Subject: Linux-Development Digest #937
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:     Fri, 22 Jul 94 22:13:04 EDT

Linux-Development Digest #937, Volume #1         Fri, 22 Jul 94 22:13:04 EDT

Contents:
  buffer cache or isofs bug? (Eberhard Moenkeberg)
  bad email: A.Cox@swansea.ac.uk (Phil Howard)
  help start info (songer)
  Re: GOTO haters ..Re: Linux Performance Enhance ? (Ian McCloghrie)
  For those lusting for threads :) (Elan Feingold)
  Re: GOTO haters ..Re: Linux Performance Enhance ? (David Holland)
  Re: Anyone else lust for threads? (Marc Fraioli)
  Mode 2 PLIP ? (niall parker)
  Re: link() system call (Grant Edwards)
  Re: buffer cache or isofs bug? (Eberhard Moenkeberg)
  New kernel message at boot time? (Bogdan Urma)
  Re: BUG in e2fs (0.5a)?  (INTERNAL ERROR) (Thomas Parmelan)
  Whats wrong ? (New kernel bug??) (Peter Enderborg)
  Re: link() system call (Gareth Newfield)
  Re: Xfree86: increase pallate? (Phil Johnson)
  Re: DOES ANYONE KNOW ABOUT THE MAILINGLISTS? (hello, is anyone there?) (Dennis Mahle)
  Re: 1.1.32 kernel broken? (Bj|rn Isaksson)
  Re: dot clocks (Dan Logue)

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

From: emoenke@gwdu03.gwdg.de (Eberhard Moenkeberg)
Subject: buffer cache or isofs bug?
Date: Fri, 22 Jul 1994 16:50:10 GMT

Hi all,

there are (with 1.1.33, but earlier, too) severe problems
with the handling of "defective" CDROMS.

For example, my issue of the new InfoMagic "Linux Developer's
Resource" double CD (both disks) has some structural defects:
Some files have their directory entries, but are unreachable.

With the following script:
==============================================================
#!/bin/sh
CD=infomagic9406-1
DIR="~"
umount /CD1
mount -t iso9660 /dev/sbpcd2 /CD1
cd /CD1
FILE=dir.${CD}
ls -alR >${DIR}/${FILE}
FILE=find.${CD}
find -noleaf -printf "%8s %Tb %Td %Ty %p %l\n" >${DIR}/${FILE}
cd ${DIR}
sort -T /tmp +4 -o ${FILE} ${FILE}
umount /CD1
==============================================================

"ls" and "find" will give some error messages.

Well, things like that happen, and the InfoMagic CDs are
worth their money, anyway.

But after getting those errors, my system is confused somehow.
I get messages about kernel NULL pointer dereferences, 
segmentation faults (with programs I ran successfully before
starting that script), etc.

It seems like the buffer cache is disturbed or destroyed somehow
after using a defective CDROM.

So, developers, get your new InfoMagic CD set and find that bug. ;-)

Cheers, -e-
--
Eberhard Moenkeberg                                GGG   W   W  DDDD    GGG
                                                  G      W   W  D   D  G
E-Mail:   emoenke@gwdg.de                         G GGG  W   W  D   D  G GGG
Phone: +49 551 201551  Fax: +49 551 21119         G   G  W W W  D   D  G   G
Mail:                                              GGG   WW WW  DDDD    GGG
Ges. fuer wissenschaftliche Datenverarbeitung mbH
Am Fassberg
D-37077 Goettingen           GERMANY
Private Email: Eberhard_Moenkeberg@rollo.central.de (FIDO: 2:2437/210.27)


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

From: phil@zeus.fasttax.com (Phil Howard)
Subject: bad email: A.Cox@swansea.ac.uk
Date: 22 Jul 1994 12:36:35 -0500

Sorry for this being posted here, but it is obviously the only
choice when one is sent a bad email address on a reply (which
was to a posting here).  Please try replying again with a better
email address.  Thanks.
-- 
Phil Howard KA9WGN      | The drive spec says the capacity is 600mb unformatted
Unix/Internet/Sys Admin | and 525mb formatted.  So where do I find an unformat
CLR/Fast-Tax            | utility?
phil@fasttax.com        |

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

From: songer@wherads.demon.co.uk (songer)
Subject: help start info
Reply-To: songer@wherads.demon.co.uk
Date: Fri, 22 Jul 1994 17:35:33 +0000

>
>Increasingly, we are approaching a domain where we want to
>develope/experiment with our own operating systems/compilers. 
>In particular
>we would like to know of 
>   MACH /LINUX
>operating systems/kernel/compilers/source codes,
>
>which are available as shareware or freeware on 
>the internet_ftp and present an up to date starting point
>(post-partial start).
>
>Any further comments, pointer to mis haps, or source to  
>valuble hints and information is very very welcomed.
>

-- 
____Songer___
. 

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

From: imcclogh@cs.ucsd.edu (Ian McCloghrie)
Subject: Re: GOTO haters ..Re: Linux Performance Enhance ?
Date: 21 Jul 94 22:00:31 GMT
Reply-To: ian@ucsd.edu

bcr@k9.via.term.none (Bill C. Riemers) writes:

>of an infinite loop.  (Very rarely the case.)  Otherwise you should
>use a for() loop (or the equivlent of it) so you have a counter to 
>insure even if something in your code is broken, eventually the
>program will exit the loop.

I disagree.  for() statements are the "right" solution when one wants
to interate something a given number of times.  Using one for a "we'll
go until X is true, or we exceed 5000 interations" is A) confusing and
therefore difficult to maintain, B) either going to cause improper
behaviour if you set the boundary too low or fail to exit in a timely
fashion if you set it too high, and C) worst of all, likely to hide
errors in the program's logic (because instead of an infinite loop
being noticed during testing, you have something that runs to the end
of the for() loop, consistantly, in production code)

--
____
\bi/  Ian McCloghrie      | FLUG:  FurryMUCK Linux User's Group
 \/   email: ian@ucsd.edu | Card Carrying Member, UCSD Secret Islandia Club
GCS (!)d-(--) p c++ l++(+++) u+ e- m+ s+/+ n+(-) h- f+ !g w+ t+ r y*



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

From: feingold@avette.zko.dec.com (Elan Feingold)
Subject: For those lusting for threads :)
Date: 22 Jul 1994 17:51:19 GMT
Reply-To: feingold@avette.zko.dec.com (Elan Feingold)


I will be releasing a threads package soon, called EThreads, which 
formed the project for my OS systems class, but I enhanced and 
optimized.  It features:

        o Conditions as synchonization primitives
        o Preemptive context switching!
        o Portable, has been tested on Ultrix, Linux, SunOS
        o Code is aesthetically pleasing, with mundo comments :)

Stay tuned, or write me for more info...

elan

--
===============================================================================
|| Elan Feingold                  || "Two of the most famous products of     ||
|| Software Engineer II           ||  Berkeley are LSD and Unix. I don't     ||
|| Digital Equipment Corporation  ||  think that is a coincidence."          ||
|| Work: 603.881.1115             ||                       - Anonymous       ||
===============================================================================

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

Subject: Re: GOTO haters ..Re: Linux Performance Enhance ?
From: dholland@husc7.harvard.edu (David Holland)
Date: 22 Jul 94 13:24:38


bcr@k9.via.term.none's message of 21 Jul 94 15:03:58 GMT said:

 > Because a goto loop, or a while() loop should only be used when you
 > are willing to face the possability of an infinite loop.  (Very
 > rarely the case.)  Otherwise you should use a for() loop (or the
 > equivlent of it) so you have a counter to insure even if something
 > in your code is broken, eventually the program will exit the loop.

Huh?
        for (i=0; i<n; i++) { ... }
is identical to
        i=0;
        while (i<n) { ... i++; }
except for some fairly minor details with "continue".

--
   - David A. Holland          | "The right to be heard does not automatically
     dholland@husc.harvard.edu |  include the right to be taken seriously."

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

From: mjf@clark.net (Marc Fraioli)
Subject: Re: Anyone else lust for threads?
Date: 22 Jul 1994 18:24:34 GMT
Reply-To: mjf@clark.net

In article ctt@quartz.ucs.ualberta.ca,  heide@ee.ualberta.ca (Cameron N. Heide) writes:
>ddt@idcube.idsoftware.com (David Taylor) writes:
>
>>I may have missed something in the FAQ's, but assuming that there
>>is no thread support in Linux, are there plans to add or document
>>it?  I'm saddened by the fact that our next generation sound code
>>can be developed under NEXTSTEP, Irix, Solaris, etc., but apparently
>>not Linux because of a lack of threads.  Am I wrong?  *hope, hope*
>
>I've actually been hacking around with the BSD libpthread package
>with moderate success (the library compiles, the application links
>and runs, but it won't actually switch threads yet.  Grrr....).
>
There are apparently two pthreads libraries for Linux, both user-level.  I
have only used one of them.  Some ftp sites:

 sipb.mit.edu (directory unknown (to me))
 ftp.ibp.fr:/pub/unix/threads/pthreads
 
  a different implementation is at the same site under:

   /pub/unix/threads/part

There is also supposedly a project under way to add threads to the kernel.
I have a libpthreads for Linux, and it compiled fine, although I haven't
really gotten around to using it yet.  I forget where I got it from, but I
can put it up somewhere for ftp again if people can't find it.  Anyway, it
_is_ out there.

        Looking forward to a multi-threaded DOOM for Linux...;-)

---
Marc Fraioli          | "They couldn't hit an elephant at this dist- " 
mjf@clark.net         |    - Last words of Union General John Sedgwick,
                      |    Battle of Spotsylvania Court House, U.S. Civil War



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

From: niallp@ee.ubc.ca (niall parker)
Subject: Mode 2 PLIP ?
Date: Thu, 21 Jul 1994 22:16:52 GMT


I saw a query posted a while back about the existence of Mode 2
PLIP (the type using bidirectional printer ports, and transfering
a byte at a time), but no response.

Does such code exist for the new kernels? If so where ? The
README's that come with the net drivers imply it is supported,
but I can't find any reference in the current plip source code.


-- 
=======
Niall Parker                            niallp@ee.ubc.ca
UBC Electrical Engineering          or  VE7HEX@VE7UBC.#VANC.BC.CAN.NOAM

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

From: grante@reddwarf.rosemount.com (Grant Edwards)
Subject: Re: link() system call
Date: Fri, 22 Jul 1994 17:02:11 GMT

A N Burton (anb@doc.ic.ac.uk) wrote:

: Would anyone know if any of the file systems available under linux
: allow 'root' to create _HARD_ links to directories?

I once created a hard link from one directory to a "niece" directory
under SunOS.  This was a Very Bad Thing(tn) to do.  It took all day to
recover from that blunder.  I was not allowed to unlink() the link
(even as root).  I finally had to clear the i-node and run fsck about
six times.  Believe me, it's not something you do again.

There are quite a few programs in the Unix world (fsck for example)
that are based on the assumption of a tree structured file system
rather than a graph structured file system.  One big problem with a
graph structure is that it's damn hard to figure out when you're done
searching for something.

--
Grant Edwards                                 |Yow!  Hmmm..  A hash-singer
Rosemount Inc.                                |and a cross-eyed guy were
                                              |SLEEPING on a deserted
grante@rosemount.com                          |island, when...

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

From: emoenke@gwdu03.gwdg.de (Eberhard Moenkeberg)
Subject: Re: buffer cache or isofs bug?
Date: Fri, 22 Jul 1994 18:04:38 GMT

Eberhard Moenkeberg (emoenke@gwdu03.gwdg.de) wrote:

: mount -t iso9660 /dev/sbpcd2 /CD1

Sorry - since 1.1.33, you must not omit the "ro" option (crazy!), so read:

  mount -t iso9660 -o ro /dev/sbpcd2 /CD1

Cheers, -e-
-- 
Eberhard Moenkeberg                                GGG   W   W  DDDD    GGG
                                                  G      W   W  D   D  G
E-Mail:   emoenke@gwdg.de                         G GGG  W   W  D   D  G GGG
Phone: +49 551 201551  Fax: +49 551 21119         G   G  W W W  D   D  G   G
Mail:                                              GGG   WW WW  DDDD    GGG
Ges. fuer wissenschaftliche Datenverarbeitung mbH
Am Fassberg
D-37077 Goettingen           GERMANY
Private Email: Eberhard_Moenkeberg@rollo.central.de (FIDO: 2:2437/210.27)


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

From: bogdan@crl.com (Bogdan Urma)
Subject: New kernel message at boot time?
Date: 22 Jul 1994 11:37:16 -0700

   With the new kernel 1.1.33 I'm getting a new kernel message at boot time.

16384 bytes for swap cache allocated

   What does this mean??

Bogdan


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

Crossposted-To: comp.os.linux.help
From: tom@darshiva.efrei.fr (Thomas Parmelan)
Subject: Re: BUG in e2fs (0.5a)?  (INTERNAL ERROR)
Date: Thu, 21 Jul 1994 18:38:20 GMT

In article <30iont$ro9@fbi-news.informatik.uni-dortmund.de>,
Martin Loebbing <loebbing@ls2.informatik.uni-dortmund.de> wrote:
>
>I have problems with my harddisk drive (Quantum LPS540S, connected to a
>Adaptec 1542CF).
[ .../... ]
>On different runs of the fsck it reports different errors.

I had the same problem, with slightly the same configuration:
AMD486DX40, Adaptec 1542CF, Micropolis 1588.

The problem appeared with kernels from 1.0.9 to 1.1.12 (this was the
latest kernel at this time).

As far as I investigated, it seems that when the adapter BIOS is
enabled, the disk doesn't write correctly the datas, but doesn't return
any errors. I ran mke2fs and e2fsck many times and each time I got
errors. Then I tried the badblocks program, just in case...

The 'badblocks -w' (write-mode) test failed only when the Adaptec BIOS
was enabled!! When I disabled it (with the dip switches on the board)
the problem disappeared.

Note that I really mean disable the _BIOS_, not only the BIOS options as
it is written in the FAQ.

Hope this helps...

Tom.

-- 
Thomas Parmelan - tom@darshiva.efrei.fr

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

From: pme@gaia.electrum.kth.se (Peter Enderborg)
Subject: Whats wrong ? (New kernel bug??)
Date: 20 Jul 94 21:48:36

Have just upgrade to 1.1.31 from 1.1.29 and this is what ps
says:
USER       PID %CPU %MEM SIZE  RSS TTY STAT START   TIME COMMAND
root         1  0.0  2.6   48  192 con S    21:33   0:00 init auto
root        61  0.0 236375.9   37  -68 p 1 S    21:34   0:00 /sbin/agetty 38400 tty1
root        54  0.0 236375.5   60  -92 con S    21:34   0:00 (rpc.ugidd)
root        39  0.0  1.1   64   80 con S    21:34   0:00 /usr/sbin/rpc.rstatd
root        21  0.0 236376.5   28  -24 con S    21:34   0:00 (update)
root        22  0.0  0.0   28    0 con SW   21:34   0:00 update (bdflush)
root        34  0.0 236375.8   57  -76con S    21:34   0:00 /usr/sbin/syslogd
root        36 t 0.0 236375.3   36 -112 con S    21:34   0:00 (klogd)
root        38  0.0 236376.2   64  -44 con S    21:34   0:00 /usr/sbin/rpc.portmap
root        55  0.0  1.2   92   88 con S    21:34   0:00 /usr/sbin/rpc.mountd
root        41  0.0 236375.4   68 -100 con S    21:34   0:00 (inetd)
root        43  0.0 236375.9  105  -68 con S    21:34   0:00 /usr/sbin/named
root        45  0.0 236375.6   68  -88 con S    21:34   0:00 (lpd)
root        47  0.0  0.2   72   20 con S    21:34   0:00 /usr/sbin/crond
root        56  0.0 236374.6  116 -156 con S    21:34   0:00 /usr/sbin/rpc.nfsd
pme         62  0.1  1.9  348  140 p 2 S    21:34   0:00 -bash
root        58  0.0 236375.4   60 -100 con S    21:34   0:00 (rpc.pcnfsd)
root        63  0.0 236375.9   37  -64 p 3 S    21:34   0:00 /sbin/agetty 38400 tty3
root        64  0.0 236375.9   37  -68 p 4 S    21:34   0:00 /sbin/agetty 38400 tty4
root        65  0.0 236375.9   37  -64 p 5 S    21:34   0:00 /sbin/agetty 38400 tty5
root        66  0.0 236375.9   37  -64 p 6 S    21:34   0:00 /sbin/agetty 38400 tty6
pme         71  0.0 236375.8  332  -76 p 2 S    21:34   0:00 sh /usr/bin/X11/startx
pme         72  0.0  1.9   48  144 p 2 S    21:34   0:00 xinit /usr/X386/lib/X11/xinit/xinitrc 
root        73  5.8 13.0 1439  952 con S    21:34   0:21 X :0
pme         75  0.0  1.4  332  104 p 2 S    21:34   0:00 sh /usr/X386/lib/X11/xinit/xinitrc
root        77  0.1  8.9  244  652 p 2 S    21:34   0:00 xterm
pme         79  0.1  1.8  353  136 pp0 S    21:34   0:00 bash
root        80  0.1  5.0  293  364 pp0 S    21:35   0:00 xterm
pme         86  0.0  4.3  461  316 pp0 S    21:35   0:00 kermit
pme         88  0.0  4.3  461  316 pp0 S    21:37   0:00 kermit
pme         83  0.0  1.9  353  140 pp1 S    21:35   0:00 bash
pme         93  0.0  1.7   81  124 pp1 R    21:40   0:00 ps -aux
bash$ 




Think some processes use to much memory :-)
The system is from 1.2.0 with 1.1.31 kernel. Is it a bug or just my
binaries that now is to old...

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

From: gareth@metl.chi.il.us (Gareth Newfield)
Subject: Re: link() system call
Date: 22 Jul 1994 19:00:28 GMT

A N Burton (anb@doc.ic.ac.uk) wrote:
: Would anyone know if any of the file systems available under linux allow
: 'root' to create _HARD_ links to directories?

Last I checked they both the ln utility and the filesystem code did
not allow hard links to directories.

I have modified the extfs2 a couple different times to allow hard links,
it is very easy to do. Unfortunately it is really easy to screw things up.
Then the big problem is that fsck will go nuts and make things worse. You
also of course need to change the unlink function.
If you just need a link here or there why not use a symbolic link? If you
want to experiment with a graph like filesystem then you might consider
using a userfs. In any case I recommend that if you experiment with hard
links to directories that you do it in a separately mounted fs and not
one that has anything important.

Gareth ( who likes hard links to directories )


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

From: johnsonp@sscnet.ucla.edu (Phil Johnson)
Subject: Re: Xfree86: increase pallate?
Reply-To: johnsonp@nicco.sscnet.ucla.edu
Date: Fri, 22 Jul 1994 04:23:21 GMT

In article <30ne0u$6gm@nic.umass.edu>,
Christopher M. May <cmay@titan.ucs.umass.edu> wrote:
>Chris Csanady (ccsanady@iastate.edu) wrote:
>: I know that implementing true 24-bit graphics in Xfree would be quite
>: a task, but how about just increasing the size of the colormap for
>: now?  Maybe from 256 to 4096 or something.  Is this possible?
>
>: -chris(ccsanady@iastate.edu)
>: -- 
>
>I heard that millions of colors would arrive with XFree 3.1

Perhaps you are joking, but someone out there will believe you.  
XFree86[tm] 3.1 will not increase the colors available.

Phil Johnson

-- 
Phil Johnson                             Email: johnsonp@nicco.sscnet.ucla.edu


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

From: dmahle@MCS.COM (Dennis Mahle)
Crossposted-To: comp.os.linux.admin
Subject: Re: DOES ANYONE KNOW ABOUT THE MAILINGLISTS? (hello, is anyone there?)
Date: 22 Jul 1994 15:32:36 -0500

In article <30elg9$60e@archsrv.rz.unibw-muenchen.de>,
Frank Derichsweiler <i31ade@applsrv.rz.unibw-muenchen.de> wrote:
>thx1139@knuth.cba.csuohio.edu (tim werner) writes:

[snip]

>
>I am on the Linux mailing list. It works fine !!
>

Can someone please tell me how to get on the Linux
mailing list, or where the FAQ is or SOMETHING?

Thanks,
Dennis


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

From: dvlbin@cs.umu.se (Bj|rn Isaksson)
Subject: Re: 1.1.32 kernel broken?
Date: Sat, 23 Jul 1994 00:30:26 GMT

In article <30kjjn$gnq@watnews1.watson.ibm.com>,
Uri Blumenthal <uri@watson.ibm.com> wrote:
>Hi,
>       After installation of 1.1.32 I got lots of
>       these:
>
>/var/log/notice:
>Jul 20 18:54:39 angmar linux: del_timer() called with timer not initialized
>Jul 20 18:55:25 angmar linux: del_timer() called with timer not initialized
>Jul 20 18:57:26 angmar last message repeated 1635 times
>Jul 20 19:07:27 angmar last message repeated 1888 times
>
>       Comments? Fixes? 

Hi!

You may want to change the debug-mode "SLOW_BUT_DEBUGGING_TIMERS" in
the file linux/kernel/sched.c.

I had the messages going to the console instead of the log-file.
Things got sort of cluttered. :)

/Bjorn
--
<dvlbin@cs.umu.se>


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

From: dlogue@starbase.neosoft.com (Dan Logue)
Subject: Re: dot clocks
Date: 22 Jul 1994 23:01:38 GMT

Holger Burde (lucy@p207.informatik.Uni-Bremen.DE) wrote:
: hi,
: I have a S3 Video Card (Spea Mirage V7 (VLB) - S3-86C805) and used the
: dot clocks from the Accel File (XFree86 2.1.1). The Problem ... i can't get
: more than 800x600. The Server 'gives up' because the dot clocks provided 
: don"t match for a higher Resolution !

: Any 'dot clocks' or other help available on this ?????


I had the same trouble and I removed the line for clocks and instead
put in the following line

 Clocks "icd2061a"

This will use the programable clock generator and I have achieved 
1280x1024.

However, the Xserver reports maximum dot clock of 110 MHz and I
don't know the command to override this maximum.  My card will
do 135Mhz, and under MS windows will do an amazing 1408x1024 256colors.
My card is a Miro Crystal 20SD-VLB and my monitor is an IBM 17P.

Does anybody know the command to override the maximum dot clock
so I can use up to 135Mhz?  That is one of the settings in the 
Xconfig file.


Dan    dlogue@starbase.neosoft.com




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


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