Connecticut Path
Framingham, Mass. Framingham, Mass.

Work: [email protected]
Home: [email protected]

Std. Disclaimer: "Mr. McKittrick, after careful consideration,
I've come to the conclusion that your new
defense system SUCKS..." -- War Games
=================================================================


[next article]
From: chojnows@osiris (Brian Chojnowski)

Subject: ESDI drives
Message-ID: <[email protected]>
Date: 15 Apr 92 21:19:17 GMT
Sender: [email protected] (News)
Organization: University of Illinois at Urbana
Lines: 5


Well I just spent a pile of time trying to install 95a onto my hardrive
becasue I thought t was a IDE drive. After getting nothing but error 2's for
mount, I asked my officemate, and he said the drive type was ESDI. Is there
anything I can do to use this drive?


[next article]
From: [email protected] (Andrew Haylett)

Subject: Re: Apology and question
Message-ID: <1593@oberon>
Date: 15 Apr 92 09:22:23 GMT
References: <[email protected]>
Organization: GEC-Marconi Research Centre, Great Baddow, UK
Lines: 11

[email protected] (Bernard Joseph Gardner) writes:

: Anyway, my real question is with regard to cut and paste using the mouse. Has
: anyone implemented such a beastie?

Look in patches/selection-1.0.tar.Z on tsx-11 for a package that implements
this, at least for serial mice, I think. Improved version on the way,
allowing selection by word and line.
--
Andrew Haylett | Inet: [email protected] | Fax: +44 245 75244
GEC-Marconi Research | Tel: +44 245 73331 x.3283 | Telex: 995016 GECRES G


[next article]
From: [email protected] (Rick Anderson)

Subject: Re: What is linux?
Keywords: linux
Message-ID: <[email protected]>
Date: 14 Apr 92 16:10:20 GMT
References: <1992Apr1.165146.1
[email protected]> <[email protected]>
Sender: [email protected]
Organization: University of Virginia
Lines: 12

I am a new user to Linux. I have gotten 95a images onto 1.2 diskettes, installed
the root image on my harddrive. Works great. But now I want to install GCC,
Mtools and other usr utilities into Linux partition. I am working from an-
other PC using MS/DOS to ftp archives to my site. How do I prepare diskettes
that linux will read and mount so I can install all these unix files?

With Linux, I was using rawrite to create diskettes from 'image' files.
Now I have unix file archives (tar) and don't know how to move them to Linux.
Can someone point me in the right direction? The issue is moving unix
archives on an MS/DOS system to the Linux system.

Thanks in advance.


[next article]
From: [email protected] (Ross Biro)

Subject: Config 95c+ uploaded to tsx-11
Summary: configc+.tar.z at tsx-11.mit.edu
Keywords: ethernet wd8003 config debugger
Message-ID: <[email protected]>
Date: 15 Apr 92 23:15:48 GMT
Sender: [email protected] (News Service)
Organization: DSG, Stanford University, CA 94305, USA
Lines: 26

I just uploaded configc+.tar.Z to tsx-11.mit.edu. It should appear
in a few hours, and it includes:

An improved configuration system which allows the specification
of arbitary constants in the devs.conf file.

The kernel debugger.

The new keyboard driver which has provisions for remapping keys.

A wd8003 ethernet driver.

most of what you need to get ka9q working with the ethernet driver.

If anyone wants to write documentation, or wants to write programs to
use some of the new features of the keyboard driver, send me some
email.




--
-----
Ross Biro [email protected]
Member League for Programming Freedom (LPF)
mail [email protected] for information


[next article]
From: [email protected] (Demian A. Johnston)

Subject: Re: ESDI drives
Message-ID:
Date: 15 Apr 92 23:59:22 GMT
Organization: Sophomore, Physics, Carnegie Mellon, Pittsburgh, PA
Lines: 13
In-Reply-To: <[email protected]>

I have 2 ESDI drives and I have successfully installed 0.95a by way of
the installation notes.. Make sure you do your pfdisk'ing' properly by
defining your hard disk geometry... I also have been successful
installing the 0.95c+ source and recompiling the kernel.. Very little
problems...

Good luck


Demian
---> [email protected]
---> [email protected]
---> [email protected]


[next article]
From: [email protected] (Matthew Jackson)
Newsgroups: comp.os.linux,alt.os.linux

Subject: How to set the path ?
Message-ID: <[email protected]>
Date: 15 Apr 92 00:42:26 GMT
Organization: University of New South Wales
Lines: 25

First off, thanks to all the work done by everyone to get Linux up and running.
Also, thanks to the writers of the new users and installation pages. I managed
to get 0.95a up and running with only one minor problem (accidentally swapping
the order for the hard disk device numbers).

Now I have a few questions (of course). When I was running the install scripts,
I checked they were flagged as executable, and tried to run them. No go. I then
tried /INSTALL/install and it worked ! OK, I thought Linux doesn't look in the
current directory for a file before searching the path (ala DOS) (is this
meant to work this way ? I am used to DOS and VMS, only just got started with
Unix). OK (he thinks again), why don't I put the current directory in the path ?
I typed set and got a list of the environment variables and there was one called
PATH, so I tried set PATH=....:. and still no go.

So can someone tell me how to change the path under Linux ?

Thanks in advance,

--
Matthew Jackson |
Computing Officer | Why didn't Noah swat those two
School of Banking and Finance | mosquitoes ?
University of New South Wales |
Australia |
[email protected] |