Subject: Linux-Development Digest #953
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, 27 Jul 94 18:13:11 EDT

Linux-Development Digest #953, Volume #1         Wed, 27 Jul 94 18:13:11 EDT

Contents:
  My updated list of programs for shadow password changes (Al Longyear)
  Re: file source? (David Monro)
  Byacc.tgz ... no liby.a ?? (aj7124323@ntuvax.ntu.ac.sg)
  PCI IDE.... (David paul Kauffman)
  Re: 1060 cyliders a problem? (Jeff Macdonald)
  On qmagic binaries.. (Jon Green)
  Re: 1060 cyliders a problem? (David Monro)
  Re: FYI -- context switching times (Gustaf Neumann)
  Driver for LMS CM206 CD-ROM? (Ed Carp)
  Re: Floppy error since 1.1.23 (Bruce Thompson)
  Re: procps-0.95 broken -- problem is in linux/sched.h (Jim Balter)
  APC UPS owners or potential buyers, trying to show user base (Craig Metz)
  Re: threads in kernel (Larry McVoy)

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

From: longyear@netcom.com (Al Longyear)
Subject: My updated list of programs for shadow password changes
Date: Sun, 24 Jul 1994 07:03:10 GMT

A few weeks ago, I asked if anyone knew which programs needed to be
changed to support the shadow password suite.

There were two reasons why I asked this.

1. I wanted to evaluate the amount of work needed to convert my system
to use the password suite. This is my selfish reason.

2. I wanted to prepare this list for others. This is my altruistic
reason. I hope the this list will help people who may be considering
using the code to realize that if you use these programs then they
must all say USE_SHADOW (or some similar define) or none must say
this. If you have only some of the programs with shadow suite then
your system will be even less secure than had you not started to
convert the programs.

Several people replied to my question. I thank all.

The programs listed below reflect the changed programs needed to
support shadow passwords. The algorithm which I used, beyond the ones
specifically known and listed, was to go to the sunsite archive, LGX,
and FreeBSD CDROMs and examine the source to the programs as much as
possible for the occurrence of the magic string 'pw_passwd'. If the
program cares about this field then there is a good chance that it
wanted to validate passwords. Of course, I verified that the programs
listed truely did want to validate the password. However, given the
nature of the software industry, additional programs will be developed
which will not be listed below.

The list:

1. Standard system utilities

chpass [1]
chsh
init.bsdi [2]
login
passwd [3]
pwdutils
simpleinit [4] 
su
useradd [3,5]

2. Serial Network drivers

dip
diplogin [6]
pppd [7]

3. Post Office Protocol related servers

mh
pop2d
pop3d
popper

4. Network programs [8]

bwnfsd
ftpd
nntpd
pcnfsd
sup
telnetd
tn3270 [9]
uucpd
yppasswd
samba
imapd

5. Remote access network programs

rexecd
rlogind
rshd [10]

6. Libraries [11]

libtelnet.a

7. Miscellaneous items

ftam2 [12]
lock
pennmush
perl [13]
screen
skey package [14] 
tcsh [15]
vlock

8. X utilities [16]

andrew
xauth
xdm
xlock
xlogin

9. Notes

[1] This is found only on the BSD source package. I have not found the
equivalent code for Linux.

[2] The init.bsdi program requests the password of root to enter the
single-user mode as a compiled security option.

[3] The shadow password suite has replacements for these programs
which are specifically designed to operate in that mode. The
non-shadow sources, those not in the suite, have no provisions for the
shadow suite. Use the programs which are included in the shadow suite
rather than the non-shadow copies.

[4] The program is included in the poeigl and utile packages.

[5] Older versions were called adduser.

[6] diplogin is an alias for the dip program.

[7] The pppd process is due to include the shadow password changes
with the next version. Mike Jagdis did the changes. If you need shadow
passwords and wish to use the pppd process for a 'server' mode for the
2.1.2 version (only, please) then ask longyear@netcom.com for a copy.

[8] Many of these programs are actually stored under a different
name. The common convention is to prefix them with 'in.' to indicate
that they are run by the inetd program. The telnet daemon (telnetd) is
commonly called 'in.telnetd'.

[9] The tn3270 client program uses a password validation only during
the processing of the server mode's shell command.

[10] The rshd process is also called 'rsh' in some systems.

[11] The libraries open the window of programs much wider than I am
able to determine. At the present time, I belive that only telnet
references this library.

[12] The ftam2 program is part of the isode package. Sunsite included
only patches to the original source on src.doc.ic.ac.uk:packages/isode
and others.

[13] The perl program has a function to retrieve the user record. This
includes the password field. As such, the password is made available
to the script. It would be possible to write a perl program which
would do password validation. It probably is not a good idea to
convert this program so that it reads the proper password. I leave it
to your discretion. (If you want perl to return the true password
then the module to change is doio.c)

[14] This is the 'one time password system'. It contains many programs
which must be changed. One program in the package is called skeyinit.

[15] tcsh uses a 'lock' function. The comment is that the lock function
only works for root if you use shadow passwords. This is due to the
fact that shells should never run setuid, and without root access, it
will not have access to the default password.

[16] My source files for the X-window system is not complete. There
may be many other programs which I did not list. Please consider this
section to be only the items known to me at this time and not
guaranteed to be the total list of all possible X programs.

-- 
Al Longyear           longyear@netcom.com

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

From: davidm@syd.dms.CSIRO.AU (David Monro)
Subject: Re: file source?
Date: Wed, 27 Jul 1994 11:42:37 GMT


> I've had mysterious problems when using the "file" command on certain files.
> It sometimes just dies with a segmentation fault.  I would very much like to
> re-compile it using the -g option and run it through gdb, but I have been
> completely unsuccessful in my attempts at locating the source code.
> 
> Can anyone tell me where I can get it?
> 
> Thanks in advance
>                 mike
> --
>                         Mike Dowling
> 
I seem to have got mine from the usr.bin section of the FreeBSD tarfiles.
Works fine for me (no coredumps that I've noticed). Not sure where you
would find FreeBSD near you, but freebsd.cdrom.com would certainly be
one place to look. Trouble is it tends to be busy.

        David Monro
--
Tomorrow will be canceled due to lack of interest.


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

From: aj7124323@ntuvax.ntu.ac.sg
Subject: Byacc.tgz ... no liby.a ??
Date: 24 Jul 94 15:14:54 +0800

Hi :
        Anybody notice that byacc.tgz that comes with slackware 2.0 package
doesn't have liby.a provided in it ... Where can I find the library ? 

regards
CY


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

From: dkauffma@tech.iupui.edu (David paul Kauffman)
Subject: PCI IDE....
Date: Mon, 25 Jul 1994 04:34:57 GMT

are there any drivers out there yet that will support the PCI IDE?

-Thanks

-Dave

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

From: jeff@zis.ziff.com (Jeff Macdonald)
Subject: Re: 1060 cyliders a problem?
Date: Wed, 27 Jul 1994 18:57:32 GMT

I have that exact drive and I'm using the UMSDOS file system. I dumped my old 
hard drive to the maxtor and have had no problems. DOS's fdisk had no 
problems. Don't you like how maxtor defines a megabyte as 1 million bytes? 
Gee.. I should define 1 dollar the same way....

In article <1994Jul26.191529.549@tora.RoBIN.de> sws@tora.RoBIN.de (Steffen W. 
Schilke) writes:>From: sws@tora.RoBIN.de (Steffen W. Schilke)
>Subject: Re: 1060 cyliders a problem?
>Date: Tue, 26 Jul 1994 19:15:29 GMT

>David Monro (davidm@dmssyd.syd.dms.CSIRO.AU) wrote:

>: In article <1994Jul20.164639.16726@cc.ntnu.edu.tw>, s21008@cc.ntnu.edu.tw
>(DL83-08) writes:
>: |> When fdisking my new Maxtor 7546AT hard drive (1060 cyl, 16 HD, 63 Sect.),
>: |> fdisk told me that having over 1024 cylinders could cause problems; it did!
>: |> When I hit 'p' to look at the new partitions, I got a wad of error messages
>: |> telling me that real and virtual 'this & that' (sorry, I don't remember --
>: |> it was a long day; see above posting). 

>: Well I run a drive with 1048 cylinders without problems. You get a silly
>message

>My Toshiba MK538 has 1120 cyl. and I have no problems at all. I just had
>to switch off the most options on my Adaptec 1542CF (exept scanning)

>Cheese          steffen
>--
>[Standard Disclaimer] in addition I would like to speak with my lawyer ....
>S. Schilke; PoBox 1213; 61102 Bad Vilbel; Germany  a.k.a  sws@tora.RoBIN.de
>                  Sokonoke Sokonoke tora-sama ga touru
>$@%9%F%U%'%s(J  $@CN2H!Z%7%k%1![(J  $@$=$3$N$1$=$3$N$18WMM$,DL$k(J
>~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~




Jeff Macdonald
Ziff Information Services
Ziff Communications
10 President's Landing
Medford, MA  
jeff@zis.ziff.com

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

From: jcgreen@iastate.edu (Jon Green)
Subject: On qmagic binaries..
Date: 27 Jul 94 18:39:07 GMT

I've seen a lot of talk recently about [qnzo]magic binaries.  Can
someone explain a little about what exactly this means, and maybe what
the advantages and disadvantages of each are?  I heard a Linux qmagic
binary would run on a NetBSD system; is this true?  If mine is a silly
question, please direct me to the appropriate document. :)

-- 
* Jon Green           *  Proud to be a member of   * 5646 Friley Hall        *
* jcgreen@iastate.edu * drunk_bastards@iastate.edu * Ames, Iowa 50012-0001   *
* Jon2@irc            *          *BURP*            * Phone (515) 296-0648    *

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

From: davidm@dmssyd.syd.dms.CSIRO.AU (David Monro)
Subject: Re: 1060 cyliders a problem?
Reply-To: davidm@dmssyd.syd.dms.CSIRO.AU (David Monro)
Date: Mon, 25 Jul 1994 04:49:43 GMT


In article <1994Jul20.164639.16726@cc.ntnu.edu.tw>, s21008@cc.ntnu.edu.tw (DL83-08) writes:
|> When fdisking my new Maxtor 7546AT hard drive (1060 cyl, 16 HD, 63 Sect.),
|> fdisk told me that having over 1024 cylinders could cause problems; it did!
|> When I hit 'p' to look at the new partitions, I got a wad of error messages
|> telling me that real and virtual 'this & that' (sorry, I don't remember --
|> it was a long day; see above posting). 

Well I run a drive with 1048 cylinders without problems. You get a silly message
from fdisk about the end of the last partition being wrong (because the
partition table truncates the cylinder value to 10 bits) but everything
works fine. Oh, make sure the partition(s) with the boot kernel and any
lilo files (well, all the files lilo needs at bootup anyway) lie wholly
within the lower 1024 cylinders, or one day you may get a nasty surprise as
the bios falls off the end of the disk it knows about...

        David
|> 
|> Now I've got the disk up and going, and it has crashed only twice, but
|> possibly due to idiocy, rather than hardware. Is this really a serious
|> problem?
|> 
|> Thanks!
|> 
|> --
|> 
|>  /\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\
|> %v  Curtis Dean Smith                      Internet: s21008@cc.ntnu.edu.tw   |
|> 0j  Institute of Chinese Language & Literature                               |
|> ?3  National Taiwan Normal University            Bitnet: NTNUS239@TWNMOE10   |
|>  \/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/

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

From: neumann@watson.ibm.com (Gustaf Neumann)
Subject: Re: FYI -- context switching times
Date: Wed, 27 Jul 1994 17:20:14 GMT

In article <313n9m$hoh@jac.zko.dec.com> from [26 Jul 1994 19:13:58 GMT]
 feingold@avette.zko.dec.com (Elan Feingold)  wrote:

 |> Got this off of a *.os.* group:
 |> 
 |> Date: Wed, 9 Feb 94 14:56:13 PST
 |> From: Larry.McVoy@eng.sun.com (Larry McVoy)
 |> To: dck@cse.nd.edu
 |> 
 |> 68     99 Mhz HP-735 HP-UX 9.01
 |> 80     33 MHz 486DX QNX 4.2
 |> 87    160 Mhz DEC 4000/610 DEC OSF/1 v1.3
 |> 98     40 MHz R3000 DEC 5000/240 Ultrix 4.2
 |> 98     66 MHz 486DX2 Linux 0.99.10          ***** Cool! *****
 |> 102     62 MHz IBM-RS6000/580 AIX 3.2
 |> 106     66 Mhz snake HP-UX 9.x
 |> 128     40 Mhz SS10 SunOS 4.1.3
 |> 154     25 MHz R3000 DEC 5000/200
 |> 158     50 MHz R4000 SGI
 |> 188     48 Mhz + 1Me$ SS10 SunOS 5.2
 |> 198     33 MHz Sparc 4/470 SunOS 4.1.1
 |> 210     33 Mhz 486 386BSD 0.1
 |> 212     50 Mhz RIOS AIX 3.2
 |> 225     40Mhz SS10 SunOS 5.2
 |> 230     40Mhz SS2 SunOS 4.1.2
 |> 281     20 MHz R3000 DEC 5000/120
 |> 345     33 MHz R3000/3010 SGI Irix 4.0.5
 |> 380     33 MHz 486 NetBSD
 |> 454     40Mhz SS2 SunOS 5.2
 |> 628     50Mhz Sparc classic SunOS 5.2
 |> 
 |> %Title Context switch times in microseconds (raw numbers)
 |> 
 |> I wonder how we do with the later kernels...
 
 well, on my 90MHz Pentium, single user mode, Larry's context switch
 timing program gives the following numbers (microseconds/switch)
 in five consecutive runs
 
 Linux 1.1.27     41; 38; 38; 38; 41
 Linux 1.1.35     40; 42; 41; 42; 42
 
 It looks, like recent Linux versions on a Pentium 90MHz have the lead
 amont the machines listed above

-gustaf
PS: i used the context switch program of Larry McVoy <lm@slovax.Eng.Sun.COM>
posted in comp.arch,comp.benchmarks on 18 Jun 1993 22:30:21 GMT 
with message id <m24gftINN3no@appserv.Eng.Sun.COM>

--
Gustaf Neumann                     neumann@watson.ibm.com
Postdoctoral/Visiting Scientist    Tel: (914) 784 7086
IBM T.J.Watson Research Center, P.O.Box 704
Yorktown Heights, New York 10598


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

Crossposted-To: comp.os.linux.help,comp.os.linux.admin
From: ecarp@netcom.com (Ed Carp)
Subject: Driver for LMS CM206 CD-ROM?
Date: Mon, 25 Jul 1994 05:34:18 GMT

Does anyone have, or is anyone working on, a driver for the LMS CM206 
double speed CD-ROM?  A friend wanted to get a double speed CD-ROM drive, 
so we went out to a computer store, found a MediaVision double speed 
package, and booted Linux.  Worked great, etc., so he goes somewhere else 
where the MediaVision was a bit cheaper, buys one, gets it home, and 
boots Linux.  No go.  :(  Is MediaVision packaging different double speed 
CD-ROM drives or what?  More importantly, is anyone working on a driver 
for this drive?

Any help appreciated.  Email, please! :)
-- 
Ed Carp, N7EKG/VE3                      Ed.Carp@linux.org, ecarp@netcom.com
Finger ecarp@netcom.com for PGP 2.5 public key          an88744@anon.penet.fi
If you want magic, let go of your armor.  Magic is so much stronger than
steel!        -- Richard Bach, "The Bridge Across Forever"

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

From: bruce@mdavcr.mda.ca (Bruce Thompson)
Subject: Re: Floppy error since 1.1.23
Date: 24 Jul 94 07:51:31 GMT

Sascha Klein (sklein@ramz.ing.tu-bs.de) wrote:
: Hi,

: I have a floppy error since kernel 1.1.23. I tested all kernels until
: 1.1.31. When I use the command

:       dd if=/dev/fd0 of=/dev/null

: I get lots of messages

:       Weird - unlocked, clean and not uptodate buffer on list ...

Just to add another data point, I've jumped from 1.1.18 to
1.1.32/33/34 and I'm getting the same behavior. The main thing I'd
like to know is if the data I'm writing to floppy is getting there!
It's tough to do backups when your backup device is non-functional!
[:-)] Time to break out vmlinuz.good (1.1.18)

        Cheers,
        Bruce
        -- 
Bruce Thompson, B.Sc.           | "A great many people think they are
Software Engineer               |  thinking when they are merely
MacDonald Dettwiler,            |  rearranging their prejudices."
13800 Commerce Parkway,         |       -- William James
Richmond, BC                    |
(604) 278-3411                  | Usual disclaimers apply
NAPRA #473                      |

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

From: jqb@netcom.com (Jim Balter)
Subject: Re: procps-0.95 broken -- problem is in linux/sched.h
Date: Mon, 25 Jul 1994 06:10:03 GMT

In article <30tkdq$id0@agate.berkeley.edu>,
Nick Kralevich <nickkral@po.EECS.Berkeley.EDU> wrote:
>
>Kernel version 1.1.34. 
>
>I created a small test file with the following in it:
>
>----- Begin -----
>#include <linux/sched.h>
>
>main ()
>{
>  printf("Hello");
>}
>----- End -----
>
>However, when I tried to compile it, I came up with these error 
>messages:
>
>In file included from /usr/include/linux/fs.h:15,
>                 from /usr/include/linux/sched.h:86,
>                 from testfile.c:1:
>/usr/include/linux/net.h:104: parse error before `select_table'
>In file included from /usr/include/linux/sched.h:86,
>                 from testfile.c:1:
>/usr/include/linux/fs.h:181: field `i_sem' has incomplete type
>In file included from /usr/include/linux/sched.h:86,
>                 from testfile.c:1:
>/usr/include/linux/fs.h:289: parse error before `select_table'
>/usr/include/linux/sched.h: In function `add_wait_queue':
>In file included from testfile.c:1:
>/usr/include/linux/sched.h:473: dereferencing pointer to incomplete type
>/usr/include/linux/sched.h:476: dereferencing pointer to incomplete type
>/usr/include/linux/sched.h:476: dereferencing pointer to incomplete type
>
>[many repetitions of above line deleted]
>
>
>From this I guessed that there was somthing wrong with the source code
>for linux, although I can't figure out what.  I tried going through the 
>source code, and I could find nothing wrong with it.
>
>The line where it first has a problem is:
>
>  int   (*select)       (struct socket *sock, int sel_type,
>                         select_table *wait);
>
>in net.h.  The error message was "parse error before select_table");
>
>This problem is keeping procps-0.95 from compiling.  Without
>a fix, the process version of ps can't compile.  I don't know 
>what was the first kernel version where the program wouldn't 
>compile, but I am testing it on 1.1.34.  It seems clear that there
>is somthing wrong with the source code, but I spent 4 hours,
>and couldn't find the first parse errror.  Agggggg!!!!!!!

I don't understand the difficulty.  The compiler told you, quite explicitly,
that testfile.c includes sched.h which includes fs.h which includes net.h,
and that net.h has an error on line 104 at the token "select"table".
select_table is obviously a typedef, and so it has to be defined before being
used.  Grepping the headers show that it is in linux/wait.h, so that needs to
be included somewhere.  Why would that take 4 hours?  Exactly what the best
fix is depends upon the Linux philosophy on nested header files.  My method
is to include a header file in any header file that requires it; this costs
nothing since, if headers are protected from multiple inclusion, gcc is clever
(tricky?) enough not to even open the header file more than once.
-- 
<J Q B>

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

From: cmetz@thor.tjhsst.edu (Craig Metz)
Crossposted-To: comp.unix.bsd
Subject: APC UPS owners or potential buyers, trying to show user base
Date: 27 Jul 1994 15:12:44 -0400

        I had a talk today with Debbie Gray (sp?) of American Power Conversion
regarding trying to get information on how to communicate with their Smart
UPS products' onboard controllers in order to write a Linux driver. APC is
one of the *many* manufacturers that plays the old NDA game, i.e., ``we
consider that to be proprietary information that we have to protect''. However,
she seemed to at least not be a brick wall to the idea of trying to change
this policy so that a freely redistributable in source form Linux driver could
be done. She told me that she will be talking to her supervisor when he's
back from Interop next week about this. 

        In cases like this, money talks. If it can be shown that opening up
to the Linux community (and the *BSD community, as well) would net them a
significant number of sales, they would be far more likely to decide to open
up than if it's just me. 

        So, I ask, could people who (or whose employer):

        * Own an APC Smart UPS used with a machine that is or will be a 
                Linux or *BSD system

        (or)

        * Are planning to purchase a smart UPS for use with a Linux or
                *BSD system 

        Please get in touch with me, letting me know what size (sizes, or
size range) UPS and how many units are involved. I will be using this 
information to show that the Linux and *BSD markets are significant enough
that it is worth their while to reconsider their policies. 

        Also, if anyone has had experiences with other manufaturers of
smart UPS units trying to get programming/driver information from them,
please let me know. If APC doesn't change their line, I want to know who
*does* do things in the open -- that'll be who gets my business.

                                                                -Craig
Newsgroups: comp.os.linux.development,comp.unix.bsd
Subject: foo
Summary: 
Followup-To: 
Distribution: 
Organization: The Thomas Jefferson High School for Science and Technology
Keywords: 
Cc: 


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

From: lm@stanford.edu (Larry McVoy)
Subject: Re: threads in kernel
Date: 27 Jul 1994 17:36:25 GMT

Sam Oscar Lantinga (slouken@cs.ucdavis.edu) wrote:
:       Maybe I'm stupid, but what's the difference between
: clone() and fork() ?

Fork() creates a new copy of your address space (maybe implemented with
copy on write tricks).  The crucial point is that a modification of any
data in one address space is not seen in the other address space.

The idea behind a clone() or tfork() system call is that the only thing
that you get is a new stack (it would be nice, I think, if the new stack
was at the same virtual address as the old stack; it saves you having to
rethread it).  The rest of the address space is shared so that both 
processes (or threads if you want to call them that) can modify data
and have those modifications seen by the other process.

As someone else pointed out there are other things that you want to consider.
These things are any operation that changes the state of the address space
or the process context.  Things like

        mmap() - does it occur in both address spaces (yes)
        open()/close() - does it affect the other process (yes)
        exit() - do all processes exit? (no)
        signal(a, b) - does this install a signal handler for all procs (no)
        etc.

I think that you probably want to have a set of defaults (I stuck my
preference for these defaults above) and a way to override the defaults.

If anyone is seriously working on this, please contact me and I will help.
--
--
Larry McVoy                     lm@cygnus.com                   (415) 821-5758

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


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