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:     Thu, 17 Feb 94 06:13:04 EST
Subject:  Linux-Development Digest #474

Linux-Development Digest #474, Volume #1         Thu, 17 Feb 94 06:13:04 EST

Contents:
  Re: Serial Driver Leaves DTR On (Rob Janssen)
  PL15e kcore fix looks a bit fishy (Frank Lofaro)
  Re: Connecting two hosts via SCSI (Dan Swartzendruber)
  Problems with mmap() ?? (Tim Liddelow)
  BSD file system for linux ? (Faisal Chowdhury)
  Re: Two kernel sound support things. (Des Herriott)
  Re: Stuff to solve before 1.0 (problems with 15d) (Byron Faber)
  BBS software for Linux NEEDED -- *HELP!* (Donald Burr)
  Re: Stuff to solve before 1.0 (Bill C. Riemers)
  DigiBoard Support (Shawn Rhoads)
  Re: [DIAGNOSIS] Kernel and telnetd bug (Karel Kubat)
  Re: GNU finger bizarro under Linux... (Ian A Murdock)
  Re: Read/Write EXT2 from DOS (Robert Sanders)
  [DIAGNOSIS] Kernel and telnetd bug (Billy L. Williams Jr.)
  Re: ANNOUNCE: Motif 1.2.3 for Linux/Coherent/FreeBSD/NetBSD/ (Tero Laakkonen)
  Is SWAPDEV (0x1F6) still used? (Rick)

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

From: rob@pe1chl.ampr.org (Rob Janssen)
Subject: Re: Serial Driver Leaves DTR On
Date: 16 Feb 94 19:31:38 GMT
Reply-To: pe1chl@rabo.nl

In <CLALzo.6qs@csc.ti.com> Andrew J. Piziali <andy@piziali.lonestar.org> writes:

>   I just started using the communications program Seyon and discovered when
>the program terminates, the modem DTR is left on.  Subsequent executions of
>cu(1) and kermit(1) will also leave DTR asserted when they terminate although
>prior to the execution of Seyon DTR would have been deasserted.

>   The only way I've been able to turn off DTR is to reboot the kernel.  Is
>there a known bug in 0.99.14's serial device driver which could account for
>this?

>   The pertinent vital signs are:

>       o  0.99.14 kernel

>       o  "Serial driver version 3.96 with no serial options enabled"
>          "tty01 at 0x02f8 (irq = 3) is a 16550A"

>                       -- from the boot messages

>       o  Telebit T1000 modem

>   The main annoyance caused is having the modem auto-answer because DTR has
>been left on ...

It is not a bug in the kernel, but in Seyon!
I have mailed it to the author back when he was still maintaining it,
and I think he had made the fix in the source but there was no release
after that.
Someone else took over the maintenance, has anyone seen any signs of
a new release?

Rob


BTW: the bug is that Seyon enables the "clocal" flag for your line and
never turns it off.  Use: stty -clocal </dev/ttyS1
-- 
=========================================================================
| Rob Janssen                | AMPRnet:   rob@pe1chl.ampr.org           |
| e-mail: pe1chl@rabo.nl     | AX.25 BBS: PE1CHL@PI8UTR.#UTR.NLD.EU     |
=========================================================================

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

From: ftlofaro@unlv.edu (Frank Lofaro)
Subject: PL15e kcore fix looks a bit fishy
Date: 17 Feb 94 01:22:29 GMT

I looked at the 0.99.15e patch and the /proc/kcore fix looks a bit fishy.
I checked my diff which I posted, and one of the (what seems to me) important 
fixes is gone.

Here is the 0.99.15e fs/proc/array.c diff:

diff -u --recursive --new-file pl15d/linux/fs/proc/array.c linux/fs/proc/array.c
--- pl15d/linux/fs/proc/array.c Mon Feb 14 19:00:30 1994
+++ linux/fs/proc/array.c       Wed Feb 16 09:34:59 1994
@@ -42,9 +42,9 @@

        if (count < 0)
                return -EINVAL;
-       if (p >= high_memory)
+       if (p >= high_memory + PAGE_SIZE)
                return 0;
-       if (count > high_memory - p)
+       if (count > high_memory + PAGE_SIZE - p)
                count = high_memory - p;
        read = 0;


Note, the line after the second +. Shouldn't it be:
count = high_memory + PAGE_SIZE - p

?

else if count is too high, it will get truncated down based on a one page 
too short kcore. In fact, if this happens when the position is in the last 
page of memory, ( i.e p > high_memory) count will be negative, which would be
bad.

I can't test this from here, but it definitely looks a bit odd to me.

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

From: dswartz@pugsley.osf.org (Dan Swartzendruber)
Subject: Re: Connecting two hosts via SCSI
Date: 16 Feb 1994 18:11:17 GMT

In article <CLBL3M.5JM@cs.utwente.nl>, heida@cs.utwente.nl (Kees K.S. Heida) writes:
> Christian Herkt (macbett@cs.tu-berlin.de) wrote:

[snip]

> So here is the question:
> 
> If anyone started with the implementation of SCSI network drivers please
> responde to this discussion.

Anyone who *does* get information on this, please post it!


-- 

#include <std_disclaimer.h>

Dan S.

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

From: tim@jaguar.catt.citri.edu.au (Tim Liddelow)
Subject: Problems with mmap() ??
Date: Wed, 16 Feb 94 00:37:53 GMT

Ok.  I have basically used the example in Stevens' 'Advanced Programming in the
Unix Environment' which mmap()'s /dev/zero and uses this to provide a shared memory
region between like processes.   Example code:

fd = open("/dev/zero", O_RDWR);
/* check for valid fd */

if ((area = mmap(0, SIZE, PROT_READ | PROT_WRITE, MAP_SHARED, fd, 0))
        == (caddr_t) -1)
   perror("mmap");

===============
Now this code returns EINVAL.  I run the process as root, so there isn't any
problems with the open of /dev/zero.  I do notice however, that MAP_PRIVATE does
seem to work.  So am I doing something wrong, or is mmap() broken?

Tim.
-- 
________________________________________________________________________________
  Tim Liddelow                                          'Out out brief candle'.
  Software Engineer / Systems Programmer
  Centre of Advanced Technology in Telecommunications   Live like there's no
  CITRI, Melbourne, Australia                           tomorrow.
  internet : tim@jaguar.catt.citri.edu.au                
  Phone : +61 3 282 2455   Fax : +61 3 282 2444         
________________________________________________________________________________

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

From: faisal@corona.cus.com (Faisal Chowdhury)
Subject: BSD file system for linux ?
Date: 16 Feb 1994 16:55:38 GMT

 Is any body working on implementing the BSD file system for linux ?

-- 
Faisal
Western Atlas Software, Houston, TX

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

From: dnh@mfltd.co.uk (Des Herriott)
Subject: Re: Two kernel sound support things.
Date: 16 Feb 1994 17:07:45 -0000

In article <FRANK.94Feb16161319@namu02.gwdg.de>,
Jens Frank 29206029 <frank@namu02.gwdg.de> wrote:
>What about porting the NetAudio system of (?) NCD. As far as I read, it
>supports
>several formats and automatic adaption of the samplerate, transparent for the
>user.  Further it supports netwide sound applications, just like X-Windows
>doesfor graphics. (client/server, and so on) .

(Tut tut - line length!)

Apparently work on this is already underway, and is at an almost
usable state.  I don't know much about it myself, but the mailing
list netaudio@ncd.com might provide you with some useful information.

Subscribe by sending a note to netaudio-request@ncd.com.

(This is second-hand information - apologies for any inaccuracies)

-- 
         Des Herriott,       \_  What goes down stairs, alone or in pairs,
       Systems Programmer,     \_  rolls over your neighbour's dog?
      Micro Focus, Newbury .     \_  What's great for a snack and fits
dnh@mfltd.co.uk / +44 0635-565354  \   on your back, it's log, log, log!

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

From: btf57346@uxa.cso.uiuc.edu (Byron Faber)
Subject: Re: Stuff to solve before 1.0 (problems with 15d)
Date: 16 Feb 1994 18:20:54 GMT

To be honest, I've run all the ALPHA series for some time, and my
system has been quite stable.  I haven't had any crashes at all.

Well.. I must however say that: 
ALPHA-pl15d now gives me problems with RPC.   Something with 
NFS has gone wrong.  Maybe the changed nfs system did it.   It occured
with pl15d though.  

I'll see what 15e does.

Byron

P.S.  The error is 

Cannot register service: RPC: Unable to send; errno = Network is unreachable
unable to register(NFS_PROGRAM, NFS_VERSION, UDP).


Everything else works fine (I think)  only had 15d running for a couple hours.
-- 
PGP 2.3 key available (in plan file) at:        Support public code:
btf57346@uxa.cso.uiuc.edu                       Use GNU software and others.
other accts at:  btf57346@sumter.cso.uiuc.edu & bf11620@coewl.cen.uiuc.edu 

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

From: dburr@sbphy.physics.ucsb.edu (Donald Burr)
Crossposted-To: comp.os.linux,comp.os.linux.admin,comp.os.linux.misc,comp.os.linux.help,comp.bbs.misc,alt.bbs,alt.bbs.internet,comp.unix.sys5.misc,comp.unix.pc-clone.32bit,comp.unix.admin,comp.unix.misc,comp.unix.programmer,comp.unix.questions,comp.unix.sysv386
Subject: BBS software for Linux NEEDED -- *HELP!*
Date: 17 Feb 94 06:29:15 GMT

I'm looking for a BBS package that will WORK under Linux.  All of the
ones I've tried (that I got from FTP sites, etc.) do not work
satisfactorially.

I'm running a pretty recent version of Linux (0.99pl15, with gcc 2.4.5)
Some of the BBS packages failed to work because they were designed for
older versions of the kernel and compiler.

PLEASE HELP ME -- I FINALLY am able to bring alive my lifelong dream of
starting a BBS; I just got my own phone line!  BUT, without a working
BBS software...  (and NO, I do NOT want to switch to <ick! blegh!> DOS!)

Thanks -- PLEASE EMAIl to picard@rain.org -- I do not read news very
often, and my news site is kinda unreliable.
--
Donald Burr
EMAIL: picard@rain.org

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

From: bcr@bohr.physics.purdue.edu (Bill C. Riemers)
Subject: Re: Stuff to solve before 1.0
Date: 16 Feb 94 15:13:01 GMT

In article <2jrg9a$b8@layla.RoBIN.de> layla.RoBIN.de wrote:
>Otherwise the system works fine for me and very stable. The mentioned
>lockup was my first for a year or so, I normally have uptimes > 30d.
>This is with normally two interactive users, >15 uucp-logins per day. 
>The box is also a pc-nfs server.

Hmmm, I still succeed in crashing my system once a day, and I'm not
using any networking execept loopback with term...  Normally the 
problems are non-reproducable.  As an example, yesterday I found that
/usr/X386/bin disappeared while the screen saver was running. OK, I 
could have killed any process locking files and then remounted root
read-only, but I found it was much easier to reboot and let e2fsck
take care of it automatically.  Not much was running at the time, and
everything running had proven themselfs safe 100's of times before...

                                Bill


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

From: shawn@eng.tridom.com (Shawn Rhoads)
Subject: DigiBoard Support
Reply-To: shawn@eng.tridom.com
Date: Wed, 16 Feb 1994 13:12:22 GMT

I posted a question about DigiBoard support that someone answered, but for some
reason it did not make it to my newsgroup feed. ?  Would someone be kind enough
to directly give me the contact information for the wonderful person working on
the DigiBoard support?

:) Thanks,

- Shawn Rhoads
shawn@eng.tridom.com
SRHOADS2@nvn.com


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

From: karel@icce.rug.nl (Karel Kubat)
Subject: Re: [DIAGNOSIS] Kernel and telnetd bug
Date: Thu, 17 Feb 1994 07:23:11 GMT

>>I'm having problems with termial servers locking up when
>>telnetted to Linux (only kernels pl14 and up).  If I telnet to
>>another machine (like a Sun) and _then_ telnet to linux,
>>everything works fine. The folks who manage our modem pool 
>>(a couple of ip terminal servers with modems attached) say that
>>I should include the '-trailers' option in my ifconfig line.

>Well, I don't know if it's this trailers thing, but I am certainly
>having my share of troubles.  Ever since about pl 14x, people have been
>unable to login to my machine from certain machines.  I finally tracked
>it down to whether or not the TERM variable is set.  If $TERM is not
>sent at the beginning of the connection (that is how it happens, I
>presume, maybe some enlightened protocol god could enlighten us), then
>the connection just sits there; with $TERM sent, the connection works
>perfectly.

Hmmm.. I don't really know what the kernel does, but I may be able to give 
you a hint: mention the (initial) terminal type in /etc/ttytype. If after 
the login the term-type gets changed, that's ok too. I found that this 
solved some of my telnet-in problems; though the problems haven't been of 
your magnitude (hence, this hint may not work for you).

Here's a my /etc/ttytype:

console tty1
console tty2
console tty3
console tty4
console tty5
console tty6
vt100 ttyp0
vt100 ttyp1
vt100 ttyp2
vt100 ttyp3
vt100 ttyS1
vt100 ttyS3

Hope this helps..
Karel.
-- 
email: K.Kubat@icce.rug.nl             "Premature optimzation is the root
phone: (+31) 50 63 36 47                of all evil.." (Knuth)
mail : ICCE, Univ. of Groningen,
  P.O. box 335, 9700 AH Groningen, Netherlands

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

Crossposted-To: comp.os.linux.help
From: imurdock@shell.portal.com (Ian A Murdock)
Subject: Re: GNU finger bizarro under Linux...
Date: Wed, 16 Feb 1994 15:09:58 GMT

In article <CLAp2B.2Hw@unix.portal.com>,
Ian A Murdock <imurdock@shell.portal.com> wrote:
>I'm having some problems getting the GNU finger server daemon `fingerd'
>to run for any extended period of time.  I've applied the patch found
>in procps-0.92 to the GNU source for version 1.37.  Everything compiles
>fine, and everything works fine for awhile... but `fingerd' crashes
>very easily.
>
>It's quite strange, actually, because it runs just fine when root is
>the only user logged into the system.  I started it last night before I
>went to bed with only root logged in, and found it still running when I
>awoke.
>
>But here's where it gets strange... the minute I log in as anyone
>besides root... crash!  It dies.  And I'm stumped.
>

With a bit more experimentation, I discovered that when logged into the
system with only one specific user (this can mean multiple logins of
the same user), `fingerd' runs as it should.  I've been running here as
imurdock all morning without a problem.  But I just logged in as root
to another VC, and, sure enough, `fingerd' died.

What could be causing this?

Crossposted to comp.os.linux.development.

Ian 
--
Ian Murdock <imurdock@shell.portal.com>

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

From: gt8134b@prism.gatech.EDU (Robert Sanders)
Subject: Re: Read/Write EXT2 from DOS
Date: 16 Feb 94 23:20:05 GMT

rob@pe1chl.ampr.org (Rob Janssen) writes:

>In <2jgvuh$1gf@gilligan.npeg.orem.novell.com> jdsmith@novell.com (J. Douglas Smith) writes:

>>Has anyone written a utility that can read and write ext2 filesystems from  
>>DOS?  Any pointers would be appreciated.

>This has been discussed before, and the general agreement was that the
>utility was not required as we already have "Linux" :-)

Although I've put off working on my ext2fsperl so I can port it
to Perl5 (when Perl5 goes beta), once the ext2lib innards are
ported to DOS, I have a couple of fsperl scripts that can read
files from an ext2fs partition just fine.  Furthermore, I've
written part of a full-screen fs debugger in fsperl.  Being
able to run that from DOS might not be too bad.

But none of that will become relevant until perl5 gets more
stable.


--
 _g,  '96 --->>>>>>>>>>   gt8134b@prism.gatech.edu  <<<<<<<<<---  CompSci  ,g_
W@@@W__        |-\      ^        | disclaimer:  <---> "Bow before ZOD!" __W@@@W
W@@@@**~~~'  ro|-<ert s/_\ nders |   who am I???  ^  from Superman  '~~~**@@@@W
`*MV' hi,ocie! |-/ad! /   \ss!!  | ooga ooga!!    |    II (cool)!         `VW*'

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

From: williams@vierzk.bates.scarolina.edu (Billy L. Williams Jr.)
Subject: [DIAGNOSIS] Kernel and telnetd bug
Date: 17 Feb 1994 00:51:25 -0600

Newsgroups: comp.os.linux.admin,comp.os.linux.help
Subject: [DIAGNOSIS] Bug between kernel and telnetd
References: <1994Feb15.163515.23807@cs.uah.edu>
Organization: University of South Carolina

In article <1994Feb15.163515.23807@cs.uah.edu>,
Mike Prevost <mprevost@cs.uah.edu> wrote:

>I'm having problems with termial servers locking up when
>telnetted to Linux (only kernels pl14 and up).  If I telnet to
>another machine (like a Sun) and _then_ telnet to linux,
>everything works fine. The folks who manage our modem pool 
>(a couple of ip terminal servers with modems attached) say that
>I should include the '-trailers' option in my ifconfig line.

Well, I don't know if it's this trailers thing, but I am certainly
having my share of troubles.  Ever since about pl 14x, people have been
unable to login to my machine from certain machines.  I finally tracked
it down to whether or not the TERM variable is set.  If $TERM is not
sent at the beginning of the connection (that is how it happens, I
presume, maybe some enlightened protocol god could enlighten us), then
the connection just sits there; with $TERM sent, the connection works
perfectly.

This MUST be a kernel problem (how so, I don't know).  I have tried
every possible combination of libs, gcc, and telnetd that I could find.
The ONLY thing that really changed from the previous working
configuration was the kernel, that is until I _really_ looked for a
solution by fiddling around my my configuration....

Here's a script from another machine showing you the exact problem and
the way to goose it into working:

[ beginning of script ]

[2] % telnet vierzk.bates.scarolina.edu
Trying 129.252.177.4 ...
Connected to vierzk.bates.scarolina.edu.
Escape character is '^]'.

^]
telnet> close
Connection closed.
[3] % setenv TERM vt100
[4] % telnet vierzk.bates.scarolina.edu
Trying 129.252.177.4 ...
Connected to vierzk.bates.scarolina.edu.
Escape character is '^]'.


Linux 0.99.15e (vierzk.bates.scarolina.edu.(none)) (ttyp2)

login: Connection closed by foreign host.

[ end of script ]

Can anybody track this problem down?  I am sure that it's screwing a lot
of people up....

Billy L. Williams, Jr.
--
  Billy L. Williams, Jr.     e-mail:  williams@vierzk.bates.scarolina.edu
  University of South Carolina Master of International Business Candidate
  Try Linux--it will be the last OS you ever install............Go Cocks!
  ..................finger me for pgp key and more info..................
Newsgroups: comp.os.linux.admin,comp.os.linux.help
Subject: [DIAGNOSIS] Bug between kernel and telnetd
References: <1994Feb15.163515.23807@cs.uah.edu>
Organization: University of South Carolina

In article <1994Feb15.163515.23807@cs.uah.edu>,
Mike Prevost <mprevost@cs.uah.edu> wrote:

>I'm having problems with termial servers locking up when
>telnetted to Linux (only kernels pl14 and up).  If I telnet to
>another machine (like a Sun) and _then_ telnet to linux,
>everything works fine. The folks who manage our modem pool 
>(a couple of ip terminal servers with modems attached) say that
>I should include the '-trailers' option in my ifconfig line.

Well, I don't know if it's this trailers thing, but I am certainly
having my share of troubles.  Ever since about pl 14x, people have been
unable to login to my machine from certain machines.  I finally tracked
it down to whether or not the TERM variable is set.  If $TERM is not
sent at the beginning of the connection (that is how it happens, I
presume, maybe some enlightened protocol god could enlighten us), then
the connection just sits there; with $TERM sent, the connection works
perfectly.

This MUST be a kernel problem (how so, I don't know).  I have tried
every possible combination of libs, gcc, and telnetd that I could find.
The ONLY thing that really changed from the previous working
configuration was the kernel, that is until I _really_ looked for a
solution by fiddling around my my configuration....

Here's a script from another machine showing you the exact problem and
the way to goose it into working:

[ beginning of script ]

[2] % telnet vierzk.bates.scarolina.edu
Trying 129.252.177.4 ...
Connected to vierzk.bates.scarolina.edu.
Escape character is '^]'.

^]
telnet> close
Connection closed.
[3] % setenv TERM vt100
[4] % telnet vierzk.bates.scarolina.edu
Trying 129.252.177.4 ...
Connected to vierzk.bates.scarolina.edu.
Escape character is '^]'.


Linux 0.99.15e (vierzk.bates.scarolina.edu.(none)) (ttyp2)

login: Connection closed by foreign host.

[ end of script ]

Can anybody track this problem down?  I am sure that it's screwing a lot
of people up....

Billy L. Williams, Jr.
--
  Billy L. Williams, Jr.     e-mail:  williams@vierzk.bates.scarolina.edu
  University of South Carolina Master of International Business Candidate
  Try Linux--it will be the last OS you ever install............Go Cocks!
  ..................finger me for pgp key and more info..................

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

From: laakkone@klaava.Helsinki.FI (Tero Laakkonen)
Subject: Re: ANNOUNCE: Motif 1.2.3 for Linux/Coherent/FreeBSD/NetBSD/
Date: 17 Feb 1994 11:54:54 +0200

In <1994Feb16.195034.614@galen.physiol.ox.ac.uk> ph@physiol.ox.ac.uk (Patrick Haggard) writes:

>Listen, there's no need to get personal.  Leo is entitled to think what he
>likes.  His views may be extreme, but he may still be highly employable, for
>all I know.  Most reasonable people believe that a commercial software 
>sector is a good thing.  What is _NOT_ a good thing is when a few 
>multinational companies monopolise the market, and charge excessive prices
>for lousy products.  The computer is too important a tool for human progress
>and advancement for humanity to allow it to be abused in such a way.

yes, one of the reasons i used to hate commercial software - especially
for messydoss - was that it wasn't "open". i hate using mycrosoft stuff
that only works with mycrosoft stuff.

motif is pretty good though, because the specs have been released to the
public and you can't be sued for imitating their user interface or
something silly like that.
-- 
"i abhor you pretentious insight. i respect conscious guessing
because it consists of two good qualities: courage and modesty."
-imre lakatos

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

From: pclink@qus102.qld.tne.oz.au (Rick)
Subject: Is SWAPDEV (0x1F6) still used?
Date: 17 Feb 94 01:10:37 GMT

I have a kernel mod that needs information passed to it during boot,
much like the values of ROOTDEV and RAMDISK are passed to the kernel
now.  The problem is finding a word in the boot block that I can use.
A check of the source code didn't reveal any obvious use of the SWAPDEV
value (offset 0x1F6 in the boot block).  Is it still used?  Can I pinch
it for my own purposes?

Thanks,
Rick.

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


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