Subject: Linux-Development Digest #783
From: Digestifier <Linux-Development-Request@senator-bedfellow.MIT.EDU>
To: Linux-Development@senator-bedfellow.MIT.EDU
Reply-To: Linux-Development@senator-bedfellow.MIT.EDU
Date:     Wed, 1 Jun 94 11:13:06 EDT

Linux-Development Digest #783, Volume #1          Wed, 1 Jun 94 11:13:06 EDT

Contents:
  Re: HELP! 20MB of mem and Linux 1.0 --swapper problems (Jonathan Kitchin)
  SCSI DTC3292 Problems. (John Kirkland)
  Install Glitch Kernel Panic? (Burton Bicksler)
  Re: Networking and Linux (Andy Beal)
  ncurses 1.8.1 (centauri)
  Re: Regarding the rlogin security hole (Nick Andrew)
  Re: HELP! 20MB of mem and Linux 1.0 --swapper problems (Wolfgang Kalthoff)
  Recompiling S3 server (Howard P. Henson)
  Kernels and Debuggers (Stephen Collyer)
  Re: kernel 1.1.14 (&13) + xiafs segfaults rpc.nfsd (Rene COUGNENC)
  Re: InterLink for Linux (GLAUDE DAVID)
  Re: HELP! 20MB of mem and Linux 1.0 --swapper problems (James Logajan)
  Re: SCSI Problems with 1.1.15 and 1.1.16 (Alan D. Peckham)
  NFS hangs system 1.1.16 & 1.1.17 (Alan D. Peckham)
  Re: SCSI Problems with 1.1.15 and 1.1.16 (Stephen Beaton-Snook)
  Re: How to print faster (Nigel Gamble)
  Writing a new file system, help (Ben Frank)
  Re: Three-button mouse, again? (Steve Kann)

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

From: jon@perth.DIALix.oz.au (Jonathan Kitchin)
Subject: Re: HELP! 20MB of mem and Linux 1.0 --swapper problems
Date: 1 Jun 1994 08:35:06 +0800

In <2sfgge$k3i@constellation.ecn.uoknor.edu> parprods@ecn.uoknor.edu (Dorwin Shields) writes:

>   HELP!!--I can hardly get Linux to boot now--I get swapper problems
>during boot up--the messages say that a nonexistant swap page was 
>trying to be duplicated--then later I get Oops:0002 and the message
>says that it is unable to fulfill a kernel paging request.  I cant
>rebuild my kernel--I tried to reinstall from scratch from the trans-
>ameritech CD but I started getting segmentation faults during installation.
>     All of this started when I took out 4 1MB simms and put 4 4MB simms in.
>I now have 20MB--4 1MB simms(3 chip) and 4 4MB(9 chip)--Can anyone help?
>Thanks,
i>Dorwii
You could try reversing the positions of the 1mb and the 4mb SIMMs
jon kitchin jon2dialix.oz.au
n

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

From: johnkirk@utxvms.cc.utexas.edu (John Kirkland)
Subject: SCSI DTC3292 Problems.
Date: 31 May 94 14:50:35 CST

I have been trying to install Linux (Slackware 1.2.0) on a 486dx50 at the University of Texas at Brownsville for about a week
now...  

It has been giving me lots of problems with the file systems and the scsi controller.  I am using a DTC3292 SCSI Host Adapter with
two Micropolis 1.2GB  Hard Drives.

Has anyone successfully used this scsi controller, and if so, what modifications, if any, had to be made in the bios setup.

Thank you,
John Kirkland

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

From: bbick@netcom.com (Burton Bicksler)
Subject: Install Glitch Kernel Panic?
Date: Wed, 1 Jun 1994 01:18:02 GMT

Hi, 
I'm just getting back into the world of Unix clones to do some 
experimenting.  Just got the Yggdrasil Summer 94 (1.0) Linux release and 
have encountered a problem during installation.  I had seen a message 
from someone else in another newsgroup who was having the same problme 
with, I think, the Slackware release.  They had 16 Megs of RAM and a 
newer system so that eliminated one of my suspects.

System configuration:
Old Micronics 20Mhz 386 Motherboard with double-sigma 386 CPU and 
Phonenix BIOS vers 1.10.0, MFM Western Digital Controller, Serial/Parr 
card, FD 850 SCSI HAB and Chinon DoubleSpeed CDROM.  This system has run 
Coherent with no problems and has no problems with DOS/Windows either, 
also can run Various protected mode DOS apps with no problems.

Even when stripped down to just hard / floppy controller, no serial/par 
or SCSI I get the same problem.

After the calibrating delay loop ... ok message ( of after the SCSI 
recognition stuff if the SCSI HAB is installed) I get a General 
Protection: 0000 error with the trap dump.
Then
Process swapper (pid:0 ..... etc)
...
Task[0] (swapper) killed: unable to recover
Kernel Panic: Trying to free up swapper memory space
In Swapper Task - Not Syncing
we are now hung (well not totally, screen blanker will kick in and the 
three fingered salute will reboot).

This is an old AT clone machine that I wanted to use to test Linux before 
commiting a full installation on my main system.
Since someone else reported a similar problem with a newer machine with 
16 Megs of RAM and with a different distribution of Linux 1.0 I suspect 
that this is a common problem.  While it may be a 'RAM' problem I doubt 
it since nothing else is failing on this same machine.

When I attempt to boot Linux on a machine at work with the Phoniex BIOS 
it makes it all the way to the complaint about not being able to mount 
root, which is ok since I don't have a CDROM on that system.  That box 
has 16 Megs of RAM on it.  I think, based on what I see on that system, 
that the GP Fault is happening about the time that Linux is testing for 
sound cards, at least the messages that pop up after calibrating delay 
loop.. are related to sound.

Do you have any suggestions on what to look for?  Oh yeah, I have an ATI 
VGA Wonder for the display card.

Thanks in advance,
Burt

-- 
                                             bbick@netcom.com

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

From: bealar@ndlc.occ.uky.edu (Andy Beal)
Subject: Re: Networking and Linux
Date: 31 May 1994 15:27:22 -0500

Scott S Critchley (scottc@alpha2.csd.uwm.edu) wrote:

: but the other day, the thought just occurred to me....  Couldn't it be
: possible to hook up nodes to a linux-based server just through plain
: everyday old serial connections....  I mean, I know that serial would
: be slower, but if all you want to do is text processing then 19,200
: baud would be plenty fast enough..  And more than that, but couldn't
: you then use ANY computer that had a serial connection (like, for
: example an Apple //GS or //e with a serial board, or a Mac plus or a 
: crappy old 286 or XT or whatever) and terminal software....  The reason
: why I am asking this is because I have several old apple //s sitting around
: collecting dust, and the thought of using those apple //s as network nodes
: is quite appealing (I mean, to access my 486 though an apple //e in a
: completely different room would be quite a feat)...  Another possibility
: could be school systems with apple //s.  The apple // is a worthless
: computer now-a-days but hook them up to a 486 and you could teach
: a programming class, or perhaps access the internet or whatever, all 
: though an old apple //.  But It could only happen if you could use 
: serial (try getting an ethernet board for an apple //e)

If this happens, I want to know about it.  I am thinking of doing the 
same, except with an Atari XE, and an Acorn Electron/BBC.  Both are 8 bit 
w/ 64kb of ram, and boot to Basic, so I can write serial interfaces for 
them.  Only blockade right now is my hard disk crashed, and I sent it 
back.  So for a few weeks, Im running a 40megger, instead of a 400.


-- 
                /^\      /^\               
___________/\  /   \    /   \  /\________  Andy Beal
             \/     \  /     \/            bealar@ndlc.occ.uky.edu
                     \/

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

From: centauri@cs.tu-berlin.de (centauri)
Subject: ncurses 1.8.1
Date: 1 Jun 1994 09:58:50 GMT

Hi,

I'm having some problems with ncurses.

We have programed a small user interface with maskfields, dropdownmenus and
listboxes with SCO and its curses. 

I tried to port it to Linux, and it compiles fine with no errors or warnings,
but:

- The menus arent shown correctly. It seames as if the first ESC-sequence is not
interpreted by the terminal. So at the beginning of the menubar something like 
'[[D' is shown. Same goes to the dropdownmenus. There is a '[[A', an the first
entry is missing.

- The last line of a listbox is not shown, instead the first ist missing. So I
get something like

        +---------------------------+
        | Line 2                    |
        | Line 3                    |
        | Line 4                    | 
        |                           |
        +---------------------------+

There should be line 1-4.

To remove the dropdowns, I copied their background in a second window, before
drawing them. Well ... To redraw the background, I try to copy the background
back, but curses doesn't do anything, though I do a wrefresh.

This problem looks to me as if the terminal doesn't understand the ESC-sequences.
Any idea what I must change ? (ncurses-version, termcap, terminfo ?)

cu Christian

==============================================================================
Christian Schroeder, TU-Berlin, FB13  
InterNet : centauri@cs.tu-berlin.de
==============================================================================

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

From: nick@kralizec.zeta.org.au (Nick Andrew)
Subject: Re: Regarding the rlogin security hole
Date: 1 Jun 1994 20:17:44 +1000

In <CqA2AL.4LG@cichlid.com> aab@cichlid.com (Andy Burgess) writes:

>Are we vulnerable if we are using tcpwrapper to restrict rlogin and telnet
>access?

Yes - to all sites which are allowed to rlogin in ...

Nick.
-- 
Kralizec Dialup Unix (Public Access)    Data: +61-2-837-1183, 837-1868
Zeta Microcomputer Software             v.42bis v.32bis 14.4k 24 hours
P.O. Box 177, Riverstone NSW 2765       Plan: To beat Gnuchess 4.0 fairly!

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

From: wo@rio70.bln.sni.de (Wolfgang Kalthoff)
Subject: Re: HELP! 20MB of mem and Linux 1.0 --swapper problems
Date: 1 Jun 1994 12:01:21 +0200
Reply-To: wo@rio70.bln.sni.de (Wolfgang Kalthoff)

In <2sfgge$k3i@constellation.ecn.uoknor.edu> parprods@ecn.uoknor.edu (Dorwin Shields) writes:

>   HELP!!--I can hardly get Linux to boot now--I get swapper problems
>during boot up--the messages say that a nonexistant swap page was 
>trying to be duplicated--then later I get Oops:0002 and the message
>says that it is unable to fulfill a kernel paging request.  I cant
>rebuild my kernel--I tried to reinstall from scratch from the trans-
>ameritech CD but I started getting segmentation faults during installation.
>     All of this started when I took out 4 1MB simms and put 4 4MB simms in.
>I now have 20MB--4 1MB simms(3 chip) and 4 4MB(9 chip)--Can anyone help?
>Thanks,
>Dorwin

first try : run only with the 4MB chips - should work 
after that: check the motherboard's manual : it might support only
            4*1MB, 8*1MB, 4*4MB, 8*4Mb 
            if it does support 4*4Mb + 4*1Mb
            make a kernel that supports > 16Mb and add the 1MB SIMMs

hope this helps
wolfgang

-- 
Wolfgang Kalthoff    | S iemens             | email: wo@rio70.bln.sni.de
Gustav-Meyer-Allee 1 | N ixdorf             | 
D-13355 Berlin       | I nformation Systems | Tel: +49-30-4673-2951 Fax: 2915
=============================================================================

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

From: hhenson@inyanga.wits.ac.za (Howard P. Henson)
Subject: Recompiling S3 server
Date: 1 Jun 1994 14:15:22 GMT

I tried to recompile my S3 server to include the PEX extentions but all I 
managed was to get a whole bunch of missing file errors, there seems to be a 
need to compile font servers as well of which I don't seem to have any code 
for. I am using the Slackware 1.2.0 distribution.

If anyone has recompiled their server successfully I would love to hear what I 
lack yet.

Thanks
Howard

--
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
~                   Rotating the object by                                 ~
~          Howard Henson <hhenson@inyanga.cs.wits.ac.za>                   ~
~           mans quest for object orientaion ends here!                    ~
~                                                                          ~
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~


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

From: stephen@dogmatix.inmos.co.uk (Stephen Collyer)
Subject: Kernels and Debuggers
Date: Wed, 1 Jun 1994 10:46:42 GMT

Is it possible to compile the kernel in some kind of debugging mode and 
run it under a debugger ? I'd like to do this to help me get to grips with
the kernel code. 

I'd appreciate any tips/hints/docs that would help me to do this. What I'd 
like to do specifically is:

1) compile the kernel in debugging mode
2) compile a user program in debugging mode
3) Using the UPS debugger if possible, do all the funky source level stuff
   one can do 
4) Hit a syscall, and jump into the kernel 
5) Do all the funky source level stuff one can do, but in the kernel this time,
   so I can see what gets called and when
6) exit from the kernel and return to user mode and carry on debugging

Is this possible ? Am I crazy ?

Thanks in advance.

Steve Collyer.

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

From: rene@renux.frmug.fr.net (Rene COUGNENC)
Subject: Re: kernel 1.1.14 (&13) + xiafs segfaults rpc.nfsd
Date: 31 May 1994 09:35:41 GMT
Reply-To: cougnenc@hsc.fr.net (Rene COUGNENC)

Ce brave Jan Niehusmann ecrit:

> Rene COUGNENC (rene@renux.frmug.fr.net) wrote:

>  > With 1.1.14 or 1.1.13:
>  >    On my 386 using xiafs, rpc.nfsd dies at any time.I can't keep
>  >    it running as soon as some operations like readdir are done.
>  >    (ls, or shell filename completion, etc).

> On my 386, using e2fs: The same problems (with 1.1.16). With 1.1.9 
> everything works fine. 
> So it shouldn't be a problem of xiafs.

Right, I received mails from several persons experiencing the same
problem, as of 1.1.13, rpc.nfsd is killed by sig 11 on 386 processors.
(and not on 486).

I have found one binary working fine, coming from  Slackware. But
we are unable to compile a working version from nfs-server 2.0 with
the current compiler and libs..
--
 linux linux linux linux -[ cougnenc@renux.frmug.fr.net ]- linux linux linux 

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

From: dglaude@vub.ac.be (GLAUDE DAVID)
Subject: Re: InterLink for Linux
Date: 31 May 1994 20:52:32 GMT

Rob Janssen (rob@pe1chl.ampr.org) wrote:
: In <2s8ju0$6l1@rc1.vub.ac.be> dglaude@vub.ac.be (GLAUDE DAVID) writes:
: >Of course all of this without Ethernet card ;-) Only with a Parallel or
: > Serial Cable.

: You must be a masochist... :-)
: Of course it is possible, but don't expect much speed.  Buy an ethernet card.

You mean *TWO* Ethernet card + a cable. It cost a bit more. ;-)

: Rob

PS: Thank you all for your respond.

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

From: jamesl@netcom.com (James Logajan)
Subject: Re: HELP! 20MB of mem and Linux 1.0 --swapper problems
Date: Tue, 31 May 1994 20:43:50 GMT

Dorwin Shields (parprods@ecn.uoknor.edu) wrote:
:    HELP!!--I can hardly get Linux to boot now--I get swapper problems
: during boot up--the messages say that a nonexistant swap page was 
: trying to be duplicated--then later I get Oops:0002 and the message
: says that it is unable to fulfill a kernel paging request.  I cant
: rebuild my kernel--I tried to reinstall from scratch from the trans-
: ameritech CD but I started getting segmentation faults during installation.
:      All of this started when I took out 4 1MB simms and put 4 4MB simms in.
: I now have 20MB--4 1MB simms(3 chip) and 4 4MB(9 chip)--Can anyone help?
: Thanks,
: Dorwin
Try pulling the 4 1MB simms and run with only 16MB. (Be sure to put the 16MB
in the correct bank). You are most likely getting memory problems. Those
3 chip simms don't coexist very well with 9 chip simms. I even had problems
with 9 chip 1MB simms coexisting with 9 chip 4MB simms. Good luck.

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

From: peckham@drei.enet.dec.com (Alan D. Peckham)
Subject: Re: SCSI Problems with 1.1.15 and 1.1.16
Date: 1 Jun 1994 13:25:16 GMT
Reply-To: peckham@drei.enet.dec.com (Alan D. Peckham)


I ran into the same problem with 1.1.16. Ultrastor 14F & 2x1GB drive.
I backed off to 1.0.9 and e2fsck found 2 shared blocks on the second drive.

Alan D. Peckham
peckham@drei.enet.dec.com

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

From: peckham@drei.enet.dec.com (Alan D. Peckham)
Subject: NFS hangs system 1.1.16 & 1.1.17
Date: 1 Jun 1994 13:37:38 GMT
Reply-To: peckham@drei.enet.dec.com (Alan D. Peckham)


I switched to 1.1 in hopes of taking advantage of fragmentation /
reassembly code in the new kernel on my comm/router/fileserver box.
This box is presently a small headless 386/40 4MB memory and
2 x 1GB disks in my basement. I nfs-mount filesystems to my 486/50
in the office. When I try to rebuild the kernel on a mounted
partition, it hangs both boxes - my suspicion is a memory deadlock
of the fileserver box. NFS guru! Could this happen with the new
8k sizes? Any way I can help track this down without dropping money
for another screen or more memory?

Alan Peckham
peckham@drei.enet.dec.com

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

From: steve1@morgan.ucs.mun.ca (Stephen Beaton-Snook)
Subject: Re: SCSI Problems with 1.1.15 and 1.1.16
Date: 1 Jun 1994 12:25:38 GMT

Kang-Jin Lee (lee@tengu.in-berlin.de) wrote:
: I had no problems with 1.1.15 and 1.1.16 but today, after a crash (caused by
: me experimenting with suid programms), the kernels boot okay, starts fsck
: and checks the filesystems but hangs during fsck with the following message:

: scsi0 : reseting for second half of retries
: Unable to reset scsi host 0 - Probably a SCSI bus hang

I am getting the same problem while booting when it checks for SCSI devices
and the same error message, my system setup is identical. I had to move
back to my last available kernel 1.1.11. It seems to run into problems
when it look at my SCSI I drive but not my SCSI II.

: 486/33 ISA, AHA1542B, ext2-fs.

: -- 
: Kang-Jin Lee
: lee@tengu.in-berlin.de

--

 Stephen Beaton-Snook                             Memorial University of NFLD
 steve1@morgan.ucs.mun.ca                         St John's, Newfoundland
                                                  Canada.


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

From: nigel@gate.net (Nigel Gamble)
Subject: Re: How to print faster
Date: Wed, 1 Jun 1994 12:22:32 GMT
Reply-To: nigel@gate.net (Nigel Gamble)

In <CqII9r.47F@pe1chl.ampr.org> rob@pe1chl.ampr.org (Rob Janssen) writes:
>In <2s6j6h$rf9@news1.svc.portal.com> foulds@shell.portal.com (David - Foulds) writes:
>>rob@pe1chl.ampr.org (Rob Janssen) writes:
>>>In <2s353s$c06@news1.svc.portal.com> foulds@shell.portal.com (David - Foulds) writes:
>>>>The printer driver lp.c seems to need a longer busy-loop,
>>>>at least on my machine (486-SX33) and with my printer
>>>>(Panasonic KX-P4410 laser with 2.5MB).  The busy loop
>>>>that waits to send the next char in lp_char_interrupt
>>>>only counts to 3 before giving up.  I increases the count
>>>>to 30 or so (and also increased LP_BUFFER_SIZE to 2048)
>>>>and reduced the time to print ghostscript 200K raster
>>>>files (or whatever you call them) i.e. images of postscript
>>>>pages after having been ghostscripted from 4 minutes to 50 seconds.
>>>>This was while crunching the postscript in another process.
>>>>On an empty machine the time per page went from 80 seconds 
>>>>to 40, cpu utilization from 90% to 30% or so.

>>>It looks like you are talking about a different version than is
>>>in the kernel now.
>>>Anyway, the current version allows tuning of these parameters to the
>>>requirements of your printer and CPU speed using the program "tunelp",
>>>without modifying lp.c

>>"tunelp" does indeed tune some parameters (by the way, the most
>>recent version on sunsite is from 3 July 1993, so I am up to
>>date this time) but it doesn't tune the parameter I am talking
>>about.  The spin loop I refer to is at the top of the function
>>lp_char_interrupt and doesn't look like a loop; it has the
>>form (a || a || a) where a is a condition based on the printer
>>status.

>Interesting...  apparently the printer delays quite a while before it
>releases busy after sending the ACK.
>The PC parallel port interrupt design is braindead, in that it sends an
>interrupt before the device is ready for more data.

It's up to the printer to indicate when it is ready for the next
data byte to be strobed in by asserting ACK and dropping BUSY.  It is
the ACK that causes the interrupt, but the fact that this comes before
the printer drops busy is not supposed to make any difference - by
asserting ACK, the printer has declared itself ready to receive the
next character.  It's the particular printer design that's brain-dead
if it's not behaving like this.

The reason for the "busy loop" in lp_char_interrupt is to avoid the
overhead of sleeping while waiting for the next interrupt if the
printer is almost immediately ready for the next character.  I found
that three tries was optimum for my HP LaserJet IIIP, but David's
Panasonic obviously needs a little longer than this to avoid
unnecessary sleeps.  Perhaps this is a candidate for a tuneable
parameter.

Cheers,
Nigel
--
Nigel Gamble                                    nigel@gate.net
Boca Raton, FL, USA.

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

From: u9219765@sys.uea.ac.uk (Ben Frank)
Subject: Writing a new file system, help
Date: Tue, 31 May 1994 21:42:51 GMT

Hi,
   I'm trying to write a new file system (Third year porject) I seem to
remember that someone was writing a "HowTo write a file system" document,
I can't find it anywhere, nor any reference to it. Does it exist, if not is
there anything that would be useful to look at ? I have read/am reading all
the stuff from the LDP
Cheers.

        Ben Frank.

--
If I hadn't seen such riches I could live with  | Ben Frank
being poor  :)                                  | u9219765@sys.uea.ac.uk


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

From: stevek@panix.com (Steve Kann)
Subject: Re: Three-button mouse, again?
Date: 1 Jun 1994 00:22:05 -0400

Jim Breen (jwb@capek.rdt.monash.edu.au) wrote:
: s21008@cc.ntnu.edu.tw (DL83-08) writes:

: >Thank you to all who wrote about the dual-protocol mouse configuration for
: >X. Obviously, my mouse is a bit different from most. The DOS driver has two
: >modes: Microsoft and PC mouse. The microsoft mode uses two buttons, and the
: >PC mouse uses all three buttons. In my Xconfig file, I can specify the
: >Microsoft protocol, of course, and the mouseman protocol. The problem is
: >that both use only two buttons. All of the responses to my last request for
: >help told me to use mousesystems with the cleardtr statement. Well, I tried
: >that, and also tried the clearrts, as well. The problem is that microsystems
: >just does not work with this mouse. Is there any other way to force the
: >middle button to be recognized in mouseman mode?

: I had the same problem. I bought another mouse. Problem solved.

Ahh, the lazy way.

I took a couple of diodes, and hooked them up between the middle button,
and the two outer buttons, so the mouse itself thinks I am pressing both
outer buttons whenever I press the middle button.  Works great under
windoze as well, since windoze uses chorded buttons in place of a third
button as well..

Have to use diodes, though, of course, or else it will ALWAYS look like
you are pressing the outer two :).


--
- Steve

stevek@panix.com
stevek@magnum.cooper.edu



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


** FOR YOUR REFERENCE **

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

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

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

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

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

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