Files
2024-02-19 00:23:35 -05:00

606 lines
22 KiB
Plaintext

From: Digestifier <Linux-Activists-Request@news-digests.mit.edu>
To: Linux-Activists@news-digests.mit.edu
Reply-To: Linux-Activists@news-digests.mit.edu
Date: Mon, 23 Mar 92 11:30:28 EST
Subject: Linux-Activists Digest #139
Linux-Activists Digest #139, Volume #1 Mon, 23 Mar 92 11:30:28 EST
Contents:
Re: bg, fg, jobs etc (Drew Eckhardt)
Bad kernel build (0.95a) with GCC 1.40! Why???? (cm445a17)
Linux 0.95a installation (Michael Haardt)
Curses library (Michael Haardt)
Keymapping under kermit ? (Dirk Hohndel)
(none) (PSYXSGP@otago.ac.nz)
Linux-Activists Digest #137 (Peter Desnoyers)
Re: Free BSD release: future of Minix/Linux? (Stephen Trier)
Wanted. UUCP compatible mailer for Linux. (abs@sunvax.sun.ac.za)
HD timeout [was Re: Linux 0.95a installation] (Joel M. Hoffman)
Re: Free BSD release: future of Minix/Linux? (Lance Pickup)
Re: Free BSD release: future of Minix/Linux? (Pierre LEWis)
Booting Linux as an OS/2 MOST system (Jim Snowden)
Re: diff for extended partitions (David Engel)
Re: patches (David Engel)
Re: Bad kernel build (0.95a) with GCC 1.40! Why???? (Al Clark)
Re: 386 BSD and linux (Alan B Clegg)
Bug in putpwent() (Peter Orbaek)
minor linux 0.95a bugs (Gerald Heim)
----------------------------------------------------------------------------
From: drew@cs.colorado.edu (Drew Eckhardt)
Subject: Re: bg, fg, jobs etc
Date: 23 Mar 92 08:25:03 GMT
In article <1992Mar23.074606.12143@athena.mit.edu> PSYXSGP@otago.ac.nz writes:
>Does Linux support job control commands like bg, fg, job etc.
>I'm not a Unix guru so excuse me if this is a silly question.
>I wonder if such commands are a function of the kernel ..
>or of the shell
>Stephen Pearce, University of Otago, Dunedin, New Zealand.
>
This is a function of the shell. BASH supports job control, ash
(the distribution shell) does not.
------------------------------
From: cm445a17@socrates.umd.edu (cm445a17)
Subject: Bad kernel build (0.95a) with GCC 1.40! Why????
Date: Mon, 23 Mar 1992 08:36:55 GMT
Today I ventured into build the kernel for Linuc 0.95a with
GCC 1.40. I edit some of the makefiles so that the -fcombine-regs
flag was set and I did the make. Everything compiled fine so I decided
to try it out. I booted up with the new image and everything seems to
work except for the fact that some of the keyboard keys produce the
wrong characters (i.e. pressing / produces -). Does anyone know why
this is happening? Am I supposed to be using GCC 2.0? If so, what
is the status of GCC 2.0? (Alpha, Beta???), and can it be used reliably
with Linux? Thanks,
Desperate Jaime
------------------------------
From: michael@gandalf.informatik.rwth-aachen.de (Michael Haardt)
Subject: Linux 0.95a installation
Date: 21 Mar 92 19:41:57 GMT
Reply-To: u31b3hs@messua.informatik.rwth-aachen.de (Michael Haardt)
After ftp'ing a few megs, I installed Linux 0.95a this weekend on my
machine: Micronics board, 386-20, 2 MB RAM, HGC, 5.25 and 3.5 inch
drives and a WD 1006 with a Micropolis 70 MB MFM winchester drive.
(and lots of dust from over three years, but I doubt that counts:)
The rootimage was very helpful, I needed only a short while until the
machine booted the first time with root fs on winchester. Everything
uncompressed fine and worked on first try.
When I copy *really* big files, like creating the swap file, I sometimes
get "HD timeout" messages. Is this a warning or an error (read: can I
ignore it)?
After years with MINIX, I had to try out what might happen, if I start
two compilers parallel. I got a trashing machine, as expected :) But
something else happened as well:
Warning, pid 1450 older sibling 10 has mismatched ys link
Warning, pid 1450 younger sibling 10 has mismatched os link
What does that mean? It happens from time to time, if the machine starts
thrashing.
When I try to compile the kernel, compiling fork.c will get aborted by
"Program got fatal signal 2615". I tried to compile this, and a a few
more file without -O and finally got an Image file of 215552 bytes. I
copied this file to /dev/at0 and rebooted the machine. After printing
the usual information Loading ..... and number of ptys, I got a kernel
panic from swapper, trying to free unused page or similar (sorry, I'm
typing from memory). I can give more information, if I didn't make a
wellknown mistake... I used gcc 1.40 (newgcc) for compiling.
btw: Doesn't Linux create a core dump when a program produces a
segmentation fault?
There seems to be no putenv and no regex in the library, right?
Everything else seems to work fine, just now I am writing this article with
Origami, wnews and wmail only needed slightly changes (grrr, MINIX specific
code). I am impressed of the speed of Linux, if I would have known that
earlier, I would have changed earlier, no matter how much bugs the older
versions had and I am really looking forward to the possibility of compiling
the kernel.
Michael
------------------------------
From: michael@gandalf.informatik.rwth-aachen.de (Michael Haardt)
Subject: Curses library
Date: 22 Mar 92 14:42:27 GMT
Reply-To: u31b3hs@messua.informatik.rwth-aachen.de (Michael Haardt)
I currently have two curses libraries, BSD curses which came with Linux
and a PD curses package, which has its origin in PC curses and ncurses.
I think BSD curses is awful old and 7 bit characters are a pain. On the
other side, it is standard.
The PD curses has a few minor bugs, but I used it under MINIX for example
to compile sokoban, rogue and a few other games. The weak points are the
print/scan functions (no ANSI way to pass arguments) and such things. Its
features are support of ACS characters like SYS V and 8 bit characters with
standout, underline, bold etc.
If there is no other curses library, I will work more on this PD curses,
because I really need a curses lib. Any comments?
Michael
------------------------------
From: hohndel@winx03.informatik.uni-wuerzburg.dbp.de (Dirk Hohndel)
Subject: Keymapping under kermit ?
Date: Mon, 23 Mar 1992 09:43:59 GMT
Hi Linuxers,
I've got a problem with the keymapping when using kermit to connect from my
Linux PC to our SUNs :
How can I remap a key ? I know it is something like set key .. .. when using
mskermit under MS-DOOF, but that doesn't work with Linux :-(
Is it possible ? How ?
Dirk
--
hohndel@informatik.uni-wuerzburg.de (Dirk Hohndel)
Lehrstuhl Informatik I, Universitaet Wuerzburg, Am Hubland, D-8720 Wuerzburg
------------------------------
From: PSYXSGP@otago.ac.nz
Subject: (none)
Reply-To: PSYXSGP@otago.ac.nz
Date: Mon, 23 Mar 1992 11:44:55 GMT
If and when x11 is ported to Linux .. what will this mean.
From my reading of the faq s in comp.windows.x the 2 major intefaces
Motif and OPEN LOOK appear to both be commercial products.
(I may be wrong!)
If they are then won't X be a bit useless without them?
Second question (and here I really display my ignorance)
how much work is involved in porting X? I suppose the OS will have
to be a bit more stable than it is now before work can really begin.
Could someone estimate a time frame ( always a dangerous
thing to do .. and a bit cheeky considering it is all pretty well
cost free!).
------------------------------
From: Peter Desnoyers <peterd@merlin.dev.cdx.mot.com>
Subject: Linux-Activists Digest #137
Reply-To: peterd@merlin.dev.cdx.mot.com
Date: Mon, 23 Mar 1992 13:14:34 GMT
[battle@cs.utk.edu (David Battle) proposes implementing TCP/IP in a
user-space process]
|From: mo@gizmo.bellcore.com (Michael O'Dell)
|Subject: Re: TCP/IP in user-mode
|Date: 22 Mar 92 19:33:27 GMT
|Reply-To: mo@bellcore.com (Michael O'Dell)
|
|Yeah, it's "elegant" and aside from most likely being slow as death,
|it should work fine.
|
|Networking goes in the kernel if you want it fast.
|If you don't care how fast it goes or how many resources it uses doing it,
|feel free to put it wherever you want.
|This is the voice of experience.
|
| -Mike O'Dell
| Resident crank
Kernel support for networking doesn't have to be all-or-nothing. A
number of people (Clark, for one) have suggested putting just
multiplexing facilities in the kernel, and then doing the rest as a
library within the destination user process. For TCP/IP this would
mean putting all of IP and the connection control block lookup for TCP
into the kernel.
Some advantages are:
- performance - you don't have to go kernel -> user process 1 -> user
process 2
- debugging is easier than full kernel implementation, although not
as easy as full user process method.
- "standard" TCP, RPC, etc. be supplied in shared library, but
applications would be able to use custom tuned implementations as
well. (see IEEE Network 1/92, article on RPC performance, for a
good example of why you might want this)
Just a thought...
Peter Desnoyers
------------------------------
From: stephen@ziggy.STUDENT.CWRU.Edu (Stephen Trier)
Subject: Re: Free BSD release: future of Minix/Linux?
Date: Mon, 23 Mar 92 13:18:58 GMT
Well, so much for my attempt to be a know-it-all. In article
<1992Mar23.034216.1001@usenet.ins.cwru.edu> I wrote:
>BSD386, when released, will cost $1000 with source code. 386bsd will be
>free.
I have now been informed that the beta BSD386 is currently priced at
US$1000 with source, but its price will go up later. If you really want
to know, contact BSDI or read comp.unix.bsd. I'll go keep my mouth shut
now. ;-)
--
Stephen Trier - Mach Hacker-in-training - stephen@ziggy.student.cwru.edu
"The attempt...to show that Nature does nothing in vain (that is to say,
nothing which is not profitable to man) seems to end in showing that
Nature, the gods, and man are alike mad." -- Benedict de Spinoza
------------------------------
From: abs@sunvax.sun.ac.za
Subject: Wanted. UUCP compatible mailer for Linux.
Date: 23 Mar 92 10:37:03 +0200
Hi Linuxers,
Iam looking for a mailer for linux. It must able to do UUCP and also allow
users on the linux machine to send mail to each other. Has anyone got this
working ? Please mail me.
BTW anonymous UUCP feature will be nice as well.
Thanks and regards
Andre B. Skarzynski
abs@sunvax.sun.ac.za
------------------------------
From: joel@wam.umd.edu (Joel M. Hoffman)
Subject: HD timeout [was Re: Linux 0.95a installation]
Date: Mon, 23 Mar 1992 14:32:48 GMT
In article <9203212659@gandalf.informatik.rwth-aachen.de> u31b3hs@messua.informatik.rwth-aachen.de (Michael Haardt) writes:
>When I copy *really* big files, like creating the swap file, I sometimes
>get "HD timeout" messages. Is this a warning or an error (read: can I
>ignore it)?
Some bizarre is going on with the HD timeouts, I think. I get the
same problem running GCC (DJGPP) under DOS, which is also 32bit.
Every so often, I get a ``not ready error reading drive C:'' Of
course, that absurd. I also get the same error when running some
programs compiled with GCC (DEMACS, e.g.). BTW, my machine is a Dell
325D ('385 at 25MHz), with a Seagate (yuk) 110ish HD (IDE).
The bottom line is there's no reason the HD should timeout, or be not
ready. I've spoken to Dell repeatedly about this matter. The first
time they told me the software was as fault. The second time they
told me there was a problem in their BIOS. When they sent me the new
BIOS, the problem changed its manifestation, but was not repaired.
(Now whenever I get the error, the HD jumps to track 470 and stays
there.) I spoke with tech support yesterday (but to my great surprise
got someone who didn't speak enough English to communicate. What's
happening to Dell?) He told me that one of their super-gurus said
it was either a problem on the motherboard or a problem with the HD.
They will replace the motherboard me.
Whatever the case, something screwy is going on. I suspect problems
of a similar nature may be plaguing Linux.
Comments? Suggestions?
-Joel
(joel@wam.umd.edu)
------------------------------
Crossposted-To: comp.os.minix
From: lpickup@xanadu.btv.ibm.com (Lance Pickup)
Subject: Re: Free BSD release: future of Minix/Linux?
Date: Mon, 23 Mar 92 14:12:14 GMT
Reply-To: lpickup@vnet.ibm.com
In article <1992Mar21.160452.28644@eecs.nwu.edu>, kaufman@eecs.nwu.edu (Michael L. Kaufman) writes:
|> Maybe not. If I get some free time, I may do a PS/2 port. I have already ported
|> another Unix to the PS/2, so it shouldn't be too big a deal.
|>
|> Michael
Michael,
What port are you thinking of doing, and what one have you already done?
--
...Lance (Model 70 owner looking for UNIX that'll run on my box)
=======================================================================
Lance Pickup VNET/IBM Internet: lpickup@btv
IBM/Vendor Systems Internet: lpickup@vnet.ibm.com
Technology Products
Burlington, VT Phone: (802) 769-7104 (tie 446)
=======================================================================
------------------------------
From: lewis@bnrmtl.bnr.ca (Pierre LEWis)
Subject: Re: Free BSD release: future of Minix/Linux?
Date: Mon, 23 Mar 92 14:47:27 GMT
In article <1992Mar23.034216.1001@usenet.ins.cwru.edu>
stephen@ziggy.STUDENT.CWRU.Edu (Stephen Trier) writes:
> I think it's clarification time. There are two different BSDs for the 386.
> One is commercial, one isn't. The commercial one is named BSD386, bsd386,
> or bsd/386, depending on who you ask. The just-released free one is named
> 386bsd.
Thanks, I was a bit confused. One question: is there any relation between
the two (beyond the flavor)?
THanks,
--
Pierre LEWIS +1 514 765-8207 ESN 852-8207
Internet: lew@bnr.ca Corwan: lewis@bmtls30
Implementing standards is like walking on water -- both work best when frozen.
------------------------------
From: snowden@cray.com (Jim Snowden)
Subject: Booting Linux as an OS/2 MOST system
Date: 23 Mar 92 14:01:22 GMT
A few days ago someone asked if ti was possible to install Linux as
a bootable parition using the MOST system supplied with OS/2.
Well, thanks to Iain (eonu24@castle.edinburgh.ac.uk), I got the
Linux boot image written to a suitable place on the Linux
hard disk partition. The partition can be added to the list of bootable
partitions using the OS/2 fdisk utility.
SO now when I boot my system, I am presented with a list as follows:
DR DOS
UNIX
OS/2
and I can just move the cursor to the one I want to boot.
I haven't setthe Linux partition to be the default boot (yet...!)
Jim Snowden
snowden@hemlock.cray.com
------------------------------
From: david@ods.com (David Engel)
Subject: Re: diff for extended partitions
Date: 23 Mar 92 15:19:10 GMT
hedrick@dartagnan.rutgers.edu (Charles Hedrick) writes:
: The following diff appears to fix extended partitions. At least on my
: system I get the right list of partitions, and making a file system on
: one causes no visible damage. Note that I've added code to print the
It seems to work for my one extended partition with one logical drive.
Using mtools, I can read and write my DOS drive d:.
-David
------------------------------
From: david@ods.com (David Engel)
Subject: Re: patches
Date: Mon, 23 Mar 1992 15:33:43 GMT
johnsonm@stolaf.edu (Michael K. Johnson) writes:
: To everyone posting patches --
:
: I am pulling all patches off the newsgroup, so don't worry about
: uploading them anywhere. However, make sure they are at least a
: little documented so that I can say something intelligent in the
: patch.doc file at tsx-11, though. So far everyone has, but I just
: thought I'd mention that.
This brings up something I've been meaning to ask. Are you (or anyone
else) planning on keeping a set of "recommended" kernel patches, and
possibly even fully patched set of sources. I'm mostly thinking about
bug-fixes here but we might want to also include an alternate set of
"expirmental" patches (and source) with the caveat that they may or
may not be included (at Linus' discretion) in the next release.
What I'm hoping to avoid is the "patch anxiety" experienced by users
like me who get a bit antsy when their kernel reaches a certain level
of patches. I remember when the patches to 0.12 started to overlap.
They appeared to work but I didn't feel very secure.
-David
------------------------------
From: aclark@netcom.com (Al Clark)
Subject: Re: Bad kernel build (0.95a) with GCC 1.40! Why????
Date: 23 Mar 92 15:40:17 GMT
In article <1992Mar23.083655.2661@socrates.umd.edu> cm445a17@socrates.umd.edu (cm445a17) writes:
>Today I ventured into build the kernel for Linuc 0.95a with
>GCC 1.40. I edit some of the makefiles so that the -fcombine-regs
>flag was set and I did the make. Everything compiled fine so I decided
>to try it out. I booted up with the new image and everything seems to
>work except for the fact that some of the keyboard keys produce the
>wrong characters (i.e. pressing / produces -). Does anyone know why
>this is happening? Am I supposed to be using GCC 2.0? If so, what
>is the status of GCC 2.0? (Alpha, Beta???), and can it be used reliably
>with Linux? Thanks,
>
>Desperate Jaime
>
Check in keyboard.s - looks like you have not specified US keyboard;
early kernel sources have specified Finnish keyboard.
--
Al - aclark@netcom.com - My opinions are my own.
*** Practice random kindness and senseless acts of beauty! ***
------------------------------
From: abc@banjo.concert.net (Alan B Clegg)
Subject: Re: 386 BSD and linux
Date: 23 Mar 92 13:59:54 GMT
In article <zhao.701144787@chilko.ucs.ubc.ca> zhao@unixg.ubc.ca (Jiansheng Zhao) writes:
>
> I certainly will stay with linux and I am happy with linux 0.95a, thanks
>to Linus and those involved.
Me too.
> I tried 386BSD boot-floppy on three different 386 machines, for two of
>them, drive A: was being read forever, the third one boot and panic
>and automatically re-boot, ...... .
I actually got 386BSD up and running. It is a *MESS*. People have compared
it to linux a year ago. Perhaps. Perhaps even worse. I recommend that
anyone interested in 386BSD test it (yeah, blow away your hard drive, see if
I care), then watch it fail, "extra interrupt", hang, core dump, etc etc.
I am sure that in time, it will develop into a "good thing", but right now..
NOT!
My thoughts: Get the WD8003 code out of 386BSD, get TCP/IP working. Linux
beats the pants off of 386BSD at this point.
> Also, I like bash under linux (alias, history etc.) and virtual console.
Bravo.
Long Live Linus and Linux.
-abc
--
abc@concert.net Alan Clegg - Network Programmer
KD4JML (just my luck!) MCNC -- Center for Communications
------------------------------
From: poe@daimi.aau.dk (Peter Orbaek)
Subject: Bug in putpwent()
Date: 23 Mar 92 15:59:32 GMT
The other day I was building a passwd utility for inclusion in my next
release of the init/getty/login package (poeigl). During this I discovered
that the putpwent() routine in libc was wrong. It fails to write the pw_dir
field.
After I fixed this it seems to work fine.
When are we going to see sources for an updated 0.95a libc? I haven't seen
sources for libc since 0.12 went out.
- Peter.
--
Peter Orbaek ----------------- poe@daimi.aau.dk | ///
Hasle Ringvej 122, DK-8200 Aarhus N, DENMARK | ///
| \\\///
"Strong typing is for people with weak memories" | \XX/
------------------------------
From: heim@frieda.informatik.uni-tuebingen.de (Gerald Heim)
Subject: minor linux 0.95a bugs
Date: 23 Mar 92 16:43:50
Hi Linuxers,
While playing around with the Linux0.95a images I found a few minor bugs.
- At boot-time, when I am asked if I wish to select any VGA mode or
just press any key to stay at 80x25, I noticed that some keys will
keep the system from reading the root-image (i.e. the arrow-down key).
BTW: In the upper right corner appeared EGAm instead of EGAc...
- The backspace-problem appears even on a virtual console where backspace
worked at first when I log out and try to log in again.
- Has anybody tried to fill up the process table yet? O.K. this is what
I tried: I started about 55 shells from the same virtual console :-)
The last shell I started neither gave me a prompt nor printed any
error message. Pressing ScrollLock showed that I had 58 processes.
Switching to another virtual console where I also was logged in I still
could start processes, but after a few seconds I got error messages
like memory fault or out of memory or out of swap space (certainly,
running from diskettes, I didn't have swapping enabled).
Maybe I just ran out of memory but I think even then, this behaviour
is, if not a bug, a little bit strange.
BTW: We should get a /bin/kill in the root image :-}
- One more bug: If I press ScrollLock and stay on the key for a few seconds,
I get a few dumps from the process table, then the kernel hangs and
prints:
general protection: 0014
EIP : 0008:0000c166
EFLAGS: 00010086
fs: 0010
base: 0076ffff, limit: 00000018
Pid: 0, process nr: 0
8e e0 ff 05 b0 a1 02 00 b0 20
Kernel Panic: free_page_tables called with wrong alignment
In swapper task - not syncing
Looks like a real bug, doesn't it?
I was running Linux on a 386sx 20Mhz board with 32kB cache and 4 Mb RAM.
Linux is really great, thank you Linus and all the others!
ciao
gerald
email: heim@peanuts.informatik.uni-tuebingen.de
------------------------------
** FOR YOUR REFERENCE **
The service address, to which questions about the list itself and requests
to be added to or deleted from it should be directed, is:
Internet: Linux-Activists-Request@NEWS-DIGESTS.MIT.EDU
You can send mail to the entire list (and alt.os.linux) via:
Internet: Linux-Activists@NEWS-DIGESTS.MIT.EDU
Linux may be obtained via one of these FTP sites:
nic.funet.fi pub/OS/Linux
tsx-11.mit.edu pub/linux
tupac-amaru.informatik.rwth-aachen.de pub/msdos/replace
The current version of Linux is 0.12, released on Jan 14, 1992
End of Linux-Activists Digest
******************************