Subject: Linux-Development Digest #759
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, 26 May 94 03:13:15 EDT

Linux-Development Digest #759, Volume #1         Thu, 26 May 94 03:13:15 EDT

Contents:
  Re: Standard I/O library inefficient? (Rob Janssen)
  8k nfs performance (Curtis Varner)
  Re: Skinny Dip (Rick Emerson)
  html version of linux 1.1 patches (Sam Shen)
  Re: CERT Advisory! - Linux & AIX !!!  (Juha Laiho)
  Re: enhanced IDE and SCSI-2 supported? (Allen Willson)
  Re: Using emx/gcc for OS/2 and Linux (Tony Coates)
  HELP!! - bad magic number in super-block. (Edward Vale)
  Re: SIGHUP - Deep Kernal Guts question! (Albert D. Cahalan)
  Re: enhanced IDE and SCSI-2 supported? (David C. Niemi)
  are sources to lockd available? (Carl G. Riches)
  Linux 1.1.13 compiling-->DDIOCSDBG undefined in sbpcd.c (Elaine Walton)
  Re: Linux/SCO compatibility? (Brandon S. Allbery)
  Re: SIGHUP - Where do we go from here? (Alan Cox)
  Re: Linux/SCO compatibility? (Alan Cox)
  Re: Virtual Consoles (Joey Gibson)
  Re: Virtual Consoles (Elaine Walton)
  Re: CORBA, OMG, Distributed Objects (Geoffrey Wyant - Sun Microsystems Labs BOS)

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

From: rob@pe1chl.ampr.org (Rob Janssen)
Subject: Re: Standard I/O library inefficient?
Reply-To: pe1chl@rabo.nl
Date: Thu, 19 May 1994 08:17:52 GMT

In <jcgreen.769315345@las1.iastate.edu> jcgreen@iastate.edu (Jon Green) writes:

>I've been reading through "Advanced Programming in the Unix Environment"
>by Stevens, and trying out various programs on my Linux system.  After
>much digging and reading, I still have a few questions.  I used the following
>program to time the results for reading a file with different buffer
>sizes:

[...]

>stream = stdin, line buffered, buffer size = 1024
>stream = stdout, line buffered, buffer size = 1024
>stream = stderr, unbuffered, buffer size = 1024
>stream = msdos file, fully buffered, buffer size = 2048
>stream = ext2fs file, fully buffered, buffer size = 1024

>The buffer size shown is taken from the st_bufsize member of the stat
>structure, and other calculations I've done seem to give the same answer.
>I am wondering why the optimal buffer size is not being used here.  Wouldn't
>using 8192 make standard I/O operations faster?

But the stat structure *does not have* an st_bufsize member!
No wonder, because it is not part of the (buffered) stdio library, but
an operating system level structure (at the level of open/read/write/close)

What the stat structure *does* have is an st_blksize member, which is
what you printed.  It is the blocksize, the unit of disk space allocation.
That is totally unrelated to the buffer size.

>I'm also wondering about the purpose of the _blksize member of the FILE
>structure, defined in <libio.h>.  This member always seems to be set to
>zero.. is it used for anything?

Hmm... you are not supposed to peek into the FILE structure. :-)
I would think it contains the value from st_blksize in the stat structure.
Check again after you have actually read something from the opened file,
as the values are only filled in at that time.

>I'm still new to programming under Unix, so if the above shows that I'm 
>really confused, I welcome anyone who wants to enlighten me. :)

Well, we all have been confused :-)
Only remember that this group is not really intended for discussions
about generic UNIX programming...

Rob
-- 
=========================================================================
| Rob Janssen                | AMPRnet:   rob@pe1chl.ampr.org           |
| e-mail: pe1chl@rabo.nl     | AX.25 BBS: PE1CHL@PI8UTR.#UTR.NLD.EU     |
=========================================================================

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

From: cvarner@corsa.ucr.edu (Curtis Varner)
Subject: 8k nfs performance
Date: 26 May 1994 00:24:40 GMT

Hello Everyone,

        I just compiled 1.1.13, the first kernel to have the 8k nfs
and I thought I would post a few timings I made.  A friend of mine,
Mike Griffith in the CS dept. here at UCR had related to me that
Sun-to-linux nfs had speeded up quite a bit with the 8k nfs, but did
not know what linux-to-linux performance increase, if any, there
would be. 
        I did two tests - nfs to a client on the same subnet, and nfs
to a client on a different subnet, and in both cases, the 8k nfs was
roughly twice as fast as 1k nfs.  Both machines must have the 8k nfs
in their kernel, by the way, if either of the kernels are old and
using the 1K nfs, it is the same as if both kernels used 1K nfs.

        Here are the timings:
        
Same Subnet:
                        Avg Throughput
        - 8K nfs        262.2 Kb/s
        - 1K nfs        145.7 Kb/s

Different Subnet:

        - 8K nfs        187.4 Kb/s
        - 1K nfs        100.9 Kb/s

        The way these experiments were run was by dd'ing a file from
an nfs-mounted directory (size = 1311748 bytes) sending the file to 
/dev/null, and timing how long it took to receive the file.  I am in
the process of updating the kernels on all the machines I am
responsible for to use 8K nfs.  Furthermore, I think 8k nfs should be
the default for mounting remote filesystems.  Good work Alan!

Curtis Varner
Bio-Ag Library, UC Riverside
cvarner@cs.ucr.edu

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

Subject: Re: Skinny Dip
From: rick.emerson@dscmail.com (Rick Emerson)
Date: Wed, 25 May 94 13:14:00 -0640

NI> [ad deleted]
NI> so i guess we are finally experiencing what is probably just the beginnin
NI> of endless tedious thousands of BULLSHIT advertisements...
NI> i guess we can do nothing but flame the hell out of these people until th
NI> disk drives explode.   on to battle!
NI> 

So why do you repeat the ad???  

Rick   
...
 * ATP/Linux 1.42 * If it's not worth doing, it's not worth doing well.



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

From: sls@mh1.lbl.gov (Sam Shen)
Subject: html version of linux 1.1 patches
Date: 26 May 1994 00:34:43 GMT

If you point your favorite WWW browser at:

        http://www.lbl.gov/~sls/linux-patches/

You will see a list of patches to Linux 1.1.  If you follow the link
you will see a list of files that have changes for each patch, and if
you follow those links you will see the diff.  Patches 1-15 are
online.

This version should be much faster than the previous version since all
the pages are generated in advance.

You can retrieve the tcl scripts I used to build this in

        http://www.lbl.gov/~sls/linux-patches/html_diffs.shar

Problems, questions, comments to me, Sam Shen (slshen@lbl.gov).

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

From: jlaiho@ichaos.nullnet.fi (Juha Laiho)
Subject: Re: CERT Advisory! - Linux & AIX !!! 
Date: Wed, 25 May 1994 06:00:33 GMT

jasonh@chineham.euro.csg.mot.com (Jason Haar) said:
>** NOTE!! **
>The security.tgz file will only fix telnetd/rlogin (as well as lp*).
>THERE IS NO PATCH THERE FOR THE REAL PROBLEM - NAMELY /bin/login.

As far as I can understand, the 'login' in util-linux-1.6 -package has
this problem corrected.
-- 
Wolf  a.k.a.  Juha Laiho     Helsinki, Finland
(Geek Code 1.0.1) GCS d? p c++ l++ u(-) e+ m+ s+/- n- h(*) f(?) !g w+ t- r y+
"...cancel my subscription to the resurrection!" (Jim Morrison)

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

From: allenw@bnr.ca (Allen Willson)
Subject: Re: enhanced IDE and SCSI-2 supported?
Date: 25 May 1994 14:52:42 -0500

In article <1994May25.110550.3486@tudedv.et.tudelft.nl>,
 <lcvanveen@et.tudelft.nl> wrote:
>In article <CqCFpC.CC7@mercury.wright.edu>, eatkins@alpha.wright.edu (Elizabeth Atkins) writes:
>> 
>>      Just out of curiosity.....are the "enhanced IDE" and SCSI-2 
>> supported already, or are there going to have to be new drivers for 
>> them.  I'm going to be purchasing a couple systems in the near future
>> for Linux specific work.  Also, I needed to know if these DX4/100 will
>> work OK with Linux.  I have heard that Linux does fine with minor
>> modifications on the Pentium and was wondering what the minor mods are 
>> and where I can get them.
>> 
>If you mean burst-mode IDE support then the answer is a solid YES.
>There is a IDE-patch that has been posted in the news-groups.
>
>Martijn.

Burst mode for IDE is supported, but. . .

I think Ms Atkins was referring to the new IDE standard - called
enhanced IDE.  It is supposed to be very competitive with SCSI-2
for throughput, up to 12 Mbytes/s is the highest rating I've seen.
I do not believe there is kernel support for it yet, as it is 
relatively new.  Vendors such as Maxtor, Conner, Western Digital,
and Seagate were supposed to be shipping late Spring, 1994.
See the following publications for details:
        BYTE Magazine, March 94
        Windows Sources, Jan 94, May 94
        PC Magazine, Nov 93
        Computer Shopper, May 94
        PC Week, Apr 94

'Sponsor' me with one of these disks and I'd be happy to write
the driver for you !! 8^)

Allen

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

From: coates@kelvin.physics.uq.oz.au (Tony Coates)
Crossposted-To: comp.os.os2.programmer.porting
Subject: Re: Using emx/gcc for OS/2 and Linux
Date: 25 May 94 17:43:36
Reply-To: coates@physics.uq.oz.au

In article <1994May21.212900.23676@midway.uchicago.edu> roe2@ellis.uchicago.edu (Cave Newt) writes:

   coates@physics.uq.zo.au writes:

   >    What would be really interesting, if it were possible, would
   >be files which were dynamically linked to equivalent shared/DLL
   >libraries under Linux and OS/2 such that the same executable file
   >would run under either Linux or OS/2, simply choosing the correct
   >shareable/dynamic library for the correct operating system.  For
   >anyone running both systems on the one machine, this would be a great
   >way to save space, I suspect, having all tools on a common drive
   >readable by both.

   Ooo, I'd really love that, but I suspect it isn't (currently) possible.
   I don't know anything about the Linux executable format, but OS/2's 
   begins with a DOS stub (the old "MZ" thing), and I doubt that Linux 
   understands that outside of dosemu.  Of course, there's no reason why 
   Linux couldn't be *made* to understand it, but I assume we're talking
   about a kernel hack at this point (or would it be a shell hack?).  I
   wonder if Linus would be receptive (always assuming that there are no
   problems with required static library routines)...

As to questions of Linux recognising the DOS stub, I know nothing of
the OS/2 or Linux executable file formats, but I assume that Linux,
like UNIX, uses a magic number to recognise executable file formats.
Thus if the OS/2 and Linux executable file formats aren't too
different (?) it might be possible to define a new magic number, which
would need to be imbedded in the DOS stub, defining a `Linux/OS/2'
merged executable.  It might be harder to make OS/2 understand any
sort of change in the stub, but I don't know at all what is possible
there.  If anyone can comment on whether what I am suggesting is
possible, or whether the two file formats are really too incompatible,
I would be interested to hear.
        Thanks in advance,
                                                Cheers,
                                                                Tony.
--
_____________________________________________________________________________
A.B.Coates, Dept. of Physics,
The University of Queensland  QLD  4072  Australia.
Email: coates@physics.uq.oz.au
Phone: (07/+617) 365-3424  Fax: (07/+617) 365-1242
Disclaimer: The University is ignorant of my
            opinions, let alone guilty ...
_____________________________________________________________________________

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

Crossposted-To: comp.os.linux.help
From: t46012@isdserv.dehavilland.ca (Edward Vale)
Subject: HELP!! - bad magic number in super-block.
Date: Wed, 25 May 1994 13:20:04 GMT

Got the following error message today:

   e2fsck 0.4a, 93/11/29 for EXT2 FS 0.4a 93/11/19
   /dev/hda1 is clean, no check
   e2fsck 0.4a, 93/11/29 for EXT2 FS 0.4a 93/11/19
   fsck.ext2: bad magic number in super-block
   Fsck failed. Please reboot

There are two partitions on the IDE drive:
  
   /dev/hda1    linux native 83
   /dev/hda2    linux swap 82 

I've tried reconstructing the filesystem from backups but I still get this
error.  Switching to a new hard disk didn't help.
I'd really appreciate ANY help.

ADVthanksANCE,
--
Edward Vale                                     INTERNET: evale@dehavilland.ca
Software Engineer, CAD/CAM Technology
de Havilland Inc.                               OFFICE:   (416) 375-3084
Garratt Boulevard MS N02-10                     FAX:      (416) 375-4511
Downsview, Ontario
CANADA M3K 1Y5

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

From: adc@bach.coe.neu.edu (Albert D. Cahalan)
Subject: Re: SIGHUP - Deep Kernal Guts question!
Date: 26 May 1994 00:55:11 GMT

In article <2ru7kj$5ua@spool.cs.wisc.edu> jimr@shorty.cs.wisc.edu (Jim Robinson) writes:
   silly to put in portable code for every non-posix system unless you
   know for sure that there is a strong interest in it running on a
   non-posix compilient system.  Again, I had been under the impression
   that almost all major systems have, or are trying to have, posix
   compilence.  The point was made in an earlier post that if you know

Some major systems: M$-DO$, $y$tem 7, O$/2    :)
--

Albert Cahalan
adc@meceng.coe.neu.edu

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

From: niemidc@pep.production (David C. Niemi)
Subject: Re: enhanced IDE and SCSI-2 supported?
Date: 25 May 1994 20:17:42 GMT
Reply-To: niemidc@pep.production

In article 2rvml5INNjed@CS.UTK.EDU, peek@duncan.cs.utk.edu (Michael Peek) writes:
...
>How about the new IDE's that support 4 drives?  Does anyone know if/when
>they will be supported?

Dunno about the *new* IDE controllers that support 4 drives.  I can vouch
for the at2 patches working fine on a an *old* IDE controller that supports
4 drives (an EISA DTC 2290).  But the new stuff probably means that there
will someday be BIOS info on the 3rd and 4th drives, rather than having
to manually enter the geometry and recompile itas it is today.
---
David C. Niemi

supervisor, Software Engineering Environment
GTE Federal Systems Division
15000 Conference Center Drive                 (703) 818-4203
Chantilly, VA 22021                   David.Niemi@oasis.gtegsc.com




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

From: cgr@poplar1.cfr.washington.edu (Carl G. Riches)
Crossposted-To: comp.protocols.nfs,comp.os.linux.admin
Subject: are sources to lockd available?
Date: 25 May 1994 19:25:17 GMT
Reply-To: cgr@poplar1.cfr.washington.edu

i have been planning to set up a pc network using nfs.  having a limited
budget, i had hoped to use linux as the nfs server.  i've just learned that
linux does not have lockd.  is this true?  and if so, are sources for lockd
available.

please e-mail your replies.  i will summarize and repost them.

thanks!

carl

---
carl g. riches
college of forest resources         internet: cgr@poplar1.cfr.washington.edu
university of washington ar-10      voice:    206-543-2764
seattle, wa 98195                   fax:      206-543-3254



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

From: ewalton@magnus.acs.ohio-state.edu (Elaine Walton)
Subject: Linux 1.1.13 compiling-->DDIOCSDBG undefined in sbpcd.c
Date: 25 May 1994 22:16:02 GMT


I downloaded the Linux 1.1.13 kernel from tsx-11 last night and tried to 
compile it just now.  I get an error DDIOCSDBG not defined.  I searched all 
over the place and could not find it in my new or old code.  Can someone please
tell me what its value should be?  Also what else might I be missing?

THX!
-Sean

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

From: bsa@kf8nh.wariat.org (Brandon S. Allbery)
Subject: Re: Linux/SCO compatibility?
Date: Wed, 25 May 1994 21:52:09 GMT

In article <d00n.769884288@crash.cts.com>, d00n@crash.cts.com (Kevin Spousta) says:
+---------------
| Does Linux support the SCO filesystem?  I am able to do a TAR -xvf on an SCO
| created floppy disk..
+------------->8

AFS and EAFS are not supported due to insufficient information about the
format.  If you make an S51k filesystem under SCO current kernels can mount
it.

| Also, will binaries from other Unix systems run under Linux?  i.e SCO?  BSD? 
| SunOS?
+------------->8

SunOS???  What makes you think a Sparc binary would work?

Many SCO and SVR4 binaries work using the ALPHA iBCS-2 emulator available on
tsx-11.mit.edu.  But Solaris 2/x86 binaries have not, to my knowledge, been
tested with it yet (for that matter, nobody seems to know if Solaris/x86
binaries work under UnixWare... see thread in comp.unix.solaris).  I don't
think anyone's written an x86 BSD emulator yet, but there's no reason in
principle why it can't be done the same way iBCS-2 support is.

++Brandon
-- 
Brandon S. Allbery         kf8nh@kf8nh.ampr.org          bsa@kf8nh.wariat.org
The FUDs at Microsoft are shouting "Kill The Wabi!"

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

From: iiitac@uk.ac.swan.pyr (Alan Cox)
Subject: Re: SIGHUP - Where do we go from here?
Date: Wed, 25 May 1994 19:06:51 GMT

In article <1994May25.135713.31927@loreli.ftl.fl.us> nate@loreli.ftl.fl.us (Sean Puckett) writes:
>Because of these facts:
>
>  1.  The kernel baseline will not change.  (The people who believe in
>      100% Posix compliance are the people who control the baseline).
>  2.  People will continue to have problems with non-compliant
>      applications.

All that is needed is to fix the shell programs that are managing the
process groups and the SIGHUP - no other program is affected.

Alan


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

From: iiitac@uk.ac.swan.pyr (Alan Cox)
Subject: Re: Linux/SCO compatibility?
Date: Wed, 25 May 1994 19:27:35 GMT

In article <d00n.769884288@crash.cts.com> d00n@crash.cts.com (Kevin Spousta) writes:
>Does Linux support the SCO filesystem?  I am able to do a TAR -xvf on an SCO
>created floppy disk..
Yep.
>Also, will binaries from other Unix systems run under Linux?  i.e SCO?  BSD? 
>SunOS?
With the iBCS module loaded yes to SCO (on the whole) yes to most Solaris
but not SunOS or BSD.

Alan


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

Subject: Re: Virtual Consoles
From: wjg@Creeper.Atl.GA.US (Joey Gibson)
Date: Thu, 26 May 1994 02:09:13 GMT

Stuart Herbert (ac3slh@sunc.sheffield.ac.uk) wrote:
: Apologies if this is to the wrong group :) but I'm looking for instructions
: on how to increase the number of virtual consoles I have available.

: Ideally, I'd like to have up to 24 :) (ALT F1-F12, and SHIFT-ALT F1-F12 if
: possible).  I've located the #define for setting the number of consoles,
: but I don't know what else I need to do to make it recognise the keypresses.

After reading this, I rebuilt my kernel, defining 24 VC in
include/linux/tty.h. To access the second 12 VCs, you use the *right*
ALT key, thus VC13 is accessed by RIGHT-ALT-F1, etc. Works great.

Joey

-- 
"Never trust a man in a blue trenchcoat, never drive | wjg@Creeper.Atl.GA.US
 a car when you're dead." - Tom Waits                | wjg@aix2.EMA.com
"They got a lot of coffee in Brazil" - Frank Sinatra | PGP public key
____________________________________________________ |  available on request


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

From: ewalton@magnus.acs.ohio-state.edu (Elaine Walton)
Subject: Re: Virtual Consoles
Date: 26 May 1994 01:10:53 GMT

In article <2rvncp$om3@uuneo.neosoft.com>,
John Lellis <lellis@dmccorp.com> wrote:
>Stuart Herbert (ac3slh@sunc.sheffield.ac.uk) wrote:
>: Ideally, I'd like to have up to 24 :) (ALT F1-F12, and SHIFT-ALT F1-F12 if
>: possible).  I've located the #define for setting the number of consoles,
>: but I don't know what else I need to do to make it recognise the keypresses.
>
>: Stuart
>
>to whatever number you want.  In my case, I made it 12.
>
>John Lellis (lellis@dmccorp.com)

I can understand his question: will Linux recognize that when I set this 
#define to a number >12 will it know to accept the ctrl-alt-F1...?  (Or is it 
shift-alt-numlock-capslock-F1?)

Seriously, how will Linux map the keys?  Also, I know how to go from Xterm to 
term using ctrl-alt-Fn.  But, how does one go back to the Xterm?

-Sean

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

From: gwyant@cloyd.east.sun.com (Geoffrey Wyant - Sun Microsystems Labs BOS)
Subject: Re: CORBA, OMG, Distributed Objects
Date: 25 May 1994 17:09:31 GMT



   From: krauss@charlie.igd.fhg.de (Jens Krauss)
   Newsgroups: comp.os.linux.development
   Date: 25 May 1994 11:20:42 GMT
   Organization: Haus der Graphischen Datenverarbeitung, 64283 Darmstadt, Germany
   Reply-To: igd.fhg.de
   Keywords: CORBA, OMG, Distributed Objects


   Hy all there!

   I'm very interested in distributed objects, perhaps I would like to 
   implement something like that. If sombody out there is working on
   dostributed objects for linux, mail me.
   If nobody is out there, is there interest for an "corba"???

   The problem is, that I have no acces to the OMG documents. But I have some 
   ideas how to implement such thing. But whats about the standards...???
   Would it be a good idea to programm such thing???

   If there is no interest, perhaps I#ll doing it for my own, without standards!!

   Ciao Jens

Well...if you don't mind venturing outside the mainstream,
there is a very elegant distributed object system already
available for Linux. It's the Modula-3 programming system
from DEC Systems Research Center. One of the components is
a library and toolkit called 'Network Objects' that allows
cross process/cross-machine invocation of methods on Modula-3
objects. It also provides a robust distributed garbage collector.

I'll be happy to provide more information if you are interested.

--geoff


-- 
Geoff Wyant
Geoff.Wyant@east.sun.com
Sun Microsystems Laboratories, Inc.

2 Elizabeth Drive
Chelmsford, Ma.
01824

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


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