752 lines
29 KiB
Plaintext
752 lines
29 KiB
Plaintext
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: Sat, 3 Sep 94 01:13:07 EDT
|
|
Subject: Linux-Development Digest #108
|
|
|
|
Linux-Development Digest #108, Volume #2 Sat, 3 Sep 94 01:13:07 EDT
|
|
|
|
Contents:
|
|
Re: how to do shared C libraries (was Re: nvi 1.34, curses and the new Linux C library) (David Engel)
|
|
Re: Pentium version of gcc? (H.J. Lu)
|
|
Re: Future of Linux (Mark A. Horton KA4YBR)
|
|
WD 90C33 and XFree86 (Alfred Hovdestad)
|
|
mmap( ..., PROT_WRITE, MAP_SHARED, ... ) WHY NOT? (Russell Leighton)
|
|
tar to st0 (Exabyte tape) causes kernel panic (Russell Leighton)
|
|
BUG in tcl/tk for Linux ?? (Norbert Kuemin)
|
|
Re: s3 964 server ? (Harald Koenig)
|
|
Re: Future of linux -- the sequel (Tor Arntsen)
|
|
Re: PCI ethernet (Donald Becker)
|
|
Re: Future of Linux (Mike Dowling)
|
|
Re: Linux Inside T-Shirts, Now Printing! (Dave Rossow)
|
|
Re: PGP Signature (Was: Suggest:SCSI Tape File System) (Darin Johnson)
|
|
Re: ext2fs corruption in 1.1.47-48 (Mark P. Nelson)
|
|
Re: What is happening with MGETTY 0.2 (Harald Milz)
|
|
Re: VLB IDE - DOS speeds up but Linux does not (Donald Becker)
|
|
Re: Future of linux -- the sequel (Bill Broadley)
|
|
Re: PRIORITY make an undelete command (Grant Edwards)
|
|
Re: Netware Client (Sander Plomp)
|
|
Re: PRIORITY make an undelete command (Remco Treffkorn)
|
|
|
|
----------------------------------------------------------------------------
|
|
|
|
From: david@ods.com (David Engel)
|
|
Subject: Re: how to do shared C libraries (was Re: nvi 1.34, curses and the new Linux C library)
|
|
Date: Fri, 2 Sep 1994 15:22:57 GMT
|
|
|
|
Davor Jadrijevic (davj@ds5000.irb.hr) wrote:
|
|
: it. For Linux, running ldconfig requires root privilegies.
|
|
: Is this necessary; can we have local ldconfig? I'd like to have
|
|
: this feature.
|
|
|
|
Try reading the man page, specifically, the -n option.
|
|
|
|
David
|
|
--
|
|
David Engel Optical Data Systems, Inc.
|
|
david@ods.com 1101 E. Arapaho Road
|
|
(214) 234-6400 Richardson, TX 75081
|
|
|
|
------------------------------
|
|
|
|
From: hjl@nynexst.com (H.J. Lu)
|
|
Subject: Re: Pentium version of gcc?
|
|
Date: 1 Sep 1994 02:18:28 GMT
|
|
|
|
M. Moeling (martijn@arbor.gds.nl) wrote:
|
|
: Well yesterday I,ve got my P66 too !!!
|
|
|
|
: of a pentium. GCC optimized for it would be the first thing i guess ?
|
|
|
|
: If someone out the 'NET' does find any working optimized gcc somewhere anytime
|
|
: Pass it on ! (please)
|
|
|
|
: martijn@arbor.gds.nl
|
|
|
|
|
|
One thing you can to do is the join the GCC channel. The pentium
|
|
support is on the way. If there is a big demand, there will be a
|
|
pentium binary snapshot along with the x86 version. But you have to
|
|
have lots of disk spaces since you have to keep the older versions
|
|
around in case of the newer one doesn't work, which is always true.
|
|
Currently I have 4 different versions of gcc on my notebook. Also
|
|
the new snapshot is made every week. You may have to ftp about 2 MB
|
|
bytes of gzipped files every week. The installation should be the same
|
|
as my public gcc binary release.
|
|
|
|
|
|
H.J.
|
|
--
|
|
First I thought he was on hunger strike. Later I was told he was
|
|
praticing YanXin QiGong.
|
|
|
|
------------------------------
|
|
|
|
From: mah@ka4ybr.com (Mark A. Horton KA4YBR)
|
|
Subject: Re: Future of Linux
|
|
Date: Fri, 2 Sep 1994 12:26:29 GMT
|
|
|
|
Mark Bolzern (mark@gcs.com) wrote:
|
|
|
|
: How about supporting & promoting them after they have? We have... now the
|
|
: World needs to know about it.....
|
|
|
|
: FlagShip sells to commercial Unix users for US$2800 to 5700, and has been
|
|
: made available for Linux at a special offer price less than a 10th of that.
|
|
: Check out our URL ftp://ftp.wgs.com/pub2/wgs/Filelist
|
|
: There is even a Free functional demo there.
|
|
: --
|
|
: Mark Bolzern : mark@gcs.com USA Tel: (303) 699-7470 Fax: (303) 699-2793
|
|
: WorkGroup Solutions, Inc. The FlagShip "CA-Clipper and XBase on Unix" People
|
|
: FlagShip is a 4GL Database Development System & XBase Porting Tool for Unix
|
|
: No Runtime Fees Info at ftp.wgs.com : /pub2/wgs/Filelist OR mail: info@wgs.com
|
|
|
|
|
|
Guys, I can't sing the praises of this product enough! Think about the THOUSANDS of
|
|
applications (read $$$$) already out there for xBASE and Clipper.... This lets you
|
|
port them to Linux _NATIVE_ (!!!) and make available real commericial solutions!
|
|
(Like a Linux box running some dumb terminals off a multi-port card or two... an
|
|
easy sell to the guy contemplating Novell plus the cost of a bunch of PCs and wiring
|
|
and all the other mess that goes with Netware!) The special introductory pricing
|
|
for the Linux community makes this one hell of a good deal (as low as $199.00;
|
|
$499.00 for an unlimited development/usage license!) NO runtime modules! Get the
|
|
demo, try it, port some stuff to it, see if you like it! I don't want to hear any
|
|
arguments about it not being "free" software... everything people are mentioning
|
|
on this thread involves COMMERCIAL products that sell for the big bucks... you want
|
|
commercial grade stuff, you pay for it... you want freebies, write it yourself or
|
|
put up with what you've got... ( I personally like troff for my word processing and
|
|
sc for my spreadsheet! :) ... Linux IS becoming mainstream and with that is
|
|
attracting the attention of commercial developers who are, after all, in it for
|
|
the money. Get used to it! Even better, get involved in it! If this one goes
|
|
over well, you can bet we'll start seeing a lot of other commercial stuff coming
|
|
over to Linux!
|
|
|
|
-- m
|
|
|
|
p.s. The 1200+ page manual is GREAT! It's available online, of course, but it's
|
|
well done, in a nice looseleaf binder, and besides, I LIKE being able to get out
|
|
of the computer room, sit in an easy chair, and read the docs. It's well worth
|
|
the hundred bucks extra... think how much the toner and paper would cost you!
|
|
|
|
--
|
|
"Linux! Guerrilla UNIX Development Venimus, Vidimus, Dolavimus."
|
|
============================================================
|
|
Mark A. Horton ka4ybr mah@ka4ybr.atl.ga.us
|
|
P.O. Box 747 Decatur GA US 30031-0747 mah@ka4ybr.com
|
|
+1.404.371.0291 33 45 31 N / 084 16 59 W
|
|
|
|
------------------------------
|
|
|
|
From: hovdesta@teapot.usask.ca (Alfred Hovdestad)
|
|
Crossposted-To: comp.os.linux.admin,comp.os.linux.help
|
|
Subject: WD 90C33 and XFree86
|
|
Date: 31 Aug 1994 19:49:22 GMT
|
|
|
|
Does anyone have XFree86 working with a WD90C33? I have a 486dx2/50
|
|
with a Rocket MultiFunction adapter and a WD90C33 video chip. I have
|
|
Linux installed by I can't get X to run.
|
|
|
|
I am sure that someone must have this working (or at least be working
|
|
on it). After all, if someone can get the Diamond Stealth to work,
|
|
then the WD90C33 should not be too difficult.
|
|
|
|
If no one is working on this, can someone give me an idea where to
|
|
start?
|
|
|
|
--
|
|
Alfred Hovdestad |e-mail: hovdesta@herald.usask.ca
|
|
Systems Programmer | or: Alfred.Hovdestad@usask.ca
|
|
Department of Computing Services | Voice: (306) 966-4819
|
|
University of Saskatchewan | FAX: (306) 966-4938
|
|
|
|
------------------------------
|
|
|
|
From: rrl@access3.digex.net (Russell Leighton)
|
|
Subject: mmap( ..., PROT_WRITE, MAP_SHARED, ... ) WHY NOT?
|
|
Date: 1 Sep 1994 15:45:11 -0400
|
|
|
|
Writing to mmap'd files does not seem to be supported under Linux:
|
|
|
|
mmap( ..., PROT_WRITE, MAP_SHARED, ... )
|
|
|
|
This seems to be a MAJOR limitation, since I am new to Linux, is this
|
|
right? If so, how to fix it?
|
|
|
|
Russ
|
|
|
|
--
|
|
Russell Leighton
|
|
Taylor Computing
|
|
russ@taylor.digex.net taylor@world.std.com
|
|
http://taylor.digex.net http://www.digex.net/~rrl/Welcome.html
|
|
|
|
------------------------------
|
|
|
|
From: rrl@access3.digex.net (Russell Leighton)
|
|
Subject: tar to st0 (Exabyte tape) causes kernel panic
|
|
Date: 1 Sep 1994 15:48:43 -0400
|
|
|
|
|
|
I am able to read from st0 (Exabyte 8200) but
|
|
when I write, the kernel panics...is this a known bug? ...
|
|
what to do? (I am new to Linux ~1day experience).
|
|
|
|
Russ
|
|
|
|
--
|
|
Russell Leighton
|
|
Taylor Computing
|
|
russ@taylor.digex.net taylor@world.std.com
|
|
http://taylor.digex.net http://www.digex.net/~rrl/Welcome.html
|
|
|
|
------------------------------
|
|
|
|
From: kuemin@srapc101.alcatel.ch (Norbert Kuemin)
|
|
Subject: BUG in tcl/tk for Linux ??
|
|
Date: 1 Sep 1994 07:16:57 GMT
|
|
Reply-To: norbert.kuemin@alcatel.ch
|
|
|
|
I've posted following article and noone could me give an answer :
|
|
====== Article postet in comp.lang.tcl =====================================
|
|
Hi all, I've included following line to my tcl application to count entries
|
|
in a file :
|
|
puts [exec cut -b5-30 $PH_DATA | grep -i $SEARCH | wc -l 2>/dev/null]
|
|
|
|
This is what i see:
|
|
1 error waiting for process to exit: No child processeserror waiting for
|
|
process to exit:No child processeserror waiting for process to exit: No
|
|
child processes
|
|
|
|
It counts correctly, but what about the messages about no chield process??
|
|
Thanks for any hints
|
|
===========================================================================
|
|
|
|
I get this errormessage(error waiting for process to exit: No child processe)
|
|
also using tkmail-1.6p5. Seams to be a linux related problem !!!
|
|
|
|
Thanks Norbert Kuemin
|
|
|
|
------------------------------
|
|
|
|
From: koenig@nova.tat.physik.uni-tuebingen.de (Harald Koenig)
|
|
Subject: Re: s3 964 server ?
|
|
Date: 2 Sep 94 15:03:41 GMT
|
|
|
|
In <cschwarz.2.2E6249FA@nyx.cs.du.edu> cschwarz@nyx.cs.du.edu (Chris Schwarzfischer) writes:
|
|
|
|
>we got a pentium 90MHz ..... (fine).
|
|
>we got a S3 card with 964 chipset (PCI) (fast but do not work with the S3
|
|
>X-Server)
|
|
|
|
>where can i find the newest server?
|
|
|
|
wrong question!
|
|
this should read: WHEN can i find the newest server (if you are interessted in
|
|
964 support).
|
|
|
|
answer for this question:
|
|
mid/end of september
|
|
|
|
Watch for the "XFree86 3.1" announcement. I'm shure it will be available
|
|
on your favourite Linux anon.ftp server WHEN it is released.
|
|
|
|
Harald
|
|
--
|
|
Harald Koenig, koenig@tat.physik.uni-tuebingen.de
|
|
_____ _ __ _____ ___ __ ___ _ _
|
|
|_ _| |_ ___ \ \/ / __| _ ___ ___( _ )/ /TM | _ \_ _ ___ (_)___ __| |_
|
|
| | | ' \/ -_) > <| _| '_/ -_) -_) _ \ '_ \ | _/ '_/ _ \| / -_) _| _|
|
|
|_| |_||_\___| /_/\_\_||_| \___\___\___/\___/ |_| |_| \___// \___\__|\__|
|
|
|
|
------------------------------
|
|
|
|
From: tor@spacetec.no (Tor Arntsen)
|
|
Subject: Re: Future of linux -- the sequel
|
|
Date: 2 Sep 1994 13:14:41 GMT
|
|
Reply-To: tor@spacetec.no
|
|
|
|
In article 12964@philabs.philips.com, nrh@philabs.philips.com (Nikolaus R. Haus) writes:
|
|
>>You can get a VL bus motherboard with MIPS R4600 processor
|
|
>>that makes Pentium look like a 4.77 8086.
|
|
>
|
|
>Can I play DOOM on it? ;)
|
|
|
|
Maybe!
|
|
|
|
DOOM runs extremely well (i.e. *fast*) on a nearby Indy R4600. That's IRIX
|
|
of course.
|
|
|
|
Tor
|
|
|
|
|
|
|
|
------------------------------
|
|
|
|
From: becker@cesdis.gsfc.nasa.gov (Donald Becker)
|
|
Subject: Re: PCI ethernet
|
|
Date: 2 Sep 1994 11:39:53 -0400
|
|
|
|
In article <33llhq$hgc@freeside.fc.net>,
|
|
Don Carroll <root@toad-temp.comland.com> wrote:
|
|
>
|
|
>Anyone Working on drivers for the Mitron PCI ethiernet card
|
|
>they said that someone was . I would be interest in testing the driver for
|
|
|
|
What chip does that board use? If it's an AMD 79C97* chip, it's good news.
|
|
A DEC 21040 might eventually be supported as well.
|
|
|
|
The updated LANCE driver (probably in 1.1.50) works around a minor bug (1) in
|
|
the PCnet32 implementation, and doesn't allocate a DMA channel for PCI and
|
|
VL boards with native bus mastering. I've tested the new driver on the VL
|
|
bus version, but the PCI version should now work as well (3).
|
|
|
|
I'm writing a PCnet32 driver to use the VL and PCI chips in 32 bit
|
|
addressing mode, but I don't expect the performance gain will be significant
|
|
with 10Mbs networks and <=16M machines. The major advantage might be that
|
|
the PCnet32 driver will be processor independent.
|
|
|
|
If I get a DEC ethernet board, I might also do a driver for the PCI 21040.
|
|
That driver will also be processor independent.
|
|
|
|
(1) FYI, the PCnet32 bug is that the Rx buffer size in the RMD is cleared,
|
|
despite the documentation clearly stating that it's left untouched. The
|
|
symptom is that the only one ring - 16 packets - can be received before the
|
|
chip runs out of the Rx resources. The trivial work-around is to always
|
|
rewrite the buffer size entry when clearing out the buffer.
|
|
|
|
(2) Boca 10baseT boards using the AMD79C965 chip are only $70-$90! We
|
|
bought ours from Aurora, but CMO lists them for $67. Both advertise in
|
|
Computer Shopper. I expect the PCI boards to be the same or only slightly
|
|
more expensive.
|
|
|
|
(3) Frank Koeck <koeck@goofy.mpi-hd.mpg.de> reported that the PCI boards
|
|
did work for 16 packets with the old driver. The work-around described above
|
|
should fix the problem.
|
|
--
|
|
Donald Becker becker@cesdis.gsfc.nasa.gov
|
|
USRA-CESDIS, Center of Excellence in Space Data and Information Sciences.
|
|
Code 930.5, Goddard Space Flight Center, Greenbelt, MD. 20771
|
|
301-286-0882 http://cesdis.gsfc.nasa.gov/pub/people/becker/whoiam.html
|
|
|
|
------------------------------
|
|
|
|
From: mike@moocow.math.nat.tu-bs.de (Mike Dowling)
|
|
Subject: Re: Future of Linux
|
|
Date: 02 Sep 1994 17:52:06 GMT
|
|
Reply-To: on.dowling@zib-berlin.de
|
|
|
|
>>> On Wed, 31 Aug 94 06:15:12 GMT, mdc@cdi.tiac.net (Marty Connor) said:
|
|
|
|
Marty> In Article <msuzio.778283297@tiamat.umd.umich.edu>,
|
|
Marty> msuzio@tiamat.umd.umich.edu (Mike Suzio) wrote:
|
|
Marty> 1. You folks who tolerate_ Mac users could learn a lot from being
|
|
Marty> forced to _use_ a Mac for a while. I use Macs, Windows boxes, OS2
|
|
Marty> boxes, and now Linux boxes. Guess which one is easiest to teach users
|
|
Marty> how to do things on? Guess which ones have word processing you don't
|
|
Marty> have to be a programmer to use? Guess which ones developers write apps
|
|
Marty> for?
|
|
|
|
|
|
I think we're all missing the point. Firstly, nobody will complain if somebody
|
|
goes to the trouble and effort to create all these GUS and things for Linux.
|
|
It would be good, because, as many have said, it will give Linux a wider
|
|
appeal. If you don't want these things, then don't install them.
|
|
|
|
The deeper issue is that Linux is free. If I am correct in my analysis, nobody
|
|
out there who is developing programs for nothing is doing so as a sort of
|
|
charity. They are producing what they see as a quality product, and are
|
|
sharing it with others, firstly, because they hope that others will do the
|
|
same, and we would thereby all benefit, and, secondly, because they are proud
|
|
of their craftsmanship. IBM, Microsoft, etc. all ask themselves the following
|
|
questions. What do the customers want? What are their applications? How well
|
|
educated are they when it comes to computers? What is the most aggressive
|
|
marketing strategy?
|
|
|
|
Once freed of the commercial considerations, though, people ask quite different
|
|
questions. What do I think is a good user interface? What do I think would be
|
|
the most desirable features for the program? These people are all expert
|
|
computer programmers, so what they invariably think is good is a product that
|
|
is very functional, and, with a minimal number of key strokes, can achieve
|
|
great power. In short, IBM, Microsoft, Apple, etc. will find that they are
|
|
catering for people who know nothing about computers, so that their main
|
|
objective is to lead the user by the nose. (And either I am just not on their
|
|
wavelength, or else they are all doing a very bad job of producing an
|
|
intuitively easy to use user interface.) Producers of free software often
|
|
even earn their money by producing GUIs for special applications under Windows,
|
|
and, in their free time, produce something that they think is the way the thing
|
|
aught to be done. It is unlikely that they will then go and replicate the
|
|
style they're paid at work to produce in their free time at home.
|
|
|
|
The consequence is that, unless somebody pays programmers to produce these
|
|
GUIs, I very much doubt that we shall see many of them with Linux. This is
|
|
sad, but I think the reality. Linux is ideal in the academic world as it
|
|
supplies many of us with all we need and want (except powerful hardware), but I
|
|
think it is very unlikely that it will make any inroads into commercial
|
|
offices. Besides, being your own system administrator with Linux alone sets a
|
|
prerequisite to the use of Linux that precludes a very large number of PC
|
|
owners from using Linux.
|
|
--
|
|
P.D. Dr. Michael L. Dowling (__) on.dowling@zib-berlin.de
|
|
Abteilung fuer Mathematische Optimierung (oo)
|
|
Institut fuer Angewandte Mathematik \/-------\
|
|
TU Braunschweig || | \
|
|
Pockelsstr. 14 ||---W|| *
|
|
38106 Braunschweig, Germany ^^ ^^ Ph.: +49 (531) 391-7553
|
|
|
|
|
|
------------------------------
|
|
|
|
From: daver@MCS.COM (Dave Rossow)
|
|
Crossposted-To: aus.computers.linux,alt.linux.sux,alt.os.linux,comp.os.linux.admin,comp.os.linux.help,comp.os.linux.misc
|
|
Subject: Re: Linux Inside T-Shirts, Now Printing!
|
|
Date: 2 Sep 1994 22:11:25 -0500
|
|
|
|
jhs@dfw.net (Justin Scott) writes:
|
|
|
|
|
|
>Any type of JPEGs, etc we can see of the shirts before we order?
|
|
|
|
>I would love to have the "Linux Inside" as will as the "GNU Generation"
|
|
>shirts, but only if I can see pics before purchase
|
|
|
|
>Justin
|
|
|
|
Likewise!
|
|
|
|
dave
|
|
daver@mcs.com
|
|
|
|
|
|
------------------------------
|
|
|
|
From: djohnson@arnold.ucsd.edu (Darin Johnson)
|
|
Crossposted-To: comp.os.linux.misc
|
|
Subject: Re: PGP Signature (Was: Suggest:SCSI Tape File System)
|
|
Date: 02 Sep 1994 05:57:14 GMT
|
|
|
|
> If I write a message and sign it with my
|
|
> public key (that you have and you *KNOW* comes from me) you can tell 2
|
|
> things: 1) that I wrote it and 2) the message text was not tampered with.
|
|
|
|
That also begs the question of whether or not your message was
|
|
important enough to even care if you really wrote it or not.
|
|
|
|
Now maybe if you were my boss telling me to get ready and head
|
|
over to meet with the European office for the week, I'd be concerned.
|
|
But I've seen nothing so far on the net (well over 10 years) where
|
|
it mattered who posted or not.
|
|
--
|
|
Darin Johnson
|
|
djohnson@ucsd.edu
|
|
Support your right to own gnus.
|
|
|
|
------------------------------
|
|
|
|
From: mpn@AlleleB.Berkeley.EDU (Mark P. Nelson)
|
|
Subject: Re: ext2fs corruption in 1.1.47-48
|
|
Date: 2 Sep 1994 20:43:53 GMT
|
|
Reply-To: mpn@alleleb.berkeley.edu
|
|
|
|
Just to let you know:
|
|
|
|
Started 10 months ago with
|
|
0.99pl12
|
|
ext2fs 0.3
|
|
2xIDE
|
|
|
|
Now got
|
|
1.1.45
|
|
ext2fs 0.5a
|
|
2xIDE
|
|
AHA1542CF
|
|
1GB SCSI
|
|
|
|
Gone through many intermediate kernels. Never had a corruption problem.
|
|
|
|
--
|
|
Mark P. Nelson (mpn@alleleb.berkeley.edu)
|
|
While I'll admit that anyone can make a mistake once,
|
|
to go on making the same lethal errors century after
|
|
century seems to me nothing short of deliberate.--V.
|
|
|
|
------------------------------
|
|
|
|
From: hm@ix.de (Harald Milz)
|
|
Subject: Re: What is happening with MGETTY 0.2
|
|
Reply-To: hm@ix.de
|
|
Date: Fri, 2 Sep 1994 16:04:23 GMT
|
|
|
|
In comp.os.linux.development, Tiger II (robisojf@uc.edu) wrote:
|
|
> What is happening to MGETTY 0.2? I have gone to several sites to pick
|
|
> it up and can't.
|
|
|
|
Look in
|
|
|
|
sunsite.unc.edu:/pub/Linux/system/Serial/mgetty+sendfax-0.20.tar.gz
|
|
tsx-11.mit.edu:/pub/linux/sources/sbin/mgetty+sendfax-0.20.tar.gz
|
|
|
|
and associated mirrors. What's "several sites" for you?
|
|
|
|
--
|
|
It's not an optical illusion, it just looks like one.
|
|
-- Phil White
|
|
--
|
|
Harald Milz (hm@ix.de) WWW: http://www.ix.de/editors/hm.html
|
|
iX Multiuser Multitasking Magazine phone +49 (511) 53 52-377
|
|
Helstorfer Str. 7, D-30625 Hannover fax +49 (511) 53 52-378
|
|
Opinions stated herein are my own, not necessarily my employer's.
|
|
|
|
------------------------------
|
|
|
|
From: becker@cesdis.gsfc.nasa.gov (Donald Becker)
|
|
Subject: Re: VLB IDE - DOS speeds up but Linux does not
|
|
Date: 2 Sep 1994 23:41:07 -0400
|
|
|
|
In article <33rp1q$ojr@nermal.cs.uoguelph.ca>,
|
|
Harry C Pulley <hpulley@uoguelph.ca> wrote:
|
|
>I'd like to thank all who said I should get 1.1.45 for better disk performance.
|
|
>I grabbed that and bdflush 1.4 and built them without any trouble at all.
|
|
>
|
|
>I know get ~1MB/sec. which is 30% faster than I had before. This provides a
|
|
>noticable performance improvement when booting, starting X and doing other disk
|
|
>intensive tasks.
|
|
|
|
In the same vein, I would like to publically thank Mark Lord (I'm guessing
|
|
he's the right person) for the IDE performance improvement code. Combined
|
|
with the bdflush (another advance), it's made a major improvement in IDE I/O
|
|
bandwidth.
|
|
|
|
As many of you know, my current project is "Beowulf Linux clusters". This
|
|
is a project to put together a off-the-shelf hardware specification and
|
|
software distribution to do very cost effective distributed computing. Last
|
|
week the hardware arrived for the first full development cluster of 16
|
|
machines. As part of the assembly/test process I configured a master disk, and
|
|
started duplicating the whole 528M disk to the other 15 disks. Finding this
|
|
an excellent opportunity to do something interesting, I started measuring the
|
|
peformance and changing the parameters.
|
|
|
|
The first attempt was using a stock 1.0.9 kernel. I tried many different
|
|
disk copy method, including:
|
|
cp /dev/hda /dev/hdb
|
|
dd if=/dev/hda of=/dev/hdb bs=<various>
|
|
cat /dev/hda > /dev/hdb
|
|
Almost independently of the method, the full disk copy took about 37
|
|
minutes. I settled on using the most obvious method,
|
|
'time cp /dev/hda /dev/hdb', for the rest of the tests.
|
|
|
|
Next I tried the latest development kernel, 1.1.49. This kernel includes
|
|
the IDE performance patches, and uses the 'bdflush' continuous-buffer-flush
|
|
code.
|
|
|
|
The first test amazed me -- an stock kernel, with the default parameters,
|
|
completed a disk-to-disk copy in just under ten minutes, 9:54 to be exact.
|
|
By using the 'hdparm' program I was able to tune the copy to the fastest
|
|
time of 6:10 before I ran out of disks to copy. The largest benefit came
|
|
(predictably) from increasing the read-ahead size from 16 sectors to 64. A
|
|
smaller improvement came from enabling multiple sector mode, and turning on
|
|
interrupts made no significant difference in the performance.
|
|
|
|
The bottom line is that for this task the new kernel is 3.5 times to 6x
|
|
(when tuned) faster than the old. I think this is a significant difference.
|
|
|
|
BTW, these numbers were measured on a 100Mhz DX4 system with a SiS '471
|
|
chipset, all memory and cache parameters set to their fastest values, and
|
|
the VL bus set for one wait state. The IDE controller was a DTC brand VLB
|
|
multi-I/O card jumpered for the fastest transfer rate, "8.3MB". The disk
|
|
was a Maxtor 528MB IDE drive ("540M" advertised), and both disks were on the
|
|
same controller.
|
|
|
|
--
|
|
Donald Becker becker@cesdis.gsfc.nasa.gov
|
|
USRA-CESDIS, Center of Excellence in Space Data and Information Sciences.
|
|
Code 930.5, Goddard Space Flight Center, Greenbelt, MD. 20771
|
|
301-286-0882 http://cesdis.gsfc.nasa.gov/pub/people/becker/whoiam.html
|
|
|
|
------------------------------
|
|
|
|
From: broadley@turing.ucdavis.edu (Bill Broadley)
|
|
Subject: Re: Future of linux -- the sequel
|
|
Date: Fri, 2 Sep 1994 20:28:51 GMT
|
|
|
|
Hmmm indy vs pentium 90 eh?
|
|
|
|
Whats your price on a indy? Which cpu? What Mhz?
|
|
|
|
I ordered the following:
|
|
p-90, 16 MB ram, 512k L2, stealh 64 2MB, 1GB scsi, 17 mag MX (sony tube),
|
|
ethernet for $4000
|
|
|
|
I'd have to add a ccd camera $500 (guess) or so, OS (linux), and 16 MB ram
|
|
$500. $100 for sound.
|
|
|
|
So we are at $5,000. with about specint 90 spec fp 72. The low end
|
|
indies (i.e. with 1 GB under $7,000 are in the specint 34-60 range.)
|
|
|
|
: 2) Pentium systems in general:
|
|
: a) have lower resolution/slower video hardware
|
|
|
|
Above can do >= 1280x1024
|
|
|
|
: b) have smaller hard disks
|
|
|
|
Has 1.0 GB
|
|
|
|
: c) have less RAM
|
|
|
|
Maybe... $500 to go from 16 MB to 32.
|
|
|
|
: Configure a Pentium system which is identical to an SGI
|
|
: Indy and they will have very similar price/performance,
|
|
: even though the Pentium PROCESSOR is more expensive
|
|
: than the MIPS processor.
|
|
|
|
Which indy can patch 100 specint like the pent 100?
|
|
|
|
: Note that, at the price shown, the PC will not do full motion video or
|
|
: capture images, nor will it be as fast overall as the SGI.
|
|
|
|
Why not as fast? Indy does over 230k xstones, gets better specs int/fp?
|
|
|
|
--
|
|
Bill Broadley Broadley@math.ucdavis.edu
|
|
Linux is great. Bike to live, live to bike. PGP-ok
|
|
|
|
------------------------------
|
|
|
|
From: grante@reddwarf.rosemount.com (Grant Edwards)
|
|
Subject: Re: PRIORITY make an undelete command
|
|
Date: Fri, 2 Sep 1994 18:20:43 GMT
|
|
|
|
Rob Janssen (rob@pe1chl.ampr.org) wrote:
|
|
|
|
: >BTW, how does DOS' UNDELETE help if you do "COPY MAIN.*", where files
|
|
: >MAIN.C and MAIN.H match the wildcard?
|
|
|
|
: The wildcard is passed to COPY, not expanded by the shell.
|
|
: COPY sees only a single argument, and won't copy the first match over
|
|
: the second.
|
|
|
|
Depends on the shell. I use ksh under dos, and if I type "copy
|
|
main.*" ("copy" is aliased to "cp") I'm screwed just like on a Unix
|
|
system.
|
|
|
|
The fact that command.com is too dim to expand wildcards before
|
|
passing parameters to a program isn't really relevant to the undelete
|
|
issue.
|
|
|
|
If you don't like they way the "cp" "rm" or shell programs work,
|
|
(under DOS, Unix, Linux, VMS/DECshell, Ultrix, whatever) then don't
|
|
use them -- find or write something else. It's not a kernel problem.
|
|
|
|
There are many thousands and millions of things I can type under DOS
|
|
and/or Unix to screw myself over. There's no way the system can
|
|
prevent you from ever hurting yourself. If you've got enough rope to
|
|
do useful work, you've got enough rope to hang yourself.
|
|
|
|
--
|
|
Grant Edwards |Yow! Look! A ladder! Maybe
|
|
Rosemount Inc. |it leads to heaven, or a
|
|
|sandwich!
|
|
grante@rosemount.com |
|
|
|
|
------------------------------
|
|
|
|
From: sander@ankh-morpork.hacktic.nl (Sander Plomp)
|
|
Subject: Re: Netware Client
|
|
Date: Fri, 02 Sep 1994 10:39:00 GMT
|
|
|
|
wcattell@netcom.com (William B. Cattell) writes:
|
|
|
|
>I have been successfully accessing multiple NetWare servers by running
|
|
>DOSEMU un Linux. Since NetWare is a DOS based OS inorder to function as
|
|
>a true NetWare client you need to run DOS (or OS/2 or a MAC). NetWare's
|
|
>support of non-DOS clients is ok but not great.
|
|
|
|
NetWare isn't really DOS based. The OS that runs on the servers is a
|
|
completely different OS. It has a weird history, because it dates back
|
|
to the days of DOS 1, and has many hacks to support DOS oddities. But deep
|
|
down it is more like UNIX than like DOS.
|
|
|
|
But NetWare clients for many OS-es exists, and it should be possible to
|
|
make one for linux just as well.
|
|
|
|
>You can mount NetWare volumes via NFS and access the server console via
|
|
>XConsole -- as you've already seen -- but to access the server as a true
|
|
>client you would need some kind of DOS. This also depends on what you
|
|
>want to do. If you just need file access NFS should suit your needs. If
|
|
>you need to do NetWare admin (users, maint., etc.) then you would need a
|
|
>DOS client.
|
|
|
|
To do admin you need a client that allows access to the bindery, the
|
|
queues, and all other 'non filesystem' features of NetWare. Native DOS
|
|
doesn't support this either, it's the programs such as syscon and pconsole
|
|
that 'speak' this part of the protocol. Again, it is possible to support
|
|
this protocol under linux.
|
|
|
|
It is an horrible amount of work, especially to support all possible
|
|
generations of NetWare.
|
|
|
|
>
|
|
>Bill Cattell
|
|
>
|
|
--
|
|
Sander Plomp || "Yes or no Humprey!"
|
|
|| "Yes *and* no minister."
|
|
sander@ankh-morpork.hacktic.nl || -- The diaries of
|
|
|| the Right Hon. James Hacker, MP
|
|
|
|
------------------------------
|
|
|
|
From: remco@emc.rvt.com (Remco Treffkorn)
|
|
Subject: Re: PRIORITY make an undelete command
|
|
Date: Thu, 1 Sep 1994 08:18:55 GMT
|
|
Reply-To: remco@emc.rvt.com
|
|
|
|
Robert Sanders (rsanders@mindspring.com) wrote:
|
|
: I know everyone will scream "bloated pig" when I mention Emacs, but
|
|
: I'm totally in love with the Emacs version control interface. Three
|
|
: keystrokes enters a file into the version control system of your
|
|
: choice (RCS or SCCS). Once you've done that, you need only two
|
|
: keystrokes to check a file in or out. While editing a file, three
|
|
: keystrokes shows you the differences between the current version and
|
|
: the last version available in your version control system. And that's
|
|
: the simple stuff; it has the usual all-singing, all-dancing Emacs
|
|
: features piled on top.
|
|
|
|
Well, two keystrokes and this thread vanished in my kill file.
|
|
|
|
This was one of the most religious wars I have ever seen.
|
|
More appropriate for alt.comp.philosophy.
|
|
|
|
My apologies to Robert. It is not his post, but the totality of posts
|
|
that upset me. After 60 odd posts without any content it is just enuff ;-)
|
|
|
|
Remco
|
|
--
|
|
|
|
Remco Treffkorn, DC2XT
|
|
remco@emc.rvt.com
|
|
(408) 685-1201
|
|
|
|
------------------------------
|
|
|
|
|
|
** 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
|
|
******************************
|