Subject: Linux-Development Digest #931
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, 20 Jul 94 14:13:14 EDT

Linux-Development Digest #931, Volume #1         Wed, 20 Jul 94 14:13:14 EDT

Contents:
  Re: URGENT: DIP - ASSIGNING IP's BASED ON TTY?? (Harald Milz)
  Re: New Linux kernels won't boot (James E. Carpenter)
  Re: New Linux kernels won't boot (James E. Carpenter)
  Motif 1.23 compilation (Basil P. Duval EPFL - CRPP 1015 Lausanne CH)
  if_arp.h in 1.0.9 and 1.1.31 (Albert Chin-A-Young)
  Re: gcc and Linux (compiling)... (Urs Thuermann)
  MGR? (Johan Rydberg)
  Re: MGR? (Harry C Pulley)
  Re: fact on linux vs sun (Don Garrett)
  Re: PORT TO ATARI (& AMIGA) (Michael Neuffer)
  re:Massisve GCC error (Bill Adams)
  Re: ping -f crashes (at least) linux-1.1.29 (Ron Smits)
  Re: Is XFree86 3.X from X11R6 ? (NetDog)
  Patchlevel 31 breaks UMSDOS (Sam Oscar Lantinga)
  Re: Kernel Hacker's Guide (Matt Welsh)
  Re: ping -f crashes (at least) linux-1.1.29 (Michael Bongartz)
  Re: ping -f crashes (at least) linux-1.1.29 (Michael Bongartz)
  Re: MAP_SHARED? (Carl Karlsson)

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

Crossposted-To: comp.os.linux.help
From: hm@seneca.ix.de (Harald Milz)
Subject: Re: URGENT: DIP - ASSIGNING IP's BASED ON TTY??
Date: Tue, 19 Jul 1994 18:46:40 GMT
Reply-To: hm@seneca.ix.de

Brian Kramer (bjkramer@pluto.njcc.com) wrote:
: > Urgent - I need a way to assign IP's based on the tty the caller logs in on.
: > Please help.

With sliplogin, you can run in slip.login whatever you want. I _think_
you should be able to run `tty` too.


-- 
Harald Milz                             office: hm@ix.de
iX Multiuser Multitasking Magazine      home:   hm@seneca.ix.de
Opinions are mine, not my employer's -- the answer is Forty-two


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

From: jimc@zachary.riva.com (James E. Carpenter)
Subject: Re: New Linux kernels won't boot
Date: 20 Jul 1994 01:25:44 -0400

In article <30fjt9$1ao@zachary.riva.com> jimc@zachary.riva.com (James E. Carpenter) writes:
>I have been using 1.0.8 for a while and finally decided to upgrade. So I went
>and got 1.1.23. When I got it to compile, it wouldn't boot. I just got 1.1.30.
>That won't boot either. I've NEVER had ANY problems booting Linux. So
>something changed between 1.0.8 and 1.1.23.
>
>When I boot the new kernel, it seems fine up to where it prints the kernel
>version. Then it just hangs. No partition check or anything. I have configured
>it all different ways, including disabling the scsi code. For hardware I have
>a 286 with that special chip that makes it a 386SX-16. It has been able to run
>Linux from the first days I started using it (whatever the version was in
>September 1992). It only has 4 megs of memory. I have an IDE and SCSI drive,
>but I tried 1.1.23 with a MFM and it still didn't work. Nothing has changed in
>the hardware, so its got to be the kernel.
>
>I know I have a crap machine here. And I am intending on upgrading ASAP. But
>it'll be a shame if the newer versions of Linux will only run on more advanced
>machines.

Update to my own problem...

The kernel IS loading and starting to boot! I didn't make that very clear,
sorry. The new kernel begins to boot, I see messages about it uncompressing,
the version of the serial driver, the network drivers, etc. When it prints out
the kernel version (Linux version 1.1.23 blah blah blah) everything STOPS! I
_think_ I have located the problem. It appears to be in move_to_user_mode().
This is located in include/asm/system.h. I'm working on it.


-- 
James E. Carpenter                           E-Mail: jimc@zachary.riva.com
23 Munroe Drive                                 Tel: (508) 643-0908
Plainville, MA  02762-1132

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

From: jimc@zachary.riva.com (James E. Carpenter)
Subject: Re: New Linux kernels won't boot
Date: 20 Jul 1994 02:58:40 -0400

In article <30icgo$bs@zachary.riva.com> jimc@zachary.riva.com (James E. Carpenter) writes:
>Update to my own problem...
>
>The kernel IS loading and starting to boot! I didn't make that very clear,
>sorry. The new kernel begins to boot, I see messages about it uncompressing,
>the version of the serial driver, the network drivers, etc. When it prints out
>the kernel version (Linux version 1.1.23 blah blah blah) everything STOPS! I
>_think_ I have located the problem. It appears to be in move_to_user_mode().
>This is located in include/asm/system.h. I'm working on it.

move_to_user_mode() is what is making my machine hang. However, 1.1.23's
include/asm/system.h, where move_to_user_mode() is located, is identical to
the system.h found in 1.0.8, which works perfect for me. So the way I see it,
move_to_user_mode() is hanging my machine but is NOT the problem, seeing that
the same code works for me in 1.0.8. Is there some kernel hacker out there
that has any idea what is going on here??? I barely know C or assembler, so I
can't really rip apart the kernel myself.

Right now I'm going to get some older kernels to try to find out when the
problem was created (for me).

Jim

-- 
James E. Carpenter                           E-Mail: jimc@zachary.riva.com
23 Munroe Drive                                 Tel: (508) 643-0908
Plainville, MA  02762-1132

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

From: duval@elpp1.epfl.ch (Basil P. Duval EPFL - CRPP 1015 Lausanne CH)
Subject: Motif 1.23 compilation
Date: 20 Jul 1994 10:21:04 GMT
Reply-To: DUVAL@ELPP1.EPFL.CH


hello world,
        Since I have a legal access to the Motif sources, I decided to try
and make the motif libraries. I have tun into a problem

Kernel 1.31 (also 1.26) Make 3.71 (also 3.70)

The interation in the make Makefiles command ( have a looong look at that file)
seems to put my machine into a loop launching the same make over and over again
until I am forced to kill the whole thing off. I can see from the messages
that it is redoing the same commands over and over, but I can not understand
why ?
        Is this a known problem. Should I just leave it well alone for a couple
of days and leave it to sort itself out. It launches so many make procedures
that I get the "unable to vfork" back from the machine which starts swapping
like doomsday has arrived (16Mbyte + 16 swap)

ok, could we put the compilation of Motif from the standard distribution into
the FAQ....

any reply welcome


Basil P. DUVAL
EPFL/CRPP
1015 Bassenges
Lausanne, Switzerland
Email: DUVAL@ELPP1.EPFL.CH

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

From: richard@fiu.edu (Albert Chin-A-Young)
Subject: if_arp.h in 1.0.9 and 1.1.31
Date: 20 Jul 1994 10:25:01 GMT

I tried compiling tcpdump v3.0 and the <net/if_arp.h> include file
couldn't find the struct arphdr (tcpdump-3.0/linux-include/netinet\
/if_ether.h). So, I took a look at the if_arp.h file in v1.0.9 and
found it to be v1.0.1 and it contained struct arphdr. However,
the if_arp.h in v1.1.31 is also v1.0.1 and it does not contain
this struct. Why would both version numbers be the same but the
contents different?

albert

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

From: thuerman@ibr.cs.tu-bs.de (Urs Thuermann)
Subject: Re: gcc and Linux (compiling)...
Date: Wed, 20 Jul 1994 10:41:57 GMT

edgar@uwast.astro.wisc.edu (Dick Edgar) writes:

>Some C compilers may assume these things [undeclared variables are ints;
>code not in a function is in main()] 

No, if it does, it is broken.  Functions must always have a name, and
variable declarations always need at least one type specifier, storage
class specifier or type modifier.

Urs

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

From: rydberg@kuai.se (Johan Rydberg)
Subject: MGR?
Date: 20 Jul 1994 10:46:01 GMT

Have anyone compiled MGR? Is it better than Xwindows??

--
Name: Johan Rydberg     email: mct@cyberspace.org
Phone: ++46 485 52175   smail: 4186 Spjutterum, 386 96 Farjestaden, Sweden.
 
 -"MS Windows ar en bug i MSDOS, tabort den, snabbt!"


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

From: hpulley@uoguelph.ca (Harry C Pulley)
Subject: Re: MGR?
Date: 20 Jul 1994 11:19:13 GMT

Johan Rydberg (rydberg@kuai.se) wrote:
: Have anyone compiled MGR? Is it better than Xwindows??

Define better...

I haven't run it since I switched to Linux, only Coherent.

It is much smaller.  On Coherent, about 500K would be the total memory taken up
for window manager, several terminal windows, clock, biff, etc.

It is also slower.  The video code is alright but it could be a lot faster.  As
well, there is no support for accelerated boards or anything like that.

There aren't that many clients.  If most of what you do is terminal windows,
clock and biff then this will be OK.

Finally, the color support doesn't work properly when I last looked at it.  If
mono is OK then you are alright.  Most cards should be able to do 800x600 mono
if you set up a header file for the registers in your card by running the
regs.exe program under DOS to read your BIOS.

It is a nice little package.  If you don't have 8MB of RAM or virtual memory
(as in the case of Coherent) then it is a nice alternative to X.

Since I have 8MB of RAM and 16MB of swap, I prefer to run accelerated X in
1024x768x256 instead.

Harry
--
<:-{}   hpulley@uoguelph.ca       |This message released|It takes all kinds,
 \      Harry C. Pulley, IV       |to the PUBLIC DOMAIN.|and to each his own.
==================================+=====================|This thought in mind,
Stay away from the DOS side, Luke!|Un*x don't play that.|I walk alone.

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

From: dgarrett@orbit.cs.engr.latech.edu (Don Garrett)
Subject: Re: fact on linux vs sun
Date: 19 Jul 1994 21:07:04 GMT


Alex Ramos (ramos@engr.latech.edu) wrote:
: Kjetil Torgrim Homme (kjetilho@ifi.uio.no), quoted out of context, wrote:
: > +--- Supat Faarungsang:
: > | I test sun4.1.1 on sparcII vs linux1.1.29 on 486dx-50 and found that
: > | in all features linux has greater performance speed is about 10%
: > | faster on linux in all kind of programs.

: > If you want to bring out Sun's forte, try to run a math problem with a
: > dataset twice the size of RAM while you do normal interactive work. I
: > think you will find response time is much better on the Sun. Suns are
: > better at handling many users, at least compared to an ISA-bus Linux
: > box.                                          ^^^^^^^^^^^^^^^^^^^^^^

: Yes, of course. They're also faster when compared to a 386/SX16.


  I used to run my 486DX/33 (12M) at home, off the network, and it
blew our IPC's and IPX's (12/16M) out of the water. Badly. But now that
I've moved the machine up to the school, onto the school's network,
run NIS, and NFS mount home dirs, etc. I find that I don't really out
perform them anymore. It's been very disappointing to me. For general
work, my PC seems to be very similar to an IPC. 

  I still prefer the PC, but I'm really suprised out how much the full
networking costs in performance. I am hoping that another 4M will help
make up for a lot of the lost speed. And maybe a high-class networking
card. (I have an old 8 bit card with DMA)

--
Don Garrett                                                   Louisiana Tech
dgarrett@engr.latech.edu                                      University
                  http://info.latech.edu/~dgarrett/

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

From: neuffer@goofy.zdv.Uni-Mainz.DE (Michael Neuffer)
Subject: Re: PORT TO ATARI (& AMIGA)
Date: 20 Jul 1994 11:33:52 GMT

In article <1994Jul16.155041.1556@radar.demon.co.uk>, richard@radar.demon.co.uk (Richard Hodson) writes:
"> "Well enough that there is now a newsgroup maus.os.linux68k for
"> "discussing it (mostly in German, with very little English).  

We just need some more English speaking folks in this newsgroup. 
As long as only 2 English speaking persons post and read in this 
newsgroup, nobody has a reason to switch from German to English.

As soon as more and more English speaking people start posting into 
maus.os.linux68k the language will change from German to English.

If the reason that you are not posting is that you don't understand
German, don't hesitate to ask for a translation of the German
text. 


---
Maus-/UseNet:Michael_Neuffer@wi2.maus.de
Usenet      :neuffer@goofy.zdv.uni-mainz.de
             linux@uni-koblenz.de
Fido        :Michael Neuffer@2:245/5530.5


 


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

From: wjadams@news.db.erau.edu (Bill Adams)
Subject: re:Massisve GCC error
Date: 20 Jul 1994 01:49:34 GMT

hey, had the same type problem, it turned out I had 3 chip simms, and
their paging algorithm is too slow for a unix like os... (however 
"speed-blazing" dos doesn't mind a bit :>).  Could be you have the
3 chip variety of simms rather than the 9 chip.  

Hope that helps.....

                        Bill D. Cat

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

From: ron@draconia.hacktic.nl (Ron Smits)
Subject: Re: ping -f crashes (at least) linux-1.1.29
Date: 20 Jul 1994 13:55:59 GMT

>>>>> "Michael" == Michael Bongartz <bongartz@cs.uni-sb.de> writes:

    Michael> Guenther Thomsen (thomsen@cs.tu-berlin.de) wrote: : Hi,

    Michael> : if I try ping -f vera from doma ( both 8Mb RAM, linux
    Michael> 1.1.29, libc-4.5.26 : connected by ppp-2.1.2a ) my system
    Michael> crashed and I got following : last messages:

    Michael> : Kernel panic: Trying to free up swapper memory space :
    Michael> in swapper task - not syncing : Oops : 0000 : EIP ...  :
    Michael> ... corrupted stack page : Process swapper ...  :
    Michael> ... Code: <1> unable to handle kernel paging request at
    Michael> kernel address ...

    Michael> : ...

    Michael> : Aiee, killing interrupt handler : task[0] (swapper)
    Michael> killed. unable to recover : Kernel panic: Trying to free
    Michael> ...

    Michael> We've this problem on a 19k2 leased-line, running pppd
    Michael> 2.1.2a, linux 1.1.27.

    Michael> It happens with older kernels too. It seems not to crash
    Michael> using SLIP - so maybe it's really a ppp related problem.

    Michael> I sent a mail to the linux activists (net and ppp), but
    Michael> there was no answer.

    Michael> Perhaps, people with a similiar setup (ppp _or_ slip)
    Michael> could try 'ping -f', so we'll see if it's really a ppp
    Michael> problem?

    Michael> But pls., type some syncs before ;-)

    Michael> Micha

I tried it out on two different machines and with two different
kernels namely 1.1.19 and 1.1.30. The 1.1.19 kernel runs on a
386SL25Mhz machine, the 1.1.30 kernel on a 486DX50 machine. Both
kernels panic I debugged the EIP address and it's in the
dev_queue_xmit function in the file .../linux/net/inet/dev.c.

Is this something that can be solved easily by the netguru's or should
I try to find out more about. 

I'm also sending this to the PPP mailing list

Regards
--



                Ron Smits
                ron@draconia.hacktic.nl
                Ron.Smits@Netherlands.NCR.COM

/*-( My opinions are my opinions, My boss's opinions are his opinions )-*/
/*-(                They might not be the same                        )-*/

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

From: cdent@honors.indiana.edu (NetDog)
Subject: Re: Is XFree86 3.X from X11R6 ?
Reply-To: cdent@indiana.edu
Date: Wed, 20 Jul 1994 06:29:04 GMT

>>>>> "FL" == Frank Lofaro <ftlofaro@unlv.edu> writes:

   FL> In article <CsxInp.GC1@aib.com> hohndel@aib.com (Dirk Hohndel)
   FL> writes:
   >>  3.0 is out (part of X11R6) but almost unusable on Linux
   >> 

   FL> Exactly what is wrong with it? I know it is beta code. I could
   FL> deal with some weirdness or bugs. Does it crash all the time or
   FL> something like that?  Is it woefully incomplete, missing key
   FL> features? Or is it dangerous?  (e.g. will I run the risk of
   FL> frying something by running it. I don't have a screwey RAMDAC,
   FL> so the card is safe, but if Xfree 3.0 tells the card to output
   FL> 150kHz horizontal, my monitor will explode, and I won't be
   FL> happy :|) Or can people use it to hack root? (since its suid
   FL> root)

I've been using R6 on my 386 with a et4000 card since a few days after
the release. No problems yet. My guess is the xfree people are just
rying to make sure they've got their asses covered....which makes
sense with a project that is this popular.

Chris

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

From: slouken@cs.ucdavis.edu (Sam Oscar Lantinga)
Subject: Patchlevel 31 breaks UMSDOS
Date: Tue, 19 Jul 1994 18:37:20 GMT


        The file linux/fs/msdos/mmap.c from the UMSDOS distribution
contains references to share_map(), which no longer exists in patchlevel 31.
It should be an easy fix, but I don't know where to start.

Laterz,

        -Sam


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

From: mdw@cs.cornell.edu (Matt Welsh)
Subject: Re: Kernel Hacker's Guide
Date: Wed, 20 Jul 1994 14:57:49 GMT

In article <jwshin.774593830@nitride.EECS.Berkeley.EDU> jwshin@nitride.EECS.Berkeley.EDU (Jinwoo Shin) writes:
>Will there be an upgrade to the Kernel Hacker's Guide by LDP? 

Eventually, yes. Once I finish a few projects that I'm currently working
on, I plan to hack the KHG full-time, with Michael's help, of course.
From what I understand, Michael's been swamped of late.

The KHG does need updating, but the groundwork is there. I'm going to
write a kernel primer/tour---which might also appear in the I&GS---and
rewrite the memory management chapter. I'll also work on moving the
architecture-dependent sections to separate chapters.

Once we get rolling on this, we'll probably solicit work from volunteers...

mdw

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

From: bongartz@cs.uni-sb.de (Michael Bongartz)
Subject: Re: ping -f crashes (at least) linux-1.1.29
Date: 20 Jul 1994 07:25:54 GMT

Guenther Thomsen (thomsen@cs.tu-berlin.de) wrote:
: Hi,

: if I try ping -f vera from doma ( both 8Mb RAM, linux 1.1.29, libc-4.5.26 
: connected by ppp-2.1.2a ) my system crashed and I got following
:  last messages:

: Kernel panic: Trying to free up swapper memory space
: in swapper task - not syncing
: Oops : 0000
: EIP ...
: ... corrupted stack page
: Process swapper ...
: ... Code: <1> unable to handle kernel paging request at kernel address ...

: ...

: Aiee, killing interrupt handler
: task[0] (swapper) killed. unable to recover
: Kernel panic: Trying to free ...

We've this problem on a 19k2 leased-line, running pppd 2.1.2a, linux 1.1.27.

It happens with older kernels too. It seems not to crash using SLIP - so maybe
it's really a ppp related problem.

I sent a mail to the linux activists (net and ppp), but there was no answer.

Perhaps, people with a similiar setup (ppp _or_ slip) could try 'ping -f',
so we'll see if it's really a ppp problem?

But pls., type some syncs before ;-)

Micha


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

From: bongartz@cs.uni-sb.de (Michael Bongartz)
Subject: Re: ping -f crashes (at least) linux-1.1.29
Date: 20 Jul 1994 07:27:07 GMT

Malcolm Reeves (reeves@pangea.usask.ca) wrote:
: Guenther Thomsen (thomsen@cs.tu-berlin.de) wrote:
: : Hi,

: : if I try ping -f vera from doma ( both 8Mb RAM, linux 1.1.29, libc-4.5.26 
: : connected by ppp-2.1.2a ) my system crashed 

: This is not a 1.1.29 problem - it works fine for me with a 3c509 ethernet
: card. Maybe your hardware can't send/receive back to back packets and there
: is some kind of buffering problem ! My 3c501 ethernet card was capable of
: crashing almost any kernel ! It is now acting as concrete agregate - on of
: the better uses for 3c501's .....

It only happens using ppp here.


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

From: d90-cka@dront.nada.kth.se (Carl Karlsson)
Subject: Re: MAP_SHARED?
Date: 20 Jul 94 09:41:08

In article <30i31l$164@search01.news.aol.com> dorwins@aol.com (DorwinS) writes:

>   I wanted to use the shared memory with MAP_ANON-- and it didn't give me an
>   error but it doesn't seem to work correctly--is this not supposed to work
>   or
>   did I just not do it right--I've checked the pointers (printing them out)
>   and the forked  children are writing to the correct location.
>   I'm using MAP_ANON and MAP_SHARED with fd set to -1 and I'm using
>   PROT_READ and PROT_WRITE-- if this won't work can someone recommend
>   another way to send data from child to parent?

Perhaps using pipes? Check out the pipe(2) system call.. Depending on what
you want to do, it's probably easier anyway. :)


   C

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


** 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
******************************
